Discover how Emburse drives real AI adoption by creating a culture of experimentation. Heidi Wurpel shares why giving teams “permission to play” accelerates agentic marketing.
In this episode, Sarah McConnell sits down with Heidi Wurpel, VP of Marketing at Emburse, to explore why the true foundation of AI adoption isn’t tools or technology. It’s culture. Heidi breaks down how giving teams permission to play with AI accelerates learning, sparks innovation, and builds the confidence needed to scale agentic marketing across an organization.
She shares how Emburse empowers marketers to experiment safely, celebrate early wins, and iterate fast, creating an environment where AI isn’t intimidating. It’s exciting! From modernizing content workflows to training AI SDR agents like Kira, Heidi explains the mindset shifts leaders need to make if they want AI to actually work inside their teams.
What you’ll learn:
TRANSCRIPT
Sarah McConnell – Qualified
Heidi, thank you so much for joining me today on The Agentic Marketer. I'm so excited to have you. Before we jump into anything, I'd love for you to tell me a little bit about yourself and the work that you're doing over at Emburse.
Heidi Wurpel – Emburse
Sure. So my job is the Vice President of Emburse and Marketing. And I basically take what is currently one of the world's largest travel and expense management companies to even greater heights by spreading our expense intelligence story powered by our incredibly innovative AI. And I've been in marketing for about 20 years now, starting in brand and corporate comms, luxury B2C during the dot-com boom and bust. So that was fun. And then I acquired my MBA, worked in market research and data, spent a little over a decade in tech and ops, and now, of course, I'm focused on leadership, including demand gen and go-to-market for Emburse.
Sarah McConnell – Qualified
Hahaha! Amazing. I love what you guys are doing over at Emburse. I've used your product before. I never realized until I got into tech how cumbersome travel expenses could be. So I love what you guys are doing over at Emburse. I've used the product. It is definitely a huge time save.
Heidi Wurpel – Emburse
We love it, and we're doing even more. It's coming down the AI roadmap to make it easier on you day in and day out, because travel can be fun, and we want it to be.
Sarah McConnell – Qualified
Amazing. Okay, so obviously, the title of the podcast, The Agentic Marketer. So Heidi, it's a new term. People aren't quite as familiar with it. How do you define agentic marketing, and what do you think it means to be an agentic marketer right now?
Heidi Wurpel – Emburse
Yeah. So I actually think agentic really is just being—like, it's a data-based way of saying the right thing to the right person at the right time. Think of it as a personalization mass multiplier. And I think that agentic marketers in particular are really fixated on getting prospective customers the answers they need as quickly as possible, as easily as possible, and truly tailored to their unique problems. And the data sets are what allow us to do that today in technology.
Sarah McConnell – Qualified
I really like that answer. I feel like it's such a—people, you know, agents, AI, it's a very loaded term, but I really like the way you phrase that, if it's more about getting people the exact right thing at the exact right time. And now with AI, we can do that so much better than we were able to before. Like, we've been talking about personalization and marketing for a really long time. That's not new. It just feels like we've entered a very new time where that personalization is next level. It's just so much different than it was.
Heidi Wurpel – Emburse
Yeah, I think that today data and technology are enabling us to do what—like, if I get on a little bit of a high horse here—like people in proximity. Again, I was in… my gray hair comes naturally. Like, I was here during the dot-com boost, right? And then a little bit of a burst. And when you think about it, in the past, before e-comm went wild, before things went huge, you generally interacted with the same people day in, day out. And they were local people, right? So they knew your preferences. They knew maybe your family's preferences, et cetera, et cetera, et cetera.
And I think we have this deep want to return to that. But we've gone so far away from it these past 30-plus years to sometimes really good things like price transparency and globalization. There's definitely positives about what we've done these past 30 years in marketing. But I just think today our databases, our data structures, and most importantly really the agents that allow us to interact with that data and those databases and structures are really what differs from traditional marketing. But it's almost a return to like traditional-traditional marketing, like the pre-e-comm days of the world, in my mind.
So I'm pretty excited to be able to connect people with the information that they need faster and more easily and let them feel seen and known and like we're adapting to their unique needs in that way.
Sarah McConnell – Qualified
Yeah, I like this return to the traditional-traditional marketing. And I always think about—over the time in my career—everything in marketing feels very cyclical. Like, you know, there's always this dead, like, “This isn't used anymore,” and then it always comes back. And the example I always use is outdoor billboards. I remember no one was paying for billboards. If you asked, “Can I go spend this money on a billboard?” they would laugh at you. Absolutely not. And all of a sudden, if you're in the Bay Area, they're everywhere. Everything does feel very cyclical.
The traditional marketing—it's just evolving, but those basics are still there, and we do need to return to those. So I really like the way that you phrase that, because it is almost like we're just going back to it.
Heidi Wurpel – Emburse
Yeah, it's just a new tool, right? Like, I think there used to be a magnet up in the marketing department at the last organization I worked at. It's always saying—this is during the app bubble, really—“An app is not a strategy; it is a tool.” And I think of AI and agentic AI in that same way. Agentic AI is not a strategy in and of itself. It is a tool that enables us to do what we want to do faster, better, more easily, and maybe getting our end users and prospects what they need quickly in that space. Yeah.
Sarah McConnell – Qualified
Yeah. And to your point of it being a tool, I'm curious—with this new tool that's agentic AI—it’s really easy for people to experiment. I think just given things like ChatGPT and your ability to build, and Gemini, and Perplexity, and all of these things, people are experimenting more heavily with this technology than we ever have with previous technology just because of the availability of it. But if you want to move from that experimentation to actually leading with it and using agentic AI as a tool across your entire organization, what do you think it takes to make that shift?
Heidi Wurpel – Emburse
Yeah, I think having extremely clear goals and strategies is important. You need to know what you want your outcomes to be and be very focused on that. And I think for many of us as marketers, that means having excellent original content repositories, which include concise, fulsome, accurate product marketing assets and sales scripts—things that you want to become external-facing, but they need to be there. They need to be concise. They need to be right.
I think you also need clean data and integration capabilities to fully scale this thing. But I think more than anything, what we've found in our organization is you also have to have a culture that's really willing to rapidly pivot and use. And I don't know if this is—again, I spent over a decade in tech, so I don't know that all marketers are comfortable with CI/CD—but continuous improvement, continuous development principles are that idea that instead of changing the whole app user interface, make the button blue with a little more rounding. It's not a whole rebrand; it's a little refresh or a little evolution on what you've been doing.
And I think if you have a team who culturally is comfortable with that—these fast pivots in small ways where maybe it's not going to come to great fanfare, maybe you're going to get pushback on “That's wrong, that's not good, people don't like that”—okay. What's cool about agentic AI is great: you can train the model rapidly and easily to adapt. It's that adaptability again—continuous improvement, continuous development—constantly checking in and making those changes to train the models and make them better and better and better over time.
And realizing—you know, like you and I were talking about billboards—right? You need to get a billboard right. You print a billboard; it's the billboard. It was expensive to paste. And these days with digital billboards, you have a little more leeway, but non-digital billboards—what's on the billboard is what is on the billboard. If you have a typo or an error, that's on you.
That is not agentic AI. It is a living, breathing, changing asset day in and day out. And so I think that culture really matters. But again, have those really clear strategies, those really clear goals, and then start thinking about the backend of what's going to feed that, and then be willing to feed it rapidly and pay attention to it like you would, I don't know, any employee or any piece of information that you're putting out.
Sarah McConnell – Qualified
And we tell people that all the time. So I really like that you hit on the point of it being continual, and you have to go back and be willing to rapidly train and test. I think in the age of AI, things are moving fast—your company, how you're putting out information. And we obviously have our own AI SDR; her name is Piper. But when we talk about onboarding her as an agent, I always caution people: this isn't a set-it-and-forget-it. You can't just give her information—or whatever you end up personifying your agent as—you can't just give it information and expect it to keep up with your rapidly changing organization.
Every time something on the marketing team changes—your product marketing, your messaging shifts, you release new assets, you have a new event—it does have to become a part of your workflow if you want to lead with AI and not just experiment with it. You have to give that new information to any of the models, and that involves any of your agents or AI that you're using internally, to utilize that information and then make sure it's using it properly and give it that feedback.
And I think, to your point, Heidi, that is where people kind of fall off on going from experimentation to leading, because they tend to try it, they forget about it, then all of a sudden it's not giving quite the right answers or it's not doing it exactly the way they wanted. But it does involve pretty good oversight from the people on your team, I think.
Heidi Wurpel – Emburse
Yeah. You know, I recently read an article, and I'm not going to get his name right. I think it's Joe Carveo. He's a marketer-operator for a PE firm. And Joe wrote this thing about how, on his hierarchy, he actually slates agentic agents in as if they're people on his hierarchical org chart, so that it's really clear who the manager of that agent—or that person, what would traditionally be known as a person—is.
I don't know. That kind of blew my mind the other night. It's not my hot take; it's more of his hot take. But I think that was a really interesting take on AI and how we train it. And I think that's part of what made us go to Qualified. And we named ours Kira—you named yours Piper. So part of what we really liked was this idea that you could give her—you know, personifying here again—but you could give her information, and the moment that we update a blog, the moment that we update our landing pages, the moment that we update our data sheets, she’s instantly learning all of that information. She's ingesting that and finding ways to share it out.
Versus the old way of things where we had to go, “All right, there’s a playbook that relies upon that information. Go and remember to update that playbook.” And you know, like you and I talked about earlier, Emburse is an incredibly innovative company. We are rapidly releasing features and products and bringing new things to market. So to keep up all of those data sheets, to remember every playbook that's ever been written and redo your script—the nuance of that? Huge time sink. Huge.
The person who was managing it was just like, “My gosh, this is going to be in 30 places. This is going to be in 60 places. This is going to be in 300 playbooks.” That's hard. That's really hard. And it's not the modern way of doing things. And so I think that that's part of what was really great about Kira—or Piper—and Qualified: we could rapidly train her with the information that we're already bringing out to market.
Also, that would be good when you launch a campaign. You want certain information to hit the market at certain times. So it kind of takes care of any worries like that too—that her information is going to be as up to date as your marketing releases and very aligned with your campaign strategy. So those were some unique things that I thought really are just part of an agent that differ slightly from a human being doing enablement and making sure they don't say anything before X, Y, Z date or whatnot.
Sarah McConnell – Qualified
Yep, that is so true. Now, you mentioned you renamed your AI SDR Kira, which is great. I'm curious: what was it about an AI SDR that had you and the Emburse team leaning into that? I know when AI SDRs hit the market, there was a lot of confusion around them. So what was it that made you and the team say, “We think this is something that will be important to incorporate into our demand gen motion and our larger marketing organization”?
Heidi Wurpel – Emburse
Yeah, I think a couple things. One, my company, Emburse, is leading with AI. We are an AI-based company. So our culture embraces AI. It believes in AI. It's bringing really cutting-edge AI to market at times. So the culture was there. The trust in AI was there. And that's really important.
I think then, from a pure marketer standpoint, saying we need to be able to do just as much quicker. How do we make sure that we train with accurate information quickly and not have to rewrite all those if/then statements that were in the old playbooks in the way that we used to do things?
Having Kira to talk about our innovative AI roadmap, as well as trusting her AI to do what she does best—which is share, again, that right information with the person who's asking at the right time on a regular basis when it's needed.
Sarah McConnell – Qualified
Yeah, that's amazing. And now, is there anything unexpected that bringing on Kira changed for your team? Was there something you weren't expecting that you've now noticed having her as your AI SDR?
Heidi Wurpel – Emburse
Yeah, you know, just the other day I was reading through the questions being asked of Kira, and I noticed it's great to get content guidance. And I think I knew this, but it wasn't what made us make the decision. But it's really nice—it shines a light for me on those areas that our marketing is not clearly answering. And that's why earlier I was like, “Be concise, be accurate.” Because if you see the market asking the same question over and over, the likelihood is what you have out there right now is not enough. It's not answering this question.
So I think I've been pleasantly surprised—or maybe pleasantly reminded—of the fact that you can see trends in the data, and that that can even help inform your content strategy and even bigger strategies if you want to explode it up a little bit more for future years in your marketing.
Sarah McConnell – Qualified
Yeah, that's one of the unexpected things we experienced onboarding Piper at our company. I got to oversee that, which was amazing. And I remember looking at the questions being asked. And I tell customers all the time: you're going to have your nine to ten questions that get asked pretty consistently on your website—that is universal. But there are going to be aggregates of questions that you don't think about that really will help inform your content team.
And that actually kind of shifted our org structure here at Qualified to have content and communications roll up under my team on demand gen, kind of for that reason. I was getting this feedback from the questions that were being asked. I was able to easily see it in Qualified. And I was like, I need to be able to move fast and make sure this content is now out there because I'm obviously seeing in real time we aren't answering that question well enough. We don't have the resources out there in a way that's making it easy for visitors to find us.
Now granted, we can train Piper on it, and that's fine. But we want to make sure across all of our different channels that we have the right information. So it was something that actually impacted our team pretty heavily that I was not expecting.
Heidi Wurpel – Emburse
Yeah. And I think a few years back, I was at an SEO conference, and it was really interesting. It was talking about looking at Google’s other questions that are asked—“People also ask.” And I know that's kind of a thing of the past now. It's way at the bottom. But at the time, that was at the top. It was sort of that “aha” moment where you're like, right, you can use AI to answer all these questions and make sure you get similar content and questions answered on your site because maybe you're thinking about this—expand it a little to that.
And I think that's what this agentic AI is really doing for you, but it's doing it in a hyper-personalized way. It's yours. It’s not like Google used to be—like, the general questions out there in the market are… Well, this is hyper-personalized to: “Hey, Emburse marketing team, ten people asked this thing.” Clearly you need some more information about that on your website and across your channels. So yeah, that was a really pleasant surprise.
Sarah McConnell – Qualified
Now, I'm curious—what metrics are you measuring to know that Kira is working, beyond obviously the content that's harder to measure? Is there anything else that you're looking at from a metrics standpoint pretty consistently?
Heidi Wurpel – Emburse
Yeah, you know, I mean, for us, I'm really proud of our ability to see attribution. And we linked Kira into our Salesforce. And so we're able to see specific MQAs that come through—our marketing qualified accounts—and be able to connect them to: this was a Kira conversation, or this was a Kira-qualified ask for a meeting booked. So we look at MQAs; we look at meetings booked.
And we've seen about a 212% improvement in the past six months alone. So really good numbers there. And then, of course, I'm watching her accuracy of responses, as is my Head of SDRs. He's really fabulous because we look at her as an SDR in that meetings-booked and sort of initial discovery, “disco” type of phase as well. We treat her a bit like an SDR meets a marketing lead source, and we measure her on both of those metrics.
Sarah McConnell – Qualified
That's amazing. And now I want to talk a little bit more. It's very obvious from this conversation that Emburse is—you mentioned—very AI-first, not only as an organization and how you're adopting technology, but also as a company in general. So I feel like you are more than qualified to answer a question for someone if they're looking to adopt AI.
So say someone's listening to this and they are not where I feel like you or Emburse are at, and they're like, “I don't know where to start. This feels very overwhelming.” What would be your advice, having gone through this journey already, that you would tell other marketing leaders? Where should they start when they're looking to adopt AI or become more adept at agentic marketing?
Heidi Wurpel – Emburse
Yeah, I think AI adoption really does come culturally. And if you're a leader in your organization, I think you can encourage your people to play, right? What's nice about being a marketer is most of the information that we share is information that we want the world to see. So I think a lot of the time people are really concerned. So you can put guardrails up about: don't share anything company-confidential. But if you want the world to see the information—share the information you want the world to see.
Go ahead and play a little bit. Sometimes you're going to flop and fail. Other times you're going to have wild success. And I think for me, one of the first things that I did was—in my digital innovation team—they really started playing with it, and we started highlighting those areas of success. We were highlighting their successes to the rest of marketers, and that gave sort of permission to play. Like, “Hey, so-and-so did this. It's not 100%. But look how great it is. This is what it allowed us to do.”
And so we took a few problem statements that we'd been putting on the back burner for a while because we thought, gosh, finding a solution to that is so big and so hairy. And then—we’re a Google shop—so we were like, “Wait, we have NotebookLM.” Let's just play with it and see what it can do. And we were like, “Wait, that really worked.” And we could then expand it here and there. What other uses could we do?
And so I literally have a monthly show-and-tell where my team brings really cool new things they've shared. I learned about Nano Banana from them before I saw any press releases on it. So it's just really neat to see the different qualities of your team. So I think allow people to play within these safe data sets and within this safe information—nothing that's going to hurt or not be able to be shared out there.
So encourage the culture and, like I said, celebrate those successes.
Sarah McConnell – Qualified
Yeah, absolutely. I really like—we’ve heard that answer quite a bit on this podcast, and I just cannot emphasize it enough. If you want to lead and be at the forefront of this, you have to not only allow your team to use AI—which, it's wild to me, but there are still companies out there that aren't letting their employees do that. But I like what you pointed out: champion them. Champion the people that are doing it well. Give them spaces to come to, to show off what they're learning and what they're doing.
I think that's such a fast way to get the team excited about something and willing to get their hands in it and start using it themselves. Being able to champion those employees, I think, is so, so important right now.
Heidi Wurpel – Emburse
Yeah, 100%. You know, I really think sometimes as leaders—we all try to be approachable, right? We want to be really approachable to our team. But on the flip side, how often do you want to go to your boss and be like, “I don't really know about this thing”? Like, teach me about this. I don't know—maybe there are people who do. But a lot of us—the A-plus students—we want to come forth always with the A-plus things.
And so I think what this allows is you create these little knowledge experts. And you don't have to spin them as such, but these cohorts that can come together and privately say, “Hey, it’s really cool what you did there. Can you show me that? Can you teach me that?” And instead of having to go up the chain, it creates those horizontal connections and networks that allow people to ask each other the silly questions and really have respect.
Because if you're someone who has literally never typed a prompt into anything—and I think there are people absolutely still out there in the world, and no shade on that—that first prompt can feel a little scary. And literally just having another colleague there next to you to type the prompt, and have them watch and ask, “And now what do I do? And what button do I push?” It feels really nice—just to give that safety and that mechanism of: you can't screw this up. It's really okay. Go ahead and try something simple.
So yeah, that's what we've created in our marketing culture. And I'm really proud of the team, because holy cow, they bring things that I would never think to bring forth, and solutions to each other. But then just iterate on one another and have saved a great deal of time and energy and wasted cycles.
Sarah McConnell – Qualified
Now to that point—are there any particular skills you've seen within the team that you think other companies really need to double down on in order to thrive in this AI-driven era? Are there particular skills that you've seen within your team or yourself that will really help teams make that shift into an AI-first company?
Heidi Wurpel – Emburse
I think skill-wise: clarity, again. Being able to really know what you want to get out of something and really clearly articulate that the outcome should be like this, or that the output is for so-and-so.
I think having creativity is really important—being willing to ask the same question in four different ways and think in that way is really important. But I really just think it's that willingness to trust that the models will present at least something that you can start with and that you don't have to end with.
I love a blank canvas, but there are days when we are moving so fast that you don't have time for a blank canvas. It's really helpful to have—and I'm super geeky in the art space—for those of you who haven't painted, you have a canvas. You paint this white or pink or whatever color layer you want on it to fill in all the little bubbles of the canvas. Yes, that's meditative and great if you're a professional artist, but if you're not—if you're just at a sip-and-paint—that’s not particularly fulfilling. So you let the AI do that.
There's a certain horizontal line you paint often on a canvas if you're doing a landscape or something, or even just to give yourself a little guidance. Similarly, let AI do that. It's not all or nothing with this. It's not “Paint me a Picasso,” and then “That doesn't look like Picasso, I hate it.” Just let it whitewash the canvas, and you paint the rest of the image on it.
So I think that's it. It's not all or nothing. It's not black and white. Always, always edit whatever your AI is putting forth or retrain it in the agentic world. And so I think it all comes down to that willingness to iterate and have clarity of your final vision—you know what you want on that canvas at the end.
Sarah McConnell – Qualified
I think that's the best analogy I've heard for how to let AI support you versus replace you. I really love that. I might have to use that in the future because it's very clear, but also very beautiful.
Okay, Heidi, I have a couple of questions now looking forward. I know it's always hard when we're like, “Get out your crystal ball and tell the future,” especially in an AI era when things are moving so fast. But I'm curious—if you are thinking ahead 12 months, what do you think, besides AI, is going to be the single biggest change to B2B marketing? Like, what is AI doing right now that's going to make the biggest shift in the next 12 months for us B2B marketers?
Heidi Wurpel – Emburse
I actually think it will be more than just about marketing. So I think it's going to have a big impact on go-to-market. What I mean by that is: I think that we see that AI is an accelerator. It's a pace accelerator; it is a volume accelerator. And I think if we do this really right, we're going to see the sales velocity also decrease. Right?
When you think about the velocity of a sale—and this is more B2B than B2C—B2C, I see this shirt, I want to buy this shirt, I buy it off the rack. Maybe it takes a little while to ship if I'm doing e-comm. But when you think about B2B marketing, it takes time often in non-product-led growth models to go through the various phases of the sales cycle. And we know people do a ton of their research ahead of time and do it within marketing.
So done right, I think we're going to see the velocity of sales motions shorten. And I think we may even see disruptions to your traditional go-to-market sales plays, with an increase in product-led growth—even for those large global enterprise companies and expensive software.
So I think it's going to impact go-to-market if marketers do this correctly and create velocity acceleration on some deals.
Sarah McConnell – Qualified
Now on the flip side—if you're again thinking 12 months out—in the next year, is there something you worry marketers are still going to be getting wrong about AI a year from now? What do you think we're still heading down the wrong path on that we might still be doing incorrectly?
Heidi Wurpel – Emburse
Yeah. I think this concept of replacing entire roles or sets is still a little off. Do you need to hire as many of the same person? Maybe not. But that doesn't mean you don't need a product marketer anymore, or you don't need a content marketer anymore, or you don't need creatives anymore—just let the AI do it.
There’s that idea of keeping to your brand and keeping to your true north—that’s really important. So I think we could move too fast and we could get that wrong, for sure.
I'm hesitant when I hear people… you know, I love putting an AI head on the hierarchy, like I said. And I think that there's a lot we can do. We can increase pace. But I think that is something that could go awry if we let it go too far in efficiency, efficiency, optimization, optimization. Don't lose those skill sets. You need somebody to babysit the model.
I don't know about you, but I'm probably not going to hang an AI Picasso on my wall. I mean, I'm not going to hang a real Picasso because I don’t have money like that, but you know what I mean. If I'm hanging a contemporary, interesting portrait on the wall, I'm probably going to choose one that's out of paint from the hands of a true artist who put those final touches on it.
Sarah McConnell – Qualified
I cannot agree more. I do like the idea of incorporating AI into your org chart only from the standpoint of: it helps to know who's in charge. Who is the person responsible for making sure the AI and the model are running the way you want them to? But I do agree—I think it's always funny to watch things right-size. People go really hard, and I get it. As things change—as marketers—we tend to be on the forefront of wanting to test things and be on the cutting edge.
My CEO always says: if you're not a little uncomfortable, you're probably not pushing yourself hard enough, and that's good marketing. You really have to be kind of out in front of your skis a little bit. But in that process, we tend to right-size as things go on. And I think there was a shift of: could it replace entire teams? Could it replace entire roles?
And it's like—well—we’ve seen internally with SDRs: yes, we have an AI SDR, and yes, she can scale infinitely and do all of these things. But we’ve still found that we have changing roles with people that are now managing the AI SDRs. We have our inbound and our outbound AI managers. But also, we still have SDRs who are doing outbounding and writing really good personalized emails for our ABM, very hyper-targeted accounts.
So I do think having them work side by side is right now the key. That's where we're going to see the most benefit.
Heidi Wurpel – Emburse
I think so too. Again—it’s a tool; it's not a strategy. Your strategy needs to be clear and concise. You need to know what you're trying to get your outcomes to be. And part of your unique selling proposition is telling people clearly what your unique selling proposition is. Don't let the AI write that for you, because it's going to write a similar thing for all companies. It's a mass aggregator.
So you need to stay unique. You need to have those qualifiers. And I think that's the human touch and that set of control.
I also think something that—along with trying to carte blanche replace whole roles—we could get wrong is that we are so used to being able to control every word, every sentence, every message that comes out. And you have to be able to go with the flow a little bit when you move into an AI world.
If you have very specific and controlled outcomes that you want to occur—like “The company desires this sentence to always be written this exact way”—that's not going to happen anymore. What you need to change your focus to is ruthlessly focusing on your customer's needs and meeting those.
Does it get down to the heart of answering their question correctly, in an accurate way? Try to avoid hallucinations, but we know AI models today hallucinate. If you're someone who's very rigid, there are industries where you have to be very careful with every single sentence.
And what's cool about AI is you can say, “Anytime you use a sentence like this, write it this exact way.” You can train your AI SDR to do that. But I just think you need a little flexibility—a little, to your point, skiing analogy. If it's an icy day or a powder day, you're going to ski it differently. And you need to trust that AI is going to ski that terrain differently too. It's going to know what to apply, but the outcome is going to be different each time.
Sarah McConnell – Qualified
I totally agree. Okay, to wrap things up, Heidi, we have our lightning round. A couple of quick questions to wrap things up that are a little more fun.
The first one: besides ChatGPT—because that tends to be most people's answer—what was the first AI tool that you experimented with?
Heidi Wurpel – Emburse
PhotoLeap AI Headshot Generator. Yeah. My hands were weird, my ears were weird. I'm not going to lie. But a few came out okay. But it was actually an AI headshot generator. In the end, I did go and get a professional photographer—a real human. So here’s—right—I explored and said, nope, I think a human’s going to be better for this. But it was actually a headshot AI generator that I played with first and foremost.
Sarah McConnell – Qualified
I like that one. Okay, what do you think is the most overrated buzzword in martech right now?
Heidi Wurpel – Emburse
Ooh. The metaverse—that we all want to be digital AI avatars always plugged in. I think that's overblown. I don't think that's going to take over the world in the next 12 months anyway.
Sarah McConnell – Qualified
I'm with you on that one. You mentioned someone earlier, but is there one marketer who you think is really ahead of the curve on AI right now that you like to follow and you'd recommend others follow on LinkedIn or other channels?
Heidi Wurpel – Emburse
Yeah. So Erin Mills, who I have worked with before, but she's doing a really great other podcast. So if you're into podcasts, check out Erin Mills' podcast as well on LinkedIn. She talks about AI on a really regular basis and brings on some interesting guests. It's always fun to see her and what she's doing in this space.
Sarah McConnell – Qualified
Amazing. And last question—if you could automate one part of your life outside of work with AI, what would it be?
Heidi Wurpel – Emburse
Okay, I'm staring at a Monstera that needs a little bit of love. Probably my houseplant care. I have a green-enough thumb. It's green, but it's not deep verdant green here—they're struggling. If I had something that pruned them just the right way at the right time and moved them into the sun and gave them the right amount of water and food, I would absolutely invest in that. Because sometimes when you travel for work—and things which I hope everybody gets the opportunity to do because it's not that bad to do your reimbursements; use Emburse, it is fun—but when you travel, your plants do that death shrink. And that would be really nice.
Sarah McConnell – Qualified
Yeah, I would also invest in that. I'm thinking about the multitude of houseplants I have in my house that are doing okay—just good enough—but not great.
Heidi Wurpel – Emburse
That's… yeah. I'm sorry, Monstera.
Sarah McConnell – Qualified
Okay, Heidi, well, thank you so much for joining us on The Agentic Marketer. It was so great to have you today.
Heidi Wurpel – Emburse
Thanks for having me. What a pleasure.
.jpg)
Discover how agentic AI is reshaping modern marketing as Emplifi CMO Susan Ganeshan shares insights on lifecycle engagement, scaling teams, and reducing churn.
.jpg)
Discover how Katy Abbaszadeh at Glean uses agentic AI to scale outcomes, boost pipeline ROI, and transform traditional marketing.
.avif)
Learn how CMO William Tyree leverages agentic marketing and an AI SDR to improve lead response, refine buyer experiences, and drive pipeline growth.
.jpg)
Learn how G2 approaches agentic marketing, speeds up inbound, and adapts to LLM-native buyers in this conversation with CMO Sydney Sloan.
Discover how agentic AI is reshaping modern marketing as Emplifi CMO Susan Ganeshan shares insights on lifecycle engagement, scaling teams, and reducing churn.
Discover how Katy Abbaszadeh at Glean uses agentic AI to scale outcomes, boost pipeline ROI, and transform traditional marketing.
Stay up to date with weekly drops of fresh B2B marketing and sales content.
Discover how Emburse drives real AI adoption by creating a culture of experimentation. Heidi Wurpel shares why giving teams “permission to play” accelerates agentic marketing.
In this episode, Sarah McConnell sits down with Heidi Wurpel, VP of Marketing at Emburse, to explore why the true foundation of AI adoption isn’t tools or technology. It’s culture. Heidi breaks down how giving teams permission to play with AI accelerates learning, sparks innovation, and builds the confidence needed to scale agentic marketing across an organization.
She shares how Emburse empowers marketers to experiment safely, celebrate early wins, and iterate fast, creating an environment where AI isn’t intimidating. It’s exciting! From modernizing content workflows to training AI SDR agents like Kira, Heidi explains the mindset shifts leaders need to make if they want AI to actually work inside their teams.
What you’ll learn:
TRANSCRIPT
Sarah McConnell – Qualified
Heidi, thank you so much for joining me today on The Agentic Marketer. I'm so excited to have you. Before we jump into anything, I'd love for you to tell me a little bit about yourself and the work that you're doing over at Emburse.
Heidi Wurpel – Emburse
Sure. So my job is the Vice President of Emburse and Marketing. And I basically take what is currently one of the world's largest travel and expense management companies to even greater heights by spreading our expense intelligence story powered by our incredibly innovative AI. And I've been in marketing for about 20 years now, starting in brand and corporate comms, luxury B2C during the dot-com boom and bust. So that was fun. And then I acquired my MBA, worked in market research and data, spent a little over a decade in tech and ops, and now, of course, I'm focused on leadership, including demand gen and go-to-market for Emburse.
Sarah McConnell – Qualified
Hahaha! Amazing. I love what you guys are doing over at Emburse. I've used your product before. I never realized until I got into tech how cumbersome travel expenses could be. So I love what you guys are doing over at Emburse. I've used the product. It is definitely a huge time save.
Heidi Wurpel – Emburse
We love it, and we're doing even more. It's coming down the AI roadmap to make it easier on you day in and day out, because travel can be fun, and we want it to be.
Sarah McConnell – Qualified
Amazing. Okay, so obviously, the title of the podcast, The Agentic Marketer. So Heidi, it's a new term. People aren't quite as familiar with it. How do you define agentic marketing, and what do you think it means to be an agentic marketer right now?
Heidi Wurpel – Emburse
Yeah. So I actually think agentic really is just being—like, it's a data-based way of saying the right thing to the right person at the right time. Think of it as a personalization mass multiplier. And I think that agentic marketers in particular are really fixated on getting prospective customers the answers they need as quickly as possible, as easily as possible, and truly tailored to their unique problems. And the data sets are what allow us to do that today in technology.
Sarah McConnell – Qualified
I really like that answer. I feel like it's such a—people, you know, agents, AI, it's a very loaded term, but I really like the way you phrase that, if it's more about getting people the exact right thing at the exact right time. And now with AI, we can do that so much better than we were able to before. Like, we've been talking about personalization and marketing for a really long time. That's not new. It just feels like we've entered a very new time where that personalization is next level. It's just so much different than it was.
Heidi Wurpel – Emburse
Yeah, I think that today data and technology are enabling us to do what—like, if I get on a little bit of a high horse here—like people in proximity. Again, I was in… my gray hair comes naturally. Like, I was here during the dot-com boost, right? And then a little bit of a burst. And when you think about it, in the past, before e-comm went wild, before things went huge, you generally interacted with the same people day in, day out. And they were local people, right? So they knew your preferences. They knew maybe your family's preferences, et cetera, et cetera, et cetera.
And I think we have this deep want to return to that. But we've gone so far away from it these past 30-plus years to sometimes really good things like price transparency and globalization. There's definitely positives about what we've done these past 30 years in marketing. But I just think today our databases, our data structures, and most importantly really the agents that allow us to interact with that data and those databases and structures are really what differs from traditional marketing. But it's almost a return to like traditional-traditional marketing, like the pre-e-comm days of the world, in my mind.
So I'm pretty excited to be able to connect people with the information that they need faster and more easily and let them feel seen and known and like we're adapting to their unique needs in that way.
Sarah McConnell – Qualified
Yeah, I like this return to the traditional-traditional marketing. And I always think about—over the time in my career—everything in marketing feels very cyclical. Like, you know, there's always this dead, like, “This isn't used anymore,” and then it always comes back. And the example I always use is outdoor billboards. I remember no one was paying for billboards. If you asked, “Can I go spend this money on a billboard?” they would laugh at you. Absolutely not. And all of a sudden, if you're in the Bay Area, they're everywhere. Everything does feel very cyclical.
The traditional marketing—it's just evolving, but those basics are still there, and we do need to return to those. So I really like the way that you phrase that, because it is almost like we're just going back to it.
Heidi Wurpel – Emburse
Yeah, it's just a new tool, right? Like, I think there used to be a magnet up in the marketing department at the last organization I worked at. It's always saying—this is during the app bubble, really—“An app is not a strategy; it is a tool.” And I think of AI and agentic AI in that same way. Agentic AI is not a strategy in and of itself. It is a tool that enables us to do what we want to do faster, better, more easily, and maybe getting our end users and prospects what they need quickly in that space. Yeah.
Sarah McConnell – Qualified
Yeah. And to your point of it being a tool, I'm curious—with this new tool that's agentic AI—it’s really easy for people to experiment. I think just given things like ChatGPT and your ability to build, and Gemini, and Perplexity, and all of these things, people are experimenting more heavily with this technology than we ever have with previous technology just because of the availability of it. But if you want to move from that experimentation to actually leading with it and using agentic AI as a tool across your entire organization, what do you think it takes to make that shift?
Heidi Wurpel – Emburse
Yeah, I think having extremely clear goals and strategies is important. You need to know what you want your outcomes to be and be very focused on that. And I think for many of us as marketers, that means having excellent original content repositories, which include concise, fulsome, accurate product marketing assets and sales scripts—things that you want to become external-facing, but they need to be there. They need to be concise. They need to be right.
I think you also need clean data and integration capabilities to fully scale this thing. But I think more than anything, what we've found in our organization is you also have to have a culture that's really willing to rapidly pivot and use. And I don't know if this is—again, I spent over a decade in tech, so I don't know that all marketers are comfortable with CI/CD—but continuous improvement, continuous development principles are that idea that instead of changing the whole app user interface, make the button blue with a little more rounding. It's not a whole rebrand; it's a little refresh or a little evolution on what you've been doing.
And I think if you have a team who culturally is comfortable with that—these fast pivots in small ways where maybe it's not going to come to great fanfare, maybe you're going to get pushback on “That's wrong, that's not good, people don't like that”—okay. What's cool about agentic AI is great: you can train the model rapidly and easily to adapt. It's that adaptability again—continuous improvement, continuous development—constantly checking in and making those changes to train the models and make them better and better and better over time.
And realizing—you know, like you and I were talking about billboards—right? You need to get a billboard right. You print a billboard; it's the billboard. It was expensive to paste. And these days with digital billboards, you have a little more leeway, but non-digital billboards—what's on the billboard is what is on the billboard. If you have a typo or an error, that's on you.
That is not agentic AI. It is a living, breathing, changing asset day in and day out. And so I think that culture really matters. But again, have those really clear strategies, those really clear goals, and then start thinking about the backend of what's going to feed that, and then be willing to feed it rapidly and pay attention to it like you would, I don't know, any employee or any piece of information that you're putting out.
Sarah McConnell – Qualified
And we tell people that all the time. So I really like that you hit on the point of it being continual, and you have to go back and be willing to rapidly train and test. I think in the age of AI, things are moving fast—your company, how you're putting out information. And we obviously have our own AI SDR; her name is Piper. But when we talk about onboarding her as an agent, I always caution people: this isn't a set-it-and-forget-it. You can't just give her information—or whatever you end up personifying your agent as—you can't just give it information and expect it to keep up with your rapidly changing organization.
Every time something on the marketing team changes—your product marketing, your messaging shifts, you release new assets, you have a new event—it does have to become a part of your workflow if you want to lead with AI and not just experiment with it. You have to give that new information to any of the models, and that involves any of your agents or AI that you're using internally, to utilize that information and then make sure it's using it properly and give it that feedback.
And I think, to your point, Heidi, that is where people kind of fall off on going from experimentation to leading, because they tend to try it, they forget about it, then all of a sudden it's not giving quite the right answers or it's not doing it exactly the way they wanted. But it does involve pretty good oversight from the people on your team, I think.
Heidi Wurpel – Emburse
Yeah. You know, I recently read an article, and I'm not going to get his name right. I think it's Joe Carveo. He's a marketer-operator for a PE firm. And Joe wrote this thing about how, on his hierarchy, he actually slates agentic agents in as if they're people on his hierarchical org chart, so that it's really clear who the manager of that agent—or that person, what would traditionally be known as a person—is.
I don't know. That kind of blew my mind the other night. It's not my hot take; it's more of his hot take. But I think that was a really interesting take on AI and how we train it. And I think that's part of what made us go to Qualified. And we named ours Kira—you named yours Piper. So part of what we really liked was this idea that you could give her—you know, personifying here again—but you could give her information, and the moment that we update a blog, the moment that we update our landing pages, the moment that we update our data sheets, she’s instantly learning all of that information. She's ingesting that and finding ways to share it out.
Versus the old way of things where we had to go, “All right, there’s a playbook that relies upon that information. Go and remember to update that playbook.” And you know, like you and I talked about earlier, Emburse is an incredibly innovative company. We are rapidly releasing features and products and bringing new things to market. So to keep up all of those data sheets, to remember every playbook that's ever been written and redo your script—the nuance of that? Huge time sink. Huge.
The person who was managing it was just like, “My gosh, this is going to be in 30 places. This is going to be in 60 places. This is going to be in 300 playbooks.” That's hard. That's really hard. And it's not the modern way of doing things. And so I think that that's part of what was really great about Kira—or Piper—and Qualified: we could rapidly train her with the information that we're already bringing out to market.
Also, that would be good when you launch a campaign. You want certain information to hit the market at certain times. So it kind of takes care of any worries like that too—that her information is going to be as up to date as your marketing releases and very aligned with your campaign strategy. So those were some unique things that I thought really are just part of an agent that differ slightly from a human being doing enablement and making sure they don't say anything before X, Y, Z date or whatnot.
Sarah McConnell – Qualified
Yep, that is so true. Now, you mentioned you renamed your AI SDR Kira, which is great. I'm curious: what was it about an AI SDR that had you and the Emburse team leaning into that? I know when AI SDRs hit the market, there was a lot of confusion around them. So what was it that made you and the team say, “We think this is something that will be important to incorporate into our demand gen motion and our larger marketing organization”?
Heidi Wurpel – Emburse
Yeah, I think a couple things. One, my company, Emburse, is leading with AI. We are an AI-based company. So our culture embraces AI. It believes in AI. It's bringing really cutting-edge AI to market at times. So the culture was there. The trust in AI was there. And that's really important.
I think then, from a pure marketer standpoint, saying we need to be able to do just as much quicker. How do we make sure that we train with accurate information quickly and not have to rewrite all those if/then statements that were in the old playbooks in the way that we used to do things?
Having Kira to talk about our innovative AI roadmap, as well as trusting her AI to do what she does best—which is share, again, that right information with the person who's asking at the right time on a regular basis when it's needed.
Sarah McConnell – Qualified
Yeah, that's amazing. And now, is there anything unexpected that bringing on Kira changed for your team? Was there something you weren't expecting that you've now noticed having her as your AI SDR?
Heidi Wurpel – Emburse
Yeah, you know, just the other day I was reading through the questions being asked of Kira, and I noticed it's great to get content guidance. And I think I knew this, but it wasn't what made us make the decision. But it's really nice—it shines a light for me on those areas that our marketing is not clearly answering. And that's why earlier I was like, “Be concise, be accurate.” Because if you see the market asking the same question over and over, the likelihood is what you have out there right now is not enough. It's not answering this question.
So I think I've been pleasantly surprised—or maybe pleasantly reminded—of the fact that you can see trends in the data, and that that can even help inform your content strategy and even bigger strategies if you want to explode it up a little bit more for future years in your marketing.
Sarah McConnell – Qualified
Yeah, that's one of the unexpected things we experienced onboarding Piper at our company. I got to oversee that, which was amazing. And I remember looking at the questions being asked. And I tell customers all the time: you're going to have your nine to ten questions that get asked pretty consistently on your website—that is universal. But there are going to be aggregates of questions that you don't think about that really will help inform your content team.
And that actually kind of shifted our org structure here at Qualified to have content and communications roll up under my team on demand gen, kind of for that reason. I was getting this feedback from the questions that were being asked. I was able to easily see it in Qualified. And I was like, I need to be able to move fast and make sure this content is now out there because I'm obviously seeing in real time we aren't answering that question well enough. We don't have the resources out there in a way that's making it easy for visitors to find us.
Now granted, we can train Piper on it, and that's fine. But we want to make sure across all of our different channels that we have the right information. So it was something that actually impacted our team pretty heavily that I was not expecting.
Heidi Wurpel – Emburse
Yeah. And I think a few years back, I was at an SEO conference, and it was really interesting. It was talking about looking at Google’s other questions that are asked—“People also ask.” And I know that's kind of a thing of the past now. It's way at the bottom. But at the time, that was at the top. It was sort of that “aha” moment where you're like, right, you can use AI to answer all these questions and make sure you get similar content and questions answered on your site because maybe you're thinking about this—expand it a little to that.
And I think that's what this agentic AI is really doing for you, but it's doing it in a hyper-personalized way. It's yours. It’s not like Google used to be—like, the general questions out there in the market are… Well, this is hyper-personalized to: “Hey, Emburse marketing team, ten people asked this thing.” Clearly you need some more information about that on your website and across your channels. So yeah, that was a really pleasant surprise.
Sarah McConnell – Qualified
Now, I'm curious—what metrics are you measuring to know that Kira is working, beyond obviously the content that's harder to measure? Is there anything else that you're looking at from a metrics standpoint pretty consistently?
Heidi Wurpel – Emburse
Yeah, you know, I mean, for us, I'm really proud of our ability to see attribution. And we linked Kira into our Salesforce. And so we're able to see specific MQAs that come through—our marketing qualified accounts—and be able to connect them to: this was a Kira conversation, or this was a Kira-qualified ask for a meeting booked. So we look at MQAs; we look at meetings booked.
And we've seen about a 212% improvement in the past six months alone. So really good numbers there. And then, of course, I'm watching her accuracy of responses, as is my Head of SDRs. He's really fabulous because we look at her as an SDR in that meetings-booked and sort of initial discovery, “disco” type of phase as well. We treat her a bit like an SDR meets a marketing lead source, and we measure her on both of those metrics.
Sarah McConnell – Qualified
That's amazing. And now I want to talk a little bit more. It's very obvious from this conversation that Emburse is—you mentioned—very AI-first, not only as an organization and how you're adopting technology, but also as a company in general. So I feel like you are more than qualified to answer a question for someone if they're looking to adopt AI.
So say someone's listening to this and they are not where I feel like you or Emburse are at, and they're like, “I don't know where to start. This feels very overwhelming.” What would be your advice, having gone through this journey already, that you would tell other marketing leaders? Where should they start when they're looking to adopt AI or become more adept at agentic marketing?
Heidi Wurpel – Emburse
Yeah, I think AI adoption really does come culturally. And if you're a leader in your organization, I think you can encourage your people to play, right? What's nice about being a marketer is most of the information that we share is information that we want the world to see. So I think a lot of the time people are really concerned. So you can put guardrails up about: don't share anything company-confidential. But if you want the world to see the information—share the information you want the world to see.
Go ahead and play a little bit. Sometimes you're going to flop and fail. Other times you're going to have wild success. And I think for me, one of the first things that I did was—in my digital innovation team—they really started playing with it, and we started highlighting those areas of success. We were highlighting their successes to the rest of marketers, and that gave sort of permission to play. Like, “Hey, so-and-so did this. It's not 100%. But look how great it is. This is what it allowed us to do.”
And so we took a few problem statements that we'd been putting on the back burner for a while because we thought, gosh, finding a solution to that is so big and so hairy. And then—we’re a Google shop—so we were like, “Wait, we have NotebookLM.” Let's just play with it and see what it can do. And we were like, “Wait, that really worked.” And we could then expand it here and there. What other uses could we do?
And so I literally have a monthly show-and-tell where my team brings really cool new things they've shared. I learned about Nano Banana from them before I saw any press releases on it. So it's just really neat to see the different qualities of your team. So I think allow people to play within these safe data sets and within this safe information—nothing that's going to hurt or not be able to be shared out there.
So encourage the culture and, like I said, celebrate those successes.
Sarah McConnell – Qualified
Yeah, absolutely. I really like—we’ve heard that answer quite a bit on this podcast, and I just cannot emphasize it enough. If you want to lead and be at the forefront of this, you have to not only allow your team to use AI—which, it's wild to me, but there are still companies out there that aren't letting their employees do that. But I like what you pointed out: champion them. Champion the people that are doing it well. Give them spaces to come to, to show off what they're learning and what they're doing.
I think that's such a fast way to get the team excited about something and willing to get their hands in it and start using it themselves. Being able to champion those employees, I think, is so, so important right now.
Heidi Wurpel – Emburse
Yeah, 100%. You know, I really think sometimes as leaders—we all try to be approachable, right? We want to be really approachable to our team. But on the flip side, how often do you want to go to your boss and be like, “I don't really know about this thing”? Like, teach me about this. I don't know—maybe there are people who do. But a lot of us—the A-plus students—we want to come forth always with the A-plus things.
And so I think what this allows is you create these little knowledge experts. And you don't have to spin them as such, but these cohorts that can come together and privately say, “Hey, it’s really cool what you did there. Can you show me that? Can you teach me that?” And instead of having to go up the chain, it creates those horizontal connections and networks that allow people to ask each other the silly questions and really have respect.
Because if you're someone who has literally never typed a prompt into anything—and I think there are people absolutely still out there in the world, and no shade on that—that first prompt can feel a little scary. And literally just having another colleague there next to you to type the prompt, and have them watch and ask, “And now what do I do? And what button do I push?” It feels really nice—just to give that safety and that mechanism of: you can't screw this up. It's really okay. Go ahead and try something simple.
So yeah, that's what we've created in our marketing culture. And I'm really proud of the team, because holy cow, they bring things that I would never think to bring forth, and solutions to each other. But then just iterate on one another and have saved a great deal of time and energy and wasted cycles.
Sarah McConnell – Qualified
Now to that point—are there any particular skills you've seen within the team that you think other companies really need to double down on in order to thrive in this AI-driven era? Are there particular skills that you've seen within your team or yourself that will really help teams make that shift into an AI-first company?
Heidi Wurpel – Emburse
I think skill-wise: clarity, again. Being able to really know what you want to get out of something and really clearly articulate that the outcome should be like this, or that the output is for so-and-so.
I think having creativity is really important—being willing to ask the same question in four different ways and think in that way is really important. But I really just think it's that willingness to trust that the models will present at least something that you can start with and that you don't have to end with.
I love a blank canvas, but there are days when we are moving so fast that you don't have time for a blank canvas. It's really helpful to have—and I'm super geeky in the art space—for those of you who haven't painted, you have a canvas. You paint this white or pink or whatever color layer you want on it to fill in all the little bubbles of the canvas. Yes, that's meditative and great if you're a professional artist, but if you're not—if you're just at a sip-and-paint—that’s not particularly fulfilling. So you let the AI do that.
There's a certain horizontal line you paint often on a canvas if you're doing a landscape or something, or even just to give yourself a little guidance. Similarly, let AI do that. It's not all or nothing with this. It's not “Paint me a Picasso,” and then “That doesn't look like Picasso, I hate it.” Just let it whitewash the canvas, and you paint the rest of the image on it.
So I think that's it. It's not all or nothing. It's not black and white. Always, always edit whatever your AI is putting forth or retrain it in the agentic world. And so I think it all comes down to that willingness to iterate and have clarity of your final vision—you know what you want on that canvas at the end.
Sarah McConnell – Qualified
I think that's the best analogy I've heard for how to let AI support you versus replace you. I really love that. I might have to use that in the future because it's very clear, but also very beautiful.
Okay, Heidi, I have a couple of questions now looking forward. I know it's always hard when we're like, “Get out your crystal ball and tell the future,” especially in an AI era when things are moving so fast. But I'm curious—if you are thinking ahead 12 months, what do you think, besides AI, is going to be the single biggest change to B2B marketing? Like, what is AI doing right now that's going to make the biggest shift in the next 12 months for us B2B marketers?
Heidi Wurpel – Emburse
I actually think it will be more than just about marketing. So I think it's going to have a big impact on go-to-market. What I mean by that is: I think that we see that AI is an accelerator. It's a pace accelerator; it is a volume accelerator. And I think if we do this really right, we're going to see the sales velocity also decrease. Right?
When you think about the velocity of a sale—and this is more B2B than B2C—B2C, I see this shirt, I want to buy this shirt, I buy it off the rack. Maybe it takes a little while to ship if I'm doing e-comm. But when you think about B2B marketing, it takes time often in non-product-led growth models to go through the various phases of the sales cycle. And we know people do a ton of their research ahead of time and do it within marketing.
So done right, I think we're going to see the velocity of sales motions shorten. And I think we may even see disruptions to your traditional go-to-market sales plays, with an increase in product-led growth—even for those large global enterprise companies and expensive software.
So I think it's going to impact go-to-market if marketers do this correctly and create velocity acceleration on some deals.
Sarah McConnell – Qualified
Now on the flip side—if you're again thinking 12 months out—in the next year, is there something you worry marketers are still going to be getting wrong about AI a year from now? What do you think we're still heading down the wrong path on that we might still be doing incorrectly?
Heidi Wurpel – Emburse
Yeah. I think this concept of replacing entire roles or sets is still a little off. Do you need to hire as many of the same person? Maybe not. But that doesn't mean you don't need a product marketer anymore, or you don't need a content marketer anymore, or you don't need creatives anymore—just let the AI do it.
There’s that idea of keeping to your brand and keeping to your true north—that’s really important. So I think we could move too fast and we could get that wrong, for sure.
I'm hesitant when I hear people… you know, I love putting an AI head on the hierarchy, like I said. And I think that there's a lot we can do. We can increase pace. But I think that is something that could go awry if we let it go too far in efficiency, efficiency, optimization, optimization. Don't lose those skill sets. You need somebody to babysit the model.
I don't know about you, but I'm probably not going to hang an AI Picasso on my wall. I mean, I'm not going to hang a real Picasso because I don’t have money like that, but you know what I mean. If I'm hanging a contemporary, interesting portrait on the wall, I'm probably going to choose one that's out of paint from the hands of a true artist who put those final touches on it.
Sarah McConnell – Qualified
I cannot agree more. I do like the idea of incorporating AI into your org chart only from the standpoint of: it helps to know who's in charge. Who is the person responsible for making sure the AI and the model are running the way you want them to? But I do agree—I think it's always funny to watch things right-size. People go really hard, and I get it. As things change—as marketers—we tend to be on the forefront of wanting to test things and be on the cutting edge.
My CEO always says: if you're not a little uncomfortable, you're probably not pushing yourself hard enough, and that's good marketing. You really have to be kind of out in front of your skis a little bit. But in that process, we tend to right-size as things go on. And I think there was a shift of: could it replace entire teams? Could it replace entire roles?
And it's like—well—we’ve seen internally with SDRs: yes, we have an AI SDR, and yes, she can scale infinitely and do all of these things. But we’ve still found that we have changing roles with people that are now managing the AI SDRs. We have our inbound and our outbound AI managers. But also, we still have SDRs who are doing outbounding and writing really good personalized emails for our ABM, very hyper-targeted accounts.
So I do think having them work side by side is right now the key. That's where we're going to see the most benefit.
Heidi Wurpel – Emburse
I think so too. Again—it’s a tool; it's not a strategy. Your strategy needs to be clear and concise. You need to know what you're trying to get your outcomes to be. And part of your unique selling proposition is telling people clearly what your unique selling proposition is. Don't let the AI write that for you, because it's going to write a similar thing for all companies. It's a mass aggregator.
So you need to stay unique. You need to have those qualifiers. And I think that's the human touch and that set of control.
I also think something that—along with trying to carte blanche replace whole roles—we could get wrong is that we are so used to being able to control every word, every sentence, every message that comes out. And you have to be able to go with the flow a little bit when you move into an AI world.
If you have very specific and controlled outcomes that you want to occur—like “The company desires this sentence to always be written this exact way”—that's not going to happen anymore. What you need to change your focus to is ruthlessly focusing on your customer's needs and meeting those.
Does it get down to the heart of answering their question correctly, in an accurate way? Try to avoid hallucinations, but we know AI models today hallucinate. If you're someone who's very rigid, there are industries where you have to be very careful with every single sentence.
And what's cool about AI is you can say, “Anytime you use a sentence like this, write it this exact way.” You can train your AI SDR to do that. But I just think you need a little flexibility—a little, to your point, skiing analogy. If it's an icy day or a powder day, you're going to ski it differently. And you need to trust that AI is going to ski that terrain differently too. It's going to know what to apply, but the outcome is going to be different each time.
Sarah McConnell – Qualified
I totally agree. Okay, to wrap things up, Heidi, we have our lightning round. A couple of quick questions to wrap things up that are a little more fun.
The first one: besides ChatGPT—because that tends to be most people's answer—what was the first AI tool that you experimented with?
Heidi Wurpel – Emburse
PhotoLeap AI Headshot Generator. Yeah. My hands were weird, my ears were weird. I'm not going to lie. But a few came out okay. But it was actually an AI headshot generator. In the end, I did go and get a professional photographer—a real human. So here’s—right—I explored and said, nope, I think a human’s going to be better for this. But it was actually a headshot AI generator that I played with first and foremost.
Sarah McConnell – Qualified
I like that one. Okay, what do you think is the most overrated buzzword in martech right now?
Heidi Wurpel – Emburse
Ooh. The metaverse—that we all want to be digital AI avatars always plugged in. I think that's overblown. I don't think that's going to take over the world in the next 12 months anyway.
Sarah McConnell – Qualified
I'm with you on that one. You mentioned someone earlier, but is there one marketer who you think is really ahead of the curve on AI right now that you like to follow and you'd recommend others follow on LinkedIn or other channels?
Heidi Wurpel – Emburse
Yeah. So Erin Mills, who I have worked with before, but she's doing a really great other podcast. So if you're into podcasts, check out Erin Mills' podcast as well on LinkedIn. She talks about AI on a really regular basis and brings on some interesting guests. It's always fun to see her and what she's doing in this space.
Sarah McConnell – Qualified
Amazing. And last question—if you could automate one part of your life outside of work with AI, what would it be?
Heidi Wurpel – Emburse
Okay, I'm staring at a Monstera that needs a little bit of love. Probably my houseplant care. I have a green-enough thumb. It's green, but it's not deep verdant green here—they're struggling. If I had something that pruned them just the right way at the right time and moved them into the sun and gave them the right amount of water and food, I would absolutely invest in that. Because sometimes when you travel for work—and things which I hope everybody gets the opportunity to do because it's not that bad to do your reimbursements; use Emburse, it is fun—but when you travel, your plants do that death shrink. And that would be really nice.
Sarah McConnell – Qualified
Yeah, I would also invest in that. I'm thinking about the multitude of houseplants I have in my house that are doing okay—just good enough—but not great.
Heidi Wurpel – Emburse
That's… yeah. I'm sorry, Monstera.
Sarah McConnell – Qualified
Okay, Heidi, well, thank you so much for joining us on The Agentic Marketer. It was so great to have you today.
Heidi Wurpel – Emburse
Thanks for having me. What a pleasure.
Stay up to date with weekly drops of fresh B2B marketing and sales content.
Discover how Emburse drives real AI adoption by creating a culture of experimentation. Heidi Wurpel shares why giving teams “permission to play” accelerates agentic marketing.
In this episode, Sarah McConnell sits down with Heidi Wurpel, VP of Marketing at Emburse, to explore why the true foundation of AI adoption isn’t tools or technology. It’s culture. Heidi breaks down how giving teams permission to play with AI accelerates learning, sparks innovation, and builds the confidence needed to scale agentic marketing across an organization.
She shares how Emburse empowers marketers to experiment safely, celebrate early wins, and iterate fast, creating an environment where AI isn’t intimidating. It’s exciting! From modernizing content workflows to training AI SDR agents like Kira, Heidi explains the mindset shifts leaders need to make if they want AI to actually work inside their teams.
What you’ll learn:
TRANSCRIPT
Sarah McConnell – Qualified
Heidi, thank you so much for joining me today on The Agentic Marketer. I'm so excited to have you. Before we jump into anything, I'd love for you to tell me a little bit about yourself and the work that you're doing over at Emburse.
Heidi Wurpel – Emburse
Sure. So my job is the Vice President of Emburse and Marketing. And I basically take what is currently one of the world's largest travel and expense management companies to even greater heights by spreading our expense intelligence story powered by our incredibly innovative AI. And I've been in marketing for about 20 years now, starting in brand and corporate comms, luxury B2C during the dot-com boom and bust. So that was fun. And then I acquired my MBA, worked in market research and data, spent a little over a decade in tech and ops, and now, of course, I'm focused on leadership, including demand gen and go-to-market for Emburse.
Sarah McConnell – Qualified
Hahaha! Amazing. I love what you guys are doing over at Emburse. I've used your product before. I never realized until I got into tech how cumbersome travel expenses could be. So I love what you guys are doing over at Emburse. I've used the product. It is definitely a huge time save.
Heidi Wurpel – Emburse
We love it, and we're doing even more. It's coming down the AI roadmap to make it easier on you day in and day out, because travel can be fun, and we want it to be.
Sarah McConnell – Qualified
Amazing. Okay, so obviously, the title of the podcast, The Agentic Marketer. So Heidi, it's a new term. People aren't quite as familiar with it. How do you define agentic marketing, and what do you think it means to be an agentic marketer right now?
Heidi Wurpel – Emburse
Yeah. So I actually think agentic really is just being—like, it's a data-based way of saying the right thing to the right person at the right time. Think of it as a personalization mass multiplier. And I think that agentic marketers in particular are really fixated on getting prospective customers the answers they need as quickly as possible, as easily as possible, and truly tailored to their unique problems. And the data sets are what allow us to do that today in technology.
Sarah McConnell – Qualified
I really like that answer. I feel like it's such a—people, you know, agents, AI, it's a very loaded term, but I really like the way you phrase that, if it's more about getting people the exact right thing at the exact right time. And now with AI, we can do that so much better than we were able to before. Like, we've been talking about personalization and marketing for a really long time. That's not new. It just feels like we've entered a very new time where that personalization is next level. It's just so much different than it was.
Heidi Wurpel – Emburse
Yeah, I think that today data and technology are enabling us to do what—like, if I get on a little bit of a high horse here—like people in proximity. Again, I was in… my gray hair comes naturally. Like, I was here during the dot-com boost, right? And then a little bit of a burst. And when you think about it, in the past, before e-comm went wild, before things went huge, you generally interacted with the same people day in, day out. And they were local people, right? So they knew your preferences. They knew maybe your family's preferences, et cetera, et cetera, et cetera.
And I think we have this deep want to return to that. But we've gone so far away from it these past 30-plus years to sometimes really good things like price transparency and globalization. There's definitely positives about what we've done these past 30 years in marketing. But I just think today our databases, our data structures, and most importantly really the agents that allow us to interact with that data and those databases and structures are really what differs from traditional marketing. But it's almost a return to like traditional-traditional marketing, like the pre-e-comm days of the world, in my mind.
So I'm pretty excited to be able to connect people with the information that they need faster and more easily and let them feel seen and known and like we're adapting to their unique needs in that way.
Sarah McConnell – Qualified
Yeah, I like this return to the traditional-traditional marketing. And I always think about—over the time in my career—everything in marketing feels very cyclical. Like, you know, there's always this dead, like, “This isn't used anymore,” and then it always comes back. And the example I always use is outdoor billboards. I remember no one was paying for billboards. If you asked, “Can I go spend this money on a billboard?” they would laugh at you. Absolutely not. And all of a sudden, if you're in the Bay Area, they're everywhere. Everything does feel very cyclical.
The traditional marketing—it's just evolving, but those basics are still there, and we do need to return to those. So I really like the way that you phrase that, because it is almost like we're just going back to it.
Heidi Wurpel – Emburse
Yeah, it's just a new tool, right? Like, I think there used to be a magnet up in the marketing department at the last organization I worked at. It's always saying—this is during the app bubble, really—“An app is not a strategy; it is a tool.” And I think of AI and agentic AI in that same way. Agentic AI is not a strategy in and of itself. It is a tool that enables us to do what we want to do faster, better, more easily, and maybe getting our end users and prospects what they need quickly in that space. Yeah.
Sarah McConnell – Qualified
Yeah. And to your point of it being a tool, I'm curious—with this new tool that's agentic AI—it’s really easy for people to experiment. I think just given things like ChatGPT and your ability to build, and Gemini, and Perplexity, and all of these things, people are experimenting more heavily with this technology than we ever have with previous technology just because of the availability of it. But if you want to move from that experimentation to actually leading with it and using agentic AI as a tool across your entire organization, what do you think it takes to make that shift?
Heidi Wurpel – Emburse
Yeah, I think having extremely clear goals and strategies is important. You need to know what you want your outcomes to be and be very focused on that. And I think for many of us as marketers, that means having excellent original content repositories, which include concise, fulsome, accurate product marketing assets and sales scripts—things that you want to become external-facing, but they need to be there. They need to be concise. They need to be right.
I think you also need clean data and integration capabilities to fully scale this thing. But I think more than anything, what we've found in our organization is you also have to have a culture that's really willing to rapidly pivot and use. And I don't know if this is—again, I spent over a decade in tech, so I don't know that all marketers are comfortable with CI/CD—but continuous improvement, continuous development principles are that idea that instead of changing the whole app user interface, make the button blue with a little more rounding. It's not a whole rebrand; it's a little refresh or a little evolution on what you've been doing.
And I think if you have a team who culturally is comfortable with that—these fast pivots in small ways where maybe it's not going to come to great fanfare, maybe you're going to get pushback on “That's wrong, that's not good, people don't like that”—okay. What's cool about agentic AI is great: you can train the model rapidly and easily to adapt. It's that adaptability again—continuous improvement, continuous development—constantly checking in and making those changes to train the models and make them better and better and better over time.
And realizing—you know, like you and I were talking about billboards—right? You need to get a billboard right. You print a billboard; it's the billboard. It was expensive to paste. And these days with digital billboards, you have a little more leeway, but non-digital billboards—what's on the billboard is what is on the billboard. If you have a typo or an error, that's on you.
That is not agentic AI. It is a living, breathing, changing asset day in and day out. And so I think that culture really matters. But again, have those really clear strategies, those really clear goals, and then start thinking about the backend of what's going to feed that, and then be willing to feed it rapidly and pay attention to it like you would, I don't know, any employee or any piece of information that you're putting out.
Sarah McConnell – Qualified
And we tell people that all the time. So I really like that you hit on the point of it being continual, and you have to go back and be willing to rapidly train and test. I think in the age of AI, things are moving fast—your company, how you're putting out information. And we obviously have our own AI SDR; her name is Piper. But when we talk about onboarding her as an agent, I always caution people: this isn't a set-it-and-forget-it. You can't just give her information—or whatever you end up personifying your agent as—you can't just give it information and expect it to keep up with your rapidly changing organization.
Every time something on the marketing team changes—your product marketing, your messaging shifts, you release new assets, you have a new event—it does have to become a part of your workflow if you want to lead with AI and not just experiment with it. You have to give that new information to any of the models, and that involves any of your agents or AI that you're using internally, to utilize that information and then make sure it's using it properly and give it that feedback.
And I think, to your point, Heidi, that is where people kind of fall off on going from experimentation to leading, because they tend to try it, they forget about it, then all of a sudden it's not giving quite the right answers or it's not doing it exactly the way they wanted. But it does involve pretty good oversight from the people on your team, I think.
Heidi Wurpel – Emburse
Yeah. You know, I recently read an article, and I'm not going to get his name right. I think it's Joe Carveo. He's a marketer-operator for a PE firm. And Joe wrote this thing about how, on his hierarchy, he actually slates agentic agents in as if they're people on his hierarchical org chart, so that it's really clear who the manager of that agent—or that person, what would traditionally be known as a person—is.
I don't know. That kind of blew my mind the other night. It's not my hot take; it's more of his hot take. But I think that was a really interesting take on AI and how we train it. And I think that's part of what made us go to Qualified. And we named ours Kira—you named yours Piper. So part of what we really liked was this idea that you could give her—you know, personifying here again—but you could give her information, and the moment that we update a blog, the moment that we update our landing pages, the moment that we update our data sheets, she’s instantly learning all of that information. She's ingesting that and finding ways to share it out.
Versus the old way of things where we had to go, “All right, there’s a playbook that relies upon that information. Go and remember to update that playbook.” And you know, like you and I talked about earlier, Emburse is an incredibly innovative company. We are rapidly releasing features and products and bringing new things to market. So to keep up all of those data sheets, to remember every playbook that's ever been written and redo your script—the nuance of that? Huge time sink. Huge.
The person who was managing it was just like, “My gosh, this is going to be in 30 places. This is going to be in 60 places. This is going to be in 300 playbooks.” That's hard. That's really hard. And it's not the modern way of doing things. And so I think that that's part of what was really great about Kira—or Piper—and Qualified: we could rapidly train her with the information that we're already bringing out to market.
Also, that would be good when you launch a campaign. You want certain information to hit the market at certain times. So it kind of takes care of any worries like that too—that her information is going to be as up to date as your marketing releases and very aligned with your campaign strategy. So those were some unique things that I thought really are just part of an agent that differ slightly from a human being doing enablement and making sure they don't say anything before X, Y, Z date or whatnot.
Sarah McConnell – Qualified
Yep, that is so true. Now, you mentioned you renamed your AI SDR Kira, which is great. I'm curious: what was it about an AI SDR that had you and the Emburse team leaning into that? I know when AI SDRs hit the market, there was a lot of confusion around them. So what was it that made you and the team say, “We think this is something that will be important to incorporate into our demand gen motion and our larger marketing organization”?
Heidi Wurpel – Emburse
Yeah, I think a couple things. One, my company, Emburse, is leading with AI. We are an AI-based company. So our culture embraces AI. It believes in AI. It's bringing really cutting-edge AI to market at times. So the culture was there. The trust in AI was there. And that's really important.
I think then, from a pure marketer standpoint, saying we need to be able to do just as much quicker. How do we make sure that we train with accurate information quickly and not have to rewrite all those if/then statements that were in the old playbooks in the way that we used to do things?
Having Kira to talk about our innovative AI roadmap, as well as trusting her AI to do what she does best—which is share, again, that right information with the person who's asking at the right time on a regular basis when it's needed.
Sarah McConnell – Qualified
Yeah, that's amazing. And now, is there anything unexpected that bringing on Kira changed for your team? Was there something you weren't expecting that you've now noticed having her as your AI SDR?
Heidi Wurpel – Emburse
Yeah, you know, just the other day I was reading through the questions being asked of Kira, and I noticed it's great to get content guidance. And I think I knew this, but it wasn't what made us make the decision. But it's really nice—it shines a light for me on those areas that our marketing is not clearly answering. And that's why earlier I was like, “Be concise, be accurate.” Because if you see the market asking the same question over and over, the likelihood is what you have out there right now is not enough. It's not answering this question.
So I think I've been pleasantly surprised—or maybe pleasantly reminded—of the fact that you can see trends in the data, and that that can even help inform your content strategy and even bigger strategies if you want to explode it up a little bit more for future years in your marketing.
Sarah McConnell – Qualified
Yeah, that's one of the unexpected things we experienced onboarding Piper at our company. I got to oversee that, which was amazing. And I remember looking at the questions being asked. And I tell customers all the time: you're going to have your nine to ten questions that get asked pretty consistently on your website—that is universal. But there are going to be aggregates of questions that you don't think about that really will help inform your content team.
And that actually kind of shifted our org structure here at Qualified to have content and communications roll up under my team on demand gen, kind of for that reason. I was getting this feedback from the questions that were being asked. I was able to easily see it in Qualified. And I was like, I need to be able to move fast and make sure this content is now out there because I'm obviously seeing in real time we aren't answering that question well enough. We don't have the resources out there in a way that's making it easy for visitors to find us.
Now granted, we can train Piper on it, and that's fine. But we want to make sure across all of our different channels that we have the right information. So it was something that actually impacted our team pretty heavily that I was not expecting.
Heidi Wurpel – Emburse
Yeah. And I think a few years back, I was at an SEO conference, and it was really interesting. It was talking about looking at Google’s other questions that are asked—“People also ask.” And I know that's kind of a thing of the past now. It's way at the bottom. But at the time, that was at the top. It was sort of that “aha” moment where you're like, right, you can use AI to answer all these questions and make sure you get similar content and questions answered on your site because maybe you're thinking about this—expand it a little to that.
And I think that's what this agentic AI is really doing for you, but it's doing it in a hyper-personalized way. It's yours. It’s not like Google used to be—like, the general questions out there in the market are… Well, this is hyper-personalized to: “Hey, Emburse marketing team, ten people asked this thing.” Clearly you need some more information about that on your website and across your channels. So yeah, that was a really pleasant surprise.
Sarah McConnell – Qualified
Now, I'm curious—what metrics are you measuring to know that Kira is working, beyond obviously the content that's harder to measure? Is there anything else that you're looking at from a metrics standpoint pretty consistently?
Heidi Wurpel – Emburse
Yeah, you know, I mean, for us, I'm really proud of our ability to see attribution. And we linked Kira into our Salesforce. And so we're able to see specific MQAs that come through—our marketing qualified accounts—and be able to connect them to: this was a Kira conversation, or this was a Kira-qualified ask for a meeting booked. So we look at MQAs; we look at meetings booked.
And we've seen about a 212% improvement in the past six months alone. So really good numbers there. And then, of course, I'm watching her accuracy of responses, as is my Head of SDRs. He's really fabulous because we look at her as an SDR in that meetings-booked and sort of initial discovery, “disco” type of phase as well. We treat her a bit like an SDR meets a marketing lead source, and we measure her on both of those metrics.
Sarah McConnell – Qualified
That's amazing. And now I want to talk a little bit more. It's very obvious from this conversation that Emburse is—you mentioned—very AI-first, not only as an organization and how you're adopting technology, but also as a company in general. So I feel like you are more than qualified to answer a question for someone if they're looking to adopt AI.
So say someone's listening to this and they are not where I feel like you or Emburse are at, and they're like, “I don't know where to start. This feels very overwhelming.” What would be your advice, having gone through this journey already, that you would tell other marketing leaders? Where should they start when they're looking to adopt AI or become more adept at agentic marketing?
Heidi Wurpel – Emburse
Yeah, I think AI adoption really does come culturally. And if you're a leader in your organization, I think you can encourage your people to play, right? What's nice about being a marketer is most of the information that we share is information that we want the world to see. So I think a lot of the time people are really concerned. So you can put guardrails up about: don't share anything company-confidential. But if you want the world to see the information—share the information you want the world to see.
Go ahead and play a little bit. Sometimes you're going to flop and fail. Other times you're going to have wild success. And I think for me, one of the first things that I did was—in my digital innovation team—they really started playing with it, and we started highlighting those areas of success. We were highlighting their successes to the rest of marketers, and that gave sort of permission to play. Like, “Hey, so-and-so did this. It's not 100%. But look how great it is. This is what it allowed us to do.”
And so we took a few problem statements that we'd been putting on the back burner for a while because we thought, gosh, finding a solution to that is so big and so hairy. And then—we’re a Google shop—so we were like, “Wait, we have NotebookLM.” Let's just play with it and see what it can do. And we were like, “Wait, that really worked.” And we could then expand it here and there. What other uses could we do?
And so I literally have a monthly show-and-tell where my team brings really cool new things they've shared. I learned about Nano Banana from them before I saw any press releases on it. So it's just really neat to see the different qualities of your team. So I think allow people to play within these safe data sets and within this safe information—nothing that's going to hurt or not be able to be shared out there.
So encourage the culture and, like I said, celebrate those successes.
Sarah McConnell – Qualified
Yeah, absolutely. I really like—we’ve heard that answer quite a bit on this podcast, and I just cannot emphasize it enough. If you want to lead and be at the forefront of this, you have to not only allow your team to use AI—which, it's wild to me, but there are still companies out there that aren't letting their employees do that. But I like what you pointed out: champion them. Champion the people that are doing it well. Give them spaces to come to, to show off what they're learning and what they're doing.
I think that's such a fast way to get the team excited about something and willing to get their hands in it and start using it themselves. Being able to champion those employees, I think, is so, so important right now.
Heidi Wurpel – Emburse
Yeah, 100%. You know, I really think sometimes as leaders—we all try to be approachable, right? We want to be really approachable to our team. But on the flip side, how often do you want to go to your boss and be like, “I don't really know about this thing”? Like, teach me about this. I don't know—maybe there are people who do. But a lot of us—the A-plus students—we want to come forth always with the A-plus things.
And so I think what this allows is you create these little knowledge experts. And you don't have to spin them as such, but these cohorts that can come together and privately say, “Hey, it’s really cool what you did there. Can you show me that? Can you teach me that?” And instead of having to go up the chain, it creates those horizontal connections and networks that allow people to ask each other the silly questions and really have respect.
Because if you're someone who has literally never typed a prompt into anything—and I think there are people absolutely still out there in the world, and no shade on that—that first prompt can feel a little scary. And literally just having another colleague there next to you to type the prompt, and have them watch and ask, “And now what do I do? And what button do I push?” It feels really nice—just to give that safety and that mechanism of: you can't screw this up. It's really okay. Go ahead and try something simple.
So yeah, that's what we've created in our marketing culture. And I'm really proud of the team, because holy cow, they bring things that I would never think to bring forth, and solutions to each other. But then just iterate on one another and have saved a great deal of time and energy and wasted cycles.
Sarah McConnell – Qualified
Now to that point—are there any particular skills you've seen within the team that you think other companies really need to double down on in order to thrive in this AI-driven era? Are there particular skills that you've seen within your team or yourself that will really help teams make that shift into an AI-first company?
Heidi Wurpel – Emburse
I think skill-wise: clarity, again. Being able to really know what you want to get out of something and really clearly articulate that the outcome should be like this, or that the output is for so-and-so.
I think having creativity is really important—being willing to ask the same question in four different ways and think in that way is really important. But I really just think it's that willingness to trust that the models will present at least something that you can start with and that you don't have to end with.
I love a blank canvas, but there are days when we are moving so fast that you don't have time for a blank canvas. It's really helpful to have—and I'm super geeky in the art space—for those of you who haven't painted, you have a canvas. You paint this white or pink or whatever color layer you want on it to fill in all the little bubbles of the canvas. Yes, that's meditative and great if you're a professional artist, but if you're not—if you're just at a sip-and-paint—that’s not particularly fulfilling. So you let the AI do that.
There's a certain horizontal line you paint often on a canvas if you're doing a landscape or something, or even just to give yourself a little guidance. Similarly, let AI do that. It's not all or nothing with this. It's not “Paint me a Picasso,” and then “That doesn't look like Picasso, I hate it.” Just let it whitewash the canvas, and you paint the rest of the image on it.
So I think that's it. It's not all or nothing. It's not black and white. Always, always edit whatever your AI is putting forth or retrain it in the agentic world. And so I think it all comes down to that willingness to iterate and have clarity of your final vision—you know what you want on that canvas at the end.
Sarah McConnell – Qualified
I think that's the best analogy I've heard for how to let AI support you versus replace you. I really love that. I might have to use that in the future because it's very clear, but also very beautiful.
Okay, Heidi, I have a couple of questions now looking forward. I know it's always hard when we're like, “Get out your crystal ball and tell the future,” especially in an AI era when things are moving so fast. But I'm curious—if you are thinking ahead 12 months, what do you think, besides AI, is going to be the single biggest change to B2B marketing? Like, what is AI doing right now that's going to make the biggest shift in the next 12 months for us B2B marketers?
Heidi Wurpel – Emburse
I actually think it will be more than just about marketing. So I think it's going to have a big impact on go-to-market. What I mean by that is: I think that we see that AI is an accelerator. It's a pace accelerator; it is a volume accelerator. And I think if we do this really right, we're going to see the sales velocity also decrease. Right?
When you think about the velocity of a sale—and this is more B2B than B2C—B2C, I see this shirt, I want to buy this shirt, I buy it off the rack. Maybe it takes a little while to ship if I'm doing e-comm. But when you think about B2B marketing, it takes time often in non-product-led growth models to go through the various phases of the sales cycle. And we know people do a ton of their research ahead of time and do it within marketing.
So done right, I think we're going to see the velocity of sales motions shorten. And I think we may even see disruptions to your traditional go-to-market sales plays, with an increase in product-led growth—even for those large global enterprise companies and expensive software.
So I think it's going to impact go-to-market if marketers do this correctly and create velocity acceleration on some deals.
Sarah McConnell – Qualified
Now on the flip side—if you're again thinking 12 months out—in the next year, is there something you worry marketers are still going to be getting wrong about AI a year from now? What do you think we're still heading down the wrong path on that we might still be doing incorrectly?
Heidi Wurpel – Emburse
Yeah. I think this concept of replacing entire roles or sets is still a little off. Do you need to hire as many of the same person? Maybe not. But that doesn't mean you don't need a product marketer anymore, or you don't need a content marketer anymore, or you don't need creatives anymore—just let the AI do it.
There’s that idea of keeping to your brand and keeping to your true north—that’s really important. So I think we could move too fast and we could get that wrong, for sure.
I'm hesitant when I hear people… you know, I love putting an AI head on the hierarchy, like I said. And I think that there's a lot we can do. We can increase pace. But I think that is something that could go awry if we let it go too far in efficiency, efficiency, optimization, optimization. Don't lose those skill sets. You need somebody to babysit the model.
I don't know about you, but I'm probably not going to hang an AI Picasso on my wall. I mean, I'm not going to hang a real Picasso because I don’t have money like that, but you know what I mean. If I'm hanging a contemporary, interesting portrait on the wall, I'm probably going to choose one that's out of paint from the hands of a true artist who put those final touches on it.
Sarah McConnell – Qualified
I cannot agree more. I do like the idea of incorporating AI into your org chart only from the standpoint of: it helps to know who's in charge. Who is the person responsible for making sure the AI and the model are running the way you want them to? But I do agree—I think it's always funny to watch things right-size. People go really hard, and I get it. As things change—as marketers—we tend to be on the forefront of wanting to test things and be on the cutting edge.
My CEO always says: if you're not a little uncomfortable, you're probably not pushing yourself hard enough, and that's good marketing. You really have to be kind of out in front of your skis a little bit. But in that process, we tend to right-size as things go on. And I think there was a shift of: could it replace entire teams? Could it replace entire roles?
And it's like—well—we’ve seen internally with SDRs: yes, we have an AI SDR, and yes, she can scale infinitely and do all of these things. But we’ve still found that we have changing roles with people that are now managing the AI SDRs. We have our inbound and our outbound AI managers. But also, we still have SDRs who are doing outbounding and writing really good personalized emails for our ABM, very hyper-targeted accounts.
So I do think having them work side by side is right now the key. That's where we're going to see the most benefit.
Heidi Wurpel – Emburse
I think so too. Again—it’s a tool; it's not a strategy. Your strategy needs to be clear and concise. You need to know what you're trying to get your outcomes to be. And part of your unique selling proposition is telling people clearly what your unique selling proposition is. Don't let the AI write that for you, because it's going to write a similar thing for all companies. It's a mass aggregator.
So you need to stay unique. You need to have those qualifiers. And I think that's the human touch and that set of control.
I also think something that—along with trying to carte blanche replace whole roles—we could get wrong is that we are so used to being able to control every word, every sentence, every message that comes out. And you have to be able to go with the flow a little bit when you move into an AI world.
If you have very specific and controlled outcomes that you want to occur—like “The company desires this sentence to always be written this exact way”—that's not going to happen anymore. What you need to change your focus to is ruthlessly focusing on your customer's needs and meeting those.
Does it get down to the heart of answering their question correctly, in an accurate way? Try to avoid hallucinations, but we know AI models today hallucinate. If you're someone who's very rigid, there are industries where you have to be very careful with every single sentence.
And what's cool about AI is you can say, “Anytime you use a sentence like this, write it this exact way.” You can train your AI SDR to do that. But I just think you need a little flexibility—a little, to your point, skiing analogy. If it's an icy day or a powder day, you're going to ski it differently. And you need to trust that AI is going to ski that terrain differently too. It's going to know what to apply, but the outcome is going to be different each time.
Sarah McConnell – Qualified
I totally agree. Okay, to wrap things up, Heidi, we have our lightning round. A couple of quick questions to wrap things up that are a little more fun.
The first one: besides ChatGPT—because that tends to be most people's answer—what was the first AI tool that you experimented with?
Heidi Wurpel – Emburse
PhotoLeap AI Headshot Generator. Yeah. My hands were weird, my ears were weird. I'm not going to lie. But a few came out okay. But it was actually an AI headshot generator. In the end, I did go and get a professional photographer—a real human. So here’s—right—I explored and said, nope, I think a human’s going to be better for this. But it was actually a headshot AI generator that I played with first and foremost.
Sarah McConnell – Qualified
I like that one. Okay, what do you think is the most overrated buzzword in martech right now?
Heidi Wurpel – Emburse
Ooh. The metaverse—that we all want to be digital AI avatars always plugged in. I think that's overblown. I don't think that's going to take over the world in the next 12 months anyway.
Sarah McConnell – Qualified
I'm with you on that one. You mentioned someone earlier, but is there one marketer who you think is really ahead of the curve on AI right now that you like to follow and you'd recommend others follow on LinkedIn or other channels?
Heidi Wurpel – Emburse
Yeah. So Erin Mills, who I have worked with before, but she's doing a really great other podcast. So if you're into podcasts, check out Erin Mills' podcast as well on LinkedIn. She talks about AI on a really regular basis and brings on some interesting guests. It's always fun to see her and what she's doing in this space.
Sarah McConnell – Qualified
Amazing. And last question—if you could automate one part of your life outside of work with AI, what would it be?
Heidi Wurpel – Emburse
Okay, I'm staring at a Monstera that needs a little bit of love. Probably my houseplant care. I have a green-enough thumb. It's green, but it's not deep verdant green here—they're struggling. If I had something that pruned them just the right way at the right time and moved them into the sun and gave them the right amount of water and food, I would absolutely invest in that. Because sometimes when you travel for work—and things which I hope everybody gets the opportunity to do because it's not that bad to do your reimbursements; use Emburse, it is fun—but when you travel, your plants do that death shrink. And that would be really nice.
Sarah McConnell – Qualified
Yeah, I would also invest in that. I'm thinking about the multitude of houseplants I have in my house that are doing okay—just good enough—but not great.
Heidi Wurpel – Emburse
That's… yeah. I'm sorry, Monstera.
Sarah McConnell – Qualified
Okay, Heidi, well, thank you so much for joining us on The Agentic Marketer. It was so great to have you today.
Heidi Wurpel – Emburse
Thanks for having me. What a pleasure.
Stay up to date with weekly drops of fresh B2B marketing and sales content.
Discover how Emburse drives real AI adoption by creating a culture of experimentation. Heidi Wurpel shares why giving teams “permission to play” accelerates agentic marketing.
In this episode, Sarah McConnell sits down with Heidi Wurpel, VP of Marketing at Emburse, to explore why the true foundation of AI adoption isn’t tools or technology. It’s culture. Heidi breaks down how giving teams permission to play with AI accelerates learning, sparks innovation, and builds the confidence needed to scale agentic marketing across an organization.
She shares how Emburse empowers marketers to experiment safely, celebrate early wins, and iterate fast, creating an environment where AI isn’t intimidating. It’s exciting! From modernizing content workflows to training AI SDR agents like Kira, Heidi explains the mindset shifts leaders need to make if they want AI to actually work inside their teams.
What you’ll learn:
TRANSCRIPT
Sarah McConnell – Qualified
Heidi, thank you so much for joining me today on The Agentic Marketer. I'm so excited to have you. Before we jump into anything, I'd love for you to tell me a little bit about yourself and the work that you're doing over at Emburse.
Heidi Wurpel – Emburse
Sure. So my job is the Vice President of Emburse and Marketing. And I basically take what is currently one of the world's largest travel and expense management companies to even greater heights by spreading our expense intelligence story powered by our incredibly innovative AI. And I've been in marketing for about 20 years now, starting in brand and corporate comms, luxury B2C during the dot-com boom and bust. So that was fun. And then I acquired my MBA, worked in market research and data, spent a little over a decade in tech and ops, and now, of course, I'm focused on leadership, including demand gen and go-to-market for Emburse.
Sarah McConnell – Qualified
Hahaha! Amazing. I love what you guys are doing over at Emburse. I've used your product before. I never realized until I got into tech how cumbersome travel expenses could be. So I love what you guys are doing over at Emburse. I've used the product. It is definitely a huge time save.
Heidi Wurpel – Emburse
We love it, and we're doing even more. It's coming down the AI roadmap to make it easier on you day in and day out, because travel can be fun, and we want it to be.
Sarah McConnell – Qualified
Amazing. Okay, so obviously, the title of the podcast, The Agentic Marketer. So Heidi, it's a new term. People aren't quite as familiar with it. How do you define agentic marketing, and what do you think it means to be an agentic marketer right now?
Heidi Wurpel – Emburse
Yeah. So I actually think agentic really is just being—like, it's a data-based way of saying the right thing to the right person at the right time. Think of it as a personalization mass multiplier. And I think that agentic marketers in particular are really fixated on getting prospective customers the answers they need as quickly as possible, as easily as possible, and truly tailored to their unique problems. And the data sets are what allow us to do that today in technology.
Sarah McConnell – Qualified
I really like that answer. I feel like it's such a—people, you know, agents, AI, it's a very loaded term, but I really like the way you phrase that, if it's more about getting people the exact right thing at the exact right time. And now with AI, we can do that so much better than we were able to before. Like, we've been talking about personalization and marketing for a really long time. That's not new. It just feels like we've entered a very new time where that personalization is next level. It's just so much different than it was.
Heidi Wurpel – Emburse
Yeah, I think that today data and technology are enabling us to do what—like, if I get on a little bit of a high horse here—like people in proximity. Again, I was in… my gray hair comes naturally. Like, I was here during the dot-com boost, right? And then a little bit of a burst. And when you think about it, in the past, before e-comm went wild, before things went huge, you generally interacted with the same people day in, day out. And they were local people, right? So they knew your preferences. They knew maybe your family's preferences, et cetera, et cetera, et cetera.
And I think we have this deep want to return to that. But we've gone so far away from it these past 30-plus years to sometimes really good things like price transparency and globalization. There's definitely positives about what we've done these past 30 years in marketing. But I just think today our databases, our data structures, and most importantly really the agents that allow us to interact with that data and those databases and structures are really what differs from traditional marketing. But it's almost a return to like traditional-traditional marketing, like the pre-e-comm days of the world, in my mind.
So I'm pretty excited to be able to connect people with the information that they need faster and more easily and let them feel seen and known and like we're adapting to their unique needs in that way.
Sarah McConnell – Qualified
Yeah, I like this return to the traditional-traditional marketing. And I always think about—over the time in my career—everything in marketing feels very cyclical. Like, you know, there's always this dead, like, “This isn't used anymore,” and then it always comes back. And the example I always use is outdoor billboards. I remember no one was paying for billboards. If you asked, “Can I go spend this money on a billboard?” they would laugh at you. Absolutely not. And all of a sudden, if you're in the Bay Area, they're everywhere. Everything does feel very cyclical.
The traditional marketing—it's just evolving, but those basics are still there, and we do need to return to those. So I really like the way that you phrase that, because it is almost like we're just going back to it.
Heidi Wurpel – Emburse
Yeah, it's just a new tool, right? Like, I think there used to be a magnet up in the marketing department at the last organization I worked at. It's always saying—this is during the app bubble, really—“An app is not a strategy; it is a tool.” And I think of AI and agentic AI in that same way. Agentic AI is not a strategy in and of itself. It is a tool that enables us to do what we want to do faster, better, more easily, and maybe getting our end users and prospects what they need quickly in that space. Yeah.
Sarah McConnell – Qualified
Yeah. And to your point of it being a tool, I'm curious—with this new tool that's agentic AI—it’s really easy for people to experiment. I think just given things like ChatGPT and your ability to build, and Gemini, and Perplexity, and all of these things, people are experimenting more heavily with this technology than we ever have with previous technology just because of the availability of it. But if you want to move from that experimentation to actually leading with it and using agentic AI as a tool across your entire organization, what do you think it takes to make that shift?
Heidi Wurpel – Emburse
Yeah, I think having extremely clear goals and strategies is important. You need to know what you want your outcomes to be and be very focused on that. And I think for many of us as marketers, that means having excellent original content repositories, which include concise, fulsome, accurate product marketing assets and sales scripts—things that you want to become external-facing, but they need to be there. They need to be concise. They need to be right.
I think you also need clean data and integration capabilities to fully scale this thing. But I think more than anything, what we've found in our organization is you also have to have a culture that's really willing to rapidly pivot and use. And I don't know if this is—again, I spent over a decade in tech, so I don't know that all marketers are comfortable with CI/CD—but continuous improvement, continuous development principles are that idea that instead of changing the whole app user interface, make the button blue with a little more rounding. It's not a whole rebrand; it's a little refresh or a little evolution on what you've been doing.
And I think if you have a team who culturally is comfortable with that—these fast pivots in small ways where maybe it's not going to come to great fanfare, maybe you're going to get pushback on “That's wrong, that's not good, people don't like that”—okay. What's cool about agentic AI is great: you can train the model rapidly and easily to adapt. It's that adaptability again—continuous improvement, continuous development—constantly checking in and making those changes to train the models and make them better and better and better over time.
And realizing—you know, like you and I were talking about billboards—right? You need to get a billboard right. You print a billboard; it's the billboard. It was expensive to paste. And these days with digital billboards, you have a little more leeway, but non-digital billboards—what's on the billboard is what is on the billboard. If you have a typo or an error, that's on you.
That is not agentic AI. It is a living, breathing, changing asset day in and day out. And so I think that culture really matters. But again, have those really clear strategies, those really clear goals, and then start thinking about the backend of what's going to feed that, and then be willing to feed it rapidly and pay attention to it like you would, I don't know, any employee or any piece of information that you're putting out.
Sarah McConnell – Qualified
And we tell people that all the time. So I really like that you hit on the point of it being continual, and you have to go back and be willing to rapidly train and test. I think in the age of AI, things are moving fast—your company, how you're putting out information. And we obviously have our own AI SDR; her name is Piper. But when we talk about onboarding her as an agent, I always caution people: this isn't a set-it-and-forget-it. You can't just give her information—or whatever you end up personifying your agent as—you can't just give it information and expect it to keep up with your rapidly changing organization.
Every time something on the marketing team changes—your product marketing, your messaging shifts, you release new assets, you have a new event—it does have to become a part of your workflow if you want to lead with AI and not just experiment with it. You have to give that new information to any of the models, and that involves any of your agents or AI that you're using internally, to utilize that information and then make sure it's using it properly and give it that feedback.
And I think, to your point, Heidi, that is where people kind of fall off on going from experimentation to leading, because they tend to try it, they forget about it, then all of a sudden it's not giving quite the right answers or it's not doing it exactly the way they wanted. But it does involve pretty good oversight from the people on your team, I think.
Heidi Wurpel – Emburse
Yeah. You know, I recently read an article, and I'm not going to get his name right. I think it's Joe Carveo. He's a marketer-operator for a PE firm. And Joe wrote this thing about how, on his hierarchy, he actually slates agentic agents in as if they're people on his hierarchical org chart, so that it's really clear who the manager of that agent—or that person, what would traditionally be known as a person—is.
I don't know. That kind of blew my mind the other night. It's not my hot take; it's more of his hot take. But I think that was a really interesting take on AI and how we train it. And I think that's part of what made us go to Qualified. And we named ours Kira—you named yours Piper. So part of what we really liked was this idea that you could give her—you know, personifying here again—but you could give her information, and the moment that we update a blog, the moment that we update our landing pages, the moment that we update our data sheets, she’s instantly learning all of that information. She's ingesting that and finding ways to share it out.
Versus the old way of things where we had to go, “All right, there’s a playbook that relies upon that information. Go and remember to update that playbook.” And you know, like you and I talked about earlier, Emburse is an incredibly innovative company. We are rapidly releasing features and products and bringing new things to market. So to keep up all of those data sheets, to remember every playbook that's ever been written and redo your script—the nuance of that? Huge time sink. Huge.
The person who was managing it was just like, “My gosh, this is going to be in 30 places. This is going to be in 60 places. This is going to be in 300 playbooks.” That's hard. That's really hard. And it's not the modern way of doing things. And so I think that that's part of what was really great about Kira—or Piper—and Qualified: we could rapidly train her with the information that we're already bringing out to market.
Also, that would be good when you launch a campaign. You want certain information to hit the market at certain times. So it kind of takes care of any worries like that too—that her information is going to be as up to date as your marketing releases and very aligned with your campaign strategy. So those were some unique things that I thought really are just part of an agent that differ slightly from a human being doing enablement and making sure they don't say anything before X, Y, Z date or whatnot.
Sarah McConnell – Qualified
Yep, that is so true. Now, you mentioned you renamed your AI SDR Kira, which is great. I'm curious: what was it about an AI SDR that had you and the Emburse team leaning into that? I know when AI SDRs hit the market, there was a lot of confusion around them. So what was it that made you and the team say, “We think this is something that will be important to incorporate into our demand gen motion and our larger marketing organization”?
Heidi Wurpel – Emburse
Yeah, I think a couple things. One, my company, Emburse, is leading with AI. We are an AI-based company. So our culture embraces AI. It believes in AI. It's bringing really cutting-edge AI to market at times. So the culture was there. The trust in AI was there. And that's really important.
I think then, from a pure marketer standpoint, saying we need to be able to do just as much quicker. How do we make sure that we train with accurate information quickly and not have to rewrite all those if/then statements that were in the old playbooks in the way that we used to do things?
Having Kira to talk about our innovative AI roadmap, as well as trusting her AI to do what she does best—which is share, again, that right information with the person who's asking at the right time on a regular basis when it's needed.
Sarah McConnell – Qualified
Yeah, that's amazing. And now, is there anything unexpected that bringing on Kira changed for your team? Was there something you weren't expecting that you've now noticed having her as your AI SDR?
Heidi Wurpel – Emburse
Yeah, you know, just the other day I was reading through the questions being asked of Kira, and I noticed it's great to get content guidance. And I think I knew this, but it wasn't what made us make the decision. But it's really nice—it shines a light for me on those areas that our marketing is not clearly answering. And that's why earlier I was like, “Be concise, be accurate.” Because if you see the market asking the same question over and over, the likelihood is what you have out there right now is not enough. It's not answering this question.
So I think I've been pleasantly surprised—or maybe pleasantly reminded—of the fact that you can see trends in the data, and that that can even help inform your content strategy and even bigger strategies if you want to explode it up a little bit more for future years in your marketing.
Sarah McConnell – Qualified
Yeah, that's one of the unexpected things we experienced onboarding Piper at our company. I got to oversee that, which was amazing. And I remember looking at the questions being asked. And I tell customers all the time: you're going to have your nine to ten questions that get asked pretty consistently on your website—that is universal. But there are going to be aggregates of questions that you don't think about that really will help inform your content team.
And that actually kind of shifted our org structure here at Qualified to have content and communications roll up under my team on demand gen, kind of for that reason. I was getting this feedback from the questions that were being asked. I was able to easily see it in Qualified. And I was like, I need to be able to move fast and make sure this content is now out there because I'm obviously seeing in real time we aren't answering that question well enough. We don't have the resources out there in a way that's making it easy for visitors to find us.
Now granted, we can train Piper on it, and that's fine. But we want to make sure across all of our different channels that we have the right information. So it was something that actually impacted our team pretty heavily that I was not expecting.
Heidi Wurpel – Emburse
Yeah. And I think a few years back, I was at an SEO conference, and it was really interesting. It was talking about looking at Google’s other questions that are asked—“People also ask.” And I know that's kind of a thing of the past now. It's way at the bottom. But at the time, that was at the top. It was sort of that “aha” moment where you're like, right, you can use AI to answer all these questions and make sure you get similar content and questions answered on your site because maybe you're thinking about this—expand it a little to that.
And I think that's what this agentic AI is really doing for you, but it's doing it in a hyper-personalized way. It's yours. It’s not like Google used to be—like, the general questions out there in the market are… Well, this is hyper-personalized to: “Hey, Emburse marketing team, ten people asked this thing.” Clearly you need some more information about that on your website and across your channels. So yeah, that was a really pleasant surprise.
Sarah McConnell – Qualified
Now, I'm curious—what metrics are you measuring to know that Kira is working, beyond obviously the content that's harder to measure? Is there anything else that you're looking at from a metrics standpoint pretty consistently?
Heidi Wurpel – Emburse
Yeah, you know, I mean, for us, I'm really proud of our ability to see attribution. And we linked Kira into our Salesforce. And so we're able to see specific MQAs that come through—our marketing qualified accounts—and be able to connect them to: this was a Kira conversation, or this was a Kira-qualified ask for a meeting booked. So we look at MQAs; we look at meetings booked.
And we've seen about a 212% improvement in the past six months alone. So really good numbers there. And then, of course, I'm watching her accuracy of responses, as is my Head of SDRs. He's really fabulous because we look at her as an SDR in that meetings-booked and sort of initial discovery, “disco” type of phase as well. We treat her a bit like an SDR meets a marketing lead source, and we measure her on both of those metrics.
Sarah McConnell – Qualified
That's amazing. And now I want to talk a little bit more. It's very obvious from this conversation that Emburse is—you mentioned—very AI-first, not only as an organization and how you're adopting technology, but also as a company in general. So I feel like you are more than qualified to answer a question for someone if they're looking to adopt AI.
So say someone's listening to this and they are not where I feel like you or Emburse are at, and they're like, “I don't know where to start. This feels very overwhelming.” What would be your advice, having gone through this journey already, that you would tell other marketing leaders? Where should they start when they're looking to adopt AI or become more adept at agentic marketing?
Heidi Wurpel – Emburse
Yeah, I think AI adoption really does come culturally. And if you're a leader in your organization, I think you can encourage your people to play, right? What's nice about being a marketer is most of the information that we share is information that we want the world to see. So I think a lot of the time people are really concerned. So you can put guardrails up about: don't share anything company-confidential. But if you want the world to see the information—share the information you want the world to see.
Go ahead and play a little bit. Sometimes you're going to flop and fail. Other times you're going to have wild success. And I think for me, one of the first things that I did was—in my digital innovation team—they really started playing with it, and we started highlighting those areas of success. We were highlighting their successes to the rest of marketers, and that gave sort of permission to play. Like, “Hey, so-and-so did this. It's not 100%. But look how great it is. This is what it allowed us to do.”
And so we took a few problem statements that we'd been putting on the back burner for a while because we thought, gosh, finding a solution to that is so big and so hairy. And then—we’re a Google shop—so we were like, “Wait, we have NotebookLM.” Let's just play with it and see what it can do. And we were like, “Wait, that really worked.” And we could then expand it here and there. What other uses could we do?
And so I literally have a monthly show-and-tell where my team brings really cool new things they've shared. I learned about Nano Banana from them before I saw any press releases on it. So it's just really neat to see the different qualities of your team. So I think allow people to play within these safe data sets and within this safe information—nothing that's going to hurt or not be able to be shared out there.
So encourage the culture and, like I said, celebrate those successes.
Sarah McConnell – Qualified
Yeah, absolutely. I really like—we’ve heard that answer quite a bit on this podcast, and I just cannot emphasize it enough. If you want to lead and be at the forefront of this, you have to not only allow your team to use AI—which, it's wild to me, but there are still companies out there that aren't letting their employees do that. But I like what you pointed out: champion them. Champion the people that are doing it well. Give them spaces to come to, to show off what they're learning and what they're doing.
I think that's such a fast way to get the team excited about something and willing to get their hands in it and start using it themselves. Being able to champion those employees, I think, is so, so important right now.
Heidi Wurpel – Emburse
Yeah, 100%. You know, I really think sometimes as leaders—we all try to be approachable, right? We want to be really approachable to our team. But on the flip side, how often do you want to go to your boss and be like, “I don't really know about this thing”? Like, teach me about this. I don't know—maybe there are people who do. But a lot of us—the A-plus students—we want to come forth always with the A-plus things.
And so I think what this allows is you create these little knowledge experts. And you don't have to spin them as such, but these cohorts that can come together and privately say, “Hey, it’s really cool what you did there. Can you show me that? Can you teach me that?” And instead of having to go up the chain, it creates those horizontal connections and networks that allow people to ask each other the silly questions and really have respect.
Because if you're someone who has literally never typed a prompt into anything—and I think there are people absolutely still out there in the world, and no shade on that—that first prompt can feel a little scary. And literally just having another colleague there next to you to type the prompt, and have them watch and ask, “And now what do I do? And what button do I push?” It feels really nice—just to give that safety and that mechanism of: you can't screw this up. It's really okay. Go ahead and try something simple.
So yeah, that's what we've created in our marketing culture. And I'm really proud of the team, because holy cow, they bring things that I would never think to bring forth, and solutions to each other. But then just iterate on one another and have saved a great deal of time and energy and wasted cycles.
Sarah McConnell – Qualified
Now to that point—are there any particular skills you've seen within the team that you think other companies really need to double down on in order to thrive in this AI-driven era? Are there particular skills that you've seen within your team or yourself that will really help teams make that shift into an AI-first company?
Heidi Wurpel – Emburse
I think skill-wise: clarity, again. Being able to really know what you want to get out of something and really clearly articulate that the outcome should be like this, or that the output is for so-and-so.
I think having creativity is really important—being willing to ask the same question in four different ways and think in that way is really important. But I really just think it's that willingness to trust that the models will present at least something that you can start with and that you don't have to end with.
I love a blank canvas, but there are days when we are moving so fast that you don't have time for a blank canvas. It's really helpful to have—and I'm super geeky in the art space—for those of you who haven't painted, you have a canvas. You paint this white or pink or whatever color layer you want on it to fill in all the little bubbles of the canvas. Yes, that's meditative and great if you're a professional artist, but if you're not—if you're just at a sip-and-paint—that’s not particularly fulfilling. So you let the AI do that.
There's a certain horizontal line you paint often on a canvas if you're doing a landscape or something, or even just to give yourself a little guidance. Similarly, let AI do that. It's not all or nothing with this. It's not “Paint me a Picasso,” and then “That doesn't look like Picasso, I hate it.” Just let it whitewash the canvas, and you paint the rest of the image on it.
So I think that's it. It's not all or nothing. It's not black and white. Always, always edit whatever your AI is putting forth or retrain it in the agentic world. And so I think it all comes down to that willingness to iterate and have clarity of your final vision—you know what you want on that canvas at the end.
Sarah McConnell – Qualified
I think that's the best analogy I've heard for how to let AI support you versus replace you. I really love that. I might have to use that in the future because it's very clear, but also very beautiful.
Okay, Heidi, I have a couple of questions now looking forward. I know it's always hard when we're like, “Get out your crystal ball and tell the future,” especially in an AI era when things are moving so fast. But I'm curious—if you are thinking ahead 12 months, what do you think, besides AI, is going to be the single biggest change to B2B marketing? Like, what is AI doing right now that's going to make the biggest shift in the next 12 months for us B2B marketers?
Heidi Wurpel – Emburse
I actually think it will be more than just about marketing. So I think it's going to have a big impact on go-to-market. What I mean by that is: I think that we see that AI is an accelerator. It's a pace accelerator; it is a volume accelerator. And I think if we do this really right, we're going to see the sales velocity also decrease. Right?
When you think about the velocity of a sale—and this is more B2B than B2C—B2C, I see this shirt, I want to buy this shirt, I buy it off the rack. Maybe it takes a little while to ship if I'm doing e-comm. But when you think about B2B marketing, it takes time often in non-product-led growth models to go through the various phases of the sales cycle. And we know people do a ton of their research ahead of time and do it within marketing.
So done right, I think we're going to see the velocity of sales motions shorten. And I think we may even see disruptions to your traditional go-to-market sales plays, with an increase in product-led growth—even for those large global enterprise companies and expensive software.
So I think it's going to impact go-to-market if marketers do this correctly and create velocity acceleration on some deals.
Sarah McConnell – Qualified
Now on the flip side—if you're again thinking 12 months out—in the next year, is there something you worry marketers are still going to be getting wrong about AI a year from now? What do you think we're still heading down the wrong path on that we might still be doing incorrectly?
Heidi Wurpel – Emburse
Yeah. I think this concept of replacing entire roles or sets is still a little off. Do you need to hire as many of the same person? Maybe not. But that doesn't mean you don't need a product marketer anymore, or you don't need a content marketer anymore, or you don't need creatives anymore—just let the AI do it.
There’s that idea of keeping to your brand and keeping to your true north—that’s really important. So I think we could move too fast and we could get that wrong, for sure.
I'm hesitant when I hear people… you know, I love putting an AI head on the hierarchy, like I said. And I think that there's a lot we can do. We can increase pace. But I think that is something that could go awry if we let it go too far in efficiency, efficiency, optimization, optimization. Don't lose those skill sets. You need somebody to babysit the model.
I don't know about you, but I'm probably not going to hang an AI Picasso on my wall. I mean, I'm not going to hang a real Picasso because I don’t have money like that, but you know what I mean. If I'm hanging a contemporary, interesting portrait on the wall, I'm probably going to choose one that's out of paint from the hands of a true artist who put those final touches on it.
Sarah McConnell – Qualified
I cannot agree more. I do like the idea of incorporating AI into your org chart only from the standpoint of: it helps to know who's in charge. Who is the person responsible for making sure the AI and the model are running the way you want them to? But I do agree—I think it's always funny to watch things right-size. People go really hard, and I get it. As things change—as marketers—we tend to be on the forefront of wanting to test things and be on the cutting edge.
My CEO always says: if you're not a little uncomfortable, you're probably not pushing yourself hard enough, and that's good marketing. You really have to be kind of out in front of your skis a little bit. But in that process, we tend to right-size as things go on. And I think there was a shift of: could it replace entire teams? Could it replace entire roles?
And it's like—well—we’ve seen internally with SDRs: yes, we have an AI SDR, and yes, she can scale infinitely and do all of these things. But we’ve still found that we have changing roles with people that are now managing the AI SDRs. We have our inbound and our outbound AI managers. But also, we still have SDRs who are doing outbounding and writing really good personalized emails for our ABM, very hyper-targeted accounts.
So I do think having them work side by side is right now the key. That's where we're going to see the most benefit.
Heidi Wurpel – Emburse
I think so too. Again—it’s a tool; it's not a strategy. Your strategy needs to be clear and concise. You need to know what you're trying to get your outcomes to be. And part of your unique selling proposition is telling people clearly what your unique selling proposition is. Don't let the AI write that for you, because it's going to write a similar thing for all companies. It's a mass aggregator.
So you need to stay unique. You need to have those qualifiers. And I think that's the human touch and that set of control.
I also think something that—along with trying to carte blanche replace whole roles—we could get wrong is that we are so used to being able to control every word, every sentence, every message that comes out. And you have to be able to go with the flow a little bit when you move into an AI world.
If you have very specific and controlled outcomes that you want to occur—like “The company desires this sentence to always be written this exact way”—that's not going to happen anymore. What you need to change your focus to is ruthlessly focusing on your customer's needs and meeting those.
Does it get down to the heart of answering their question correctly, in an accurate way? Try to avoid hallucinations, but we know AI models today hallucinate. If you're someone who's very rigid, there are industries where you have to be very careful with every single sentence.
And what's cool about AI is you can say, “Anytime you use a sentence like this, write it this exact way.” You can train your AI SDR to do that. But I just think you need a little flexibility—a little, to your point, skiing analogy. If it's an icy day or a powder day, you're going to ski it differently. And you need to trust that AI is going to ski that terrain differently too. It's going to know what to apply, but the outcome is going to be different each time.
Sarah McConnell – Qualified
I totally agree. Okay, to wrap things up, Heidi, we have our lightning round. A couple of quick questions to wrap things up that are a little more fun.
The first one: besides ChatGPT—because that tends to be most people's answer—what was the first AI tool that you experimented with?
Heidi Wurpel – Emburse
PhotoLeap AI Headshot Generator. Yeah. My hands were weird, my ears were weird. I'm not going to lie. But a few came out okay. But it was actually an AI headshot generator. In the end, I did go and get a professional photographer—a real human. So here’s—right—I explored and said, nope, I think a human’s going to be better for this. But it was actually a headshot AI generator that I played with first and foremost.
Sarah McConnell – Qualified
I like that one. Okay, what do you think is the most overrated buzzword in martech right now?
Heidi Wurpel – Emburse
Ooh. The metaverse—that we all want to be digital AI avatars always plugged in. I think that's overblown. I don't think that's going to take over the world in the next 12 months anyway.
Sarah McConnell – Qualified
I'm with you on that one. You mentioned someone earlier, but is there one marketer who you think is really ahead of the curve on AI right now that you like to follow and you'd recommend others follow on LinkedIn or other channels?
Heidi Wurpel – Emburse
Yeah. So Erin Mills, who I have worked with before, but she's doing a really great other podcast. So if you're into podcasts, check out Erin Mills' podcast as well on LinkedIn. She talks about AI on a really regular basis and brings on some interesting guests. It's always fun to see her and what she's doing in this space.
Sarah McConnell – Qualified
Amazing. And last question—if you could automate one part of your life outside of work with AI, what would it be?
Heidi Wurpel – Emburse
Okay, I'm staring at a Monstera that needs a little bit of love. Probably my houseplant care. I have a green-enough thumb. It's green, but it's not deep verdant green here—they're struggling. If I had something that pruned them just the right way at the right time and moved them into the sun and gave them the right amount of water and food, I would absolutely invest in that. Because sometimes when you travel for work—and things which I hope everybody gets the opportunity to do because it's not that bad to do your reimbursements; use Emburse, it is fun—but when you travel, your plants do that death shrink. And that would be really nice.
Sarah McConnell – Qualified
Yeah, I would also invest in that. I'm thinking about the multitude of houseplants I have in my house that are doing okay—just good enough—but not great.
Heidi Wurpel – Emburse
That's… yeah. I'm sorry, Monstera.
Sarah McConnell – Qualified
Okay, Heidi, well, thank you so much for joining us on The Agentic Marketer. It was so great to have you today.
Heidi Wurpel – Emburse
Thanks for having me. What a pleasure.
Discover how we can help you convert more prospects into pipeline–right from your website.
.svg)
