Sean Whiteley & Ashley Kramer 21 min

When Marketing & Engineering Collide: Navigating Alignment in the Age of AI


The relationship between engineers and marketers has never been so central as it is in the Age of AI. As we all seek to learn and educate our customers on new tech, how can we partner with our product engineering org to maximize everyone's knowledge and success?



0:00

Welcome everybody. Welcome to the other edition of Pipeline Summit.

0:03

Today we're talking all things marketing, pipeline, demand generation, and we

0:08

're really

0:08

lucky to have Ashley Kramer here with us today. Ashley Kramer joins us from Git

0:14

Lab.

0:14

She is the chief marketing and strategy officer and she's been so kind to spend

0:19

some time with us

0:20

today. Ashley, given this audience are marketers, can you talk a little bit

0:24

about GitLab? If you're

0:26

a developer, everybody knows GitLab. But marketers might need a little bit of

0:30

an introduction,

0:31

so tell us a little bit about GitLab. Yeah, no problem. Thanks for having me on

0:34

today.

0:35

So basically at a nutshell, GitLab has a software platform we call it a DevSec

0:40

Ops

0:41

software platform that companies across the world use to deliver their software

0:46

in a secure

0:47

and fast way. And so we provide the software that people use to build their

0:51

software.

0:52

And so it's a pretty cool concept because as a marketer, we also use GitLab to

0:56

build GitLab.

0:57

So it's a really big opportunity to tell that story to the world, help people

1:02

figure out how to transform their DevSecOps practices. And that's a little bit

1:06

about GitLab.

1:07

Thank you and developer security operations. So you are selling into the DevSec

1:14

Ops persona.

1:15

It's appropriate to hear with us today because the title of the session is when

1:20

engineering and marketing collide. You're perfect for this session given your

1:25

journey.

1:25

Because usually when you see marketing colliding with something, it's typically

1:29

sales.

1:30

But given the rise of artificial intelligence and everything we're talking

1:36

about today,

1:37

you're perfect given that you started your career as a working at NASA, quite

1:43

literally,

1:43

and you've moved throughout engineering organizations. And now you're the CMO

1:48

and Chief

1:49

Strategy Officer at GitLab. So tell us a little bit about what that journey has

1:52

been like.

1:53

Yeah, it was interesting. I did. I was a computer science major. And as you

1:57

mentioned,

1:57

work at Det NASA or Oracle, some other companies in the capacity of a software

2:02

developer.

2:02

And then it turned out there were two problems with that. One, I wasn't very

2:05

good at it. And two,

2:06

I wasn't passionate about it. And I think that's okay because it gave me the

2:10

space level knowledge

2:12

of what it takes to develop software. And so I made some pivots throughout the

2:15

years into product

2:17

and then into product and marketing, product and engineering, as far as

2:20

leadership capacity.

2:21

And when I landed at GitLab, it was kind of perfect because what I market and

2:27

sell to our customers

2:28

is something that would have removed that pain that I experienced as a

2:32

developer. So I have that

2:34

empathy for who we're selling to and the challenges. And so it was sort of a

2:40

perfect circle for me to

2:42

to go through that journey, become the CMO of a company that sells to this

2:45

audience. And the last

2:48

piece I'll say, which is super cool, is I bought GitLab when I was at a company

2:52

called Altarix.

2:53

I was running product and a portion of engineering at the time. We were having

2:57

issues with quality

2:58

in our software, other things. And we brought GitLab in. There's still a GitLab

3:03

publicly

3:03

referenceable company today. And so I've been a buyer too, which is kind of a

3:08

cool part to the story.

3:10

Very cool. Very cool. So when you think about-- So GitLab's obviously been

3:15

leveraging AI and

3:17

machine learning. And you've been building AI into your workflows and your

3:20

product for a long

3:21

time. I'm somewhat familiar with your product. And by the way, thank you so

3:24

much for being a customer

3:25

as well. Absolutely. When you think about this sort of rapid-- What I like to

3:31

call sort of

3:33

more accessible AI, generative, it's funny to think that it was only November

3:37

30th, 2022,

3:40

is when OpenAI released or GA Open Chat TPD, which is amazing. So it's not that

3:46

long ago,

3:47

but it seems to have been the hottest topic on the planet. We talk about it as

3:51

a platform shift.

3:52

We've both been working this industry long enough to have experienced platform

3:57

shifts

3:58

from on-premise to cloud and mobile and social and all these platform shifts.

4:04

This one is extraordinary. Can you talk a little bit about how you think about

4:09

what this is going

4:11

to mean to businesses and how they're going to start to adopt some of these

4:16

rapidly emerging

4:17

technologies? Yeah, I mean, I think there's two ways to look at it. On the

4:22

first hand,

4:23

everybody right now seems to be when it comes to generative AI obsessed about

4:27

two things in particular. You mentioned one, chat TPD. I'm just going to go

4:31

type something and

4:32

magically exactly what I once going to appear, which doesn't always happen.

4:35

And the second piece is something's going to help me write my code for me. And

4:40

that one's

4:40

specific to the GitLab use case. And so we have competitors, a big competitor

4:46

in the market,

4:46

that came out with that one, a really flashy way to, hey, we're going to make

4:49

developers more

4:50

successful because we're going to help them write code. And so that's the first

4:55

piece in where you

4:56

have to start sort of peeling back and saying, but is that how software is

5:00

delivered? Is it just

5:02

about writing the code? Because our perception at GitLab is it's way more. We

5:06

talked about being

5:07

a DevSecOps platform. You have to plan. The product managers have to plan what

5:10

's going to happen.

5:11

Fellovers have to write it. They have to test it. They have to integrate

5:14

security. You have to get

5:15

it to market. And so while there's a lot of hype and buzz around these certain

5:20

features or

5:21

capability, we are grateful that we've been a platform since we were founded in

5:25

2011. We have

5:28

one end-to-end workflow that can make these professionals really successful.

5:33

And so we want

5:34

to integrate AI throughout, which means we might not be first to market with

5:38

some of these capabilities

5:39

like code completion, but we want to be the most secure, fast follower. And we

5:44

want to make the

5:45

entire process engineers successful across the entire process, not just allow

5:51

them to produce

5:52

a ton of code faster because if it's not secure, it's not at quality, it's not

5:55

merging with other

5:56

code. It's just going to sit their still. So that's the first piece. The second

6:00

piece is when you're

6:02

typing things into chat, GPT, we've seen the articles. Are you sure you know

6:07

where your

6:09

confidential information, if it is, is going? Is it being used to train other

6:13

models? When you're

6:14

using some kind of code completion, are they going to take the code that you're

6:19

putting in and the

6:19

refactored output and use it for maybe a model that your competitor might go

6:25

leverage? And so

6:26

that's the second piece is we spend a lot of time at GitLab talking about the

6:31

large language

6:32

models, the LLMs, and can we keep, we sell to highly regulated, very big

6:40

enterprises that this

6:42

isn't an option to leak their IP? And so the second piece is once we break

6:47

through the hype,

6:48

what is happening with your data? You can put this in your personal life too.

6:52

You can put this in

6:53

pipeline generation. Like, do you deeply understand if you're sharing any of

6:57

this in the world and if

6:58

that's okay? Yeah, you make a lot of really good points. You're talking about

7:04

the reality of

7:05

leveraging artificial intelligence in enterprise, really, you're thinking about

7:09

things like governance,

7:10

you're thinking about privacy, data security, transparency. These are all like

7:16

things that you

7:17

got to have to leverage this in the enterprise, right? That's right. I also

7:20

completely agree that

7:22

the infrastructure between whatever sort of corpus of data you have in those

7:28

workflows and these

7:29

large language models, like that's where you're building kind of defensible AI.

7:32

That's where you're

7:33

building something very unique. We share the same sentiment and we're learning

7:37

here as well and

7:38

we're learning there's a lot to do. It's not as easy as making an API call to

7:42

open AI.

7:42

So as you think about this world for marketers, right? I mean, there's a lot of

7:49

very clear use cases

7:50

for developers. They've been using this for a while, right? But as it relates

7:56

to marketers and

7:57

some of the capabilities that this unleashes, like, what are the things that

8:00

immediately come to mind

8:01

for you? I think of all of the things that a marketer, no matter what role you

8:06

have in marketing,

8:08

all of the things you wake up and you think, I have to go do this today, right?

8:12

The mundane,

8:13

the, I think about all of the things that you can automate because, you know,

8:19

when this all

8:20

started coming out, I got the question all the time, is this going to replace

8:22

my job, particularly

8:23

in certain roles? And my answer is no, it's going to make you a lot more

8:28

effective in your job because

8:29

you're going to automate all of those things that you don't like to do. And

8:32

then you're going to

8:33

actually use your brain for what it's there for, which is the creative side,

8:37

the thing that AI can't,

8:39

you know, cut through the personal interactions, the whatever it is depending

8:44

on your role,

8:44

you'll be able to spend more time on that and differentiate that way when it

8:50

comes to your role

8:51

in marketing. Agree. And that's a lot of people have sort of used this concept

8:56

of a co-pilot,

8:57

whether you're coding, whether you're writing some content, whether you're

9:01

generating imagery or

9:03

graphic design, you know, a co-pilot is, I think it's really a good term. It's

9:07

also a great way to

9:09

ease into AI because it's human in the loop. And it's a great way to start

9:12

really testing

9:14

as LLMs as powerful as they are. They make shit up. That happens. We all know

9:20

this hallucinations

9:21

are pretty commonplace. And the first sort of realization I had around, you

9:25

touched on something

9:26

I think pretty important, which is like, say, I'm going to come for all of our

9:29

jobs, right?

9:30

And, you know, my first experience with mid-journey, which I'm a, I'm a junkie.

9:35

I love it. Stable

9:36

diffusion in mid-journey. You can usually find me popping around in there. You

9:40

very quickly

9:41

realize that somebody who has a firm grasp on a lot of design concepts is going

9:48

to be able to do

9:48

a lot more with this tool. It's not about, you know, I think everything's going

9:52

to have a natural

9:52

language interface, apps and data. You're going to use a natural language

9:56

interface to communicate

9:57

with it. But the more you know about something, the more leverage you're going

10:01

to get from it. So,

10:03

I do agree with the fact that it's going to do a lot of the work that you roll

10:08

your eyes at in the

10:08

morning. That's ideal. But I also think that people who have a firm grasp or

10:13

domain expertise

10:14

around something are going to be able to be that much more effective with the

10:18

technology.

10:19

At Qualified, we've really challenged everybody in our business to think about

10:25

how they're going

10:26

to use it in their specific business function. And I've seen you're an advisor

10:31

to a lot of companies.

10:32

A lot of those are Martec companies. What's the thing you say to that? When you

10:38

talk to companies

10:39

and marketing leaders, what's the thing that you drive home in terms of if you

10:44

're not thinking

10:44

about this, you're making a huge mistake? Well, for me, number one is if you're

10:52

not thinking about

10:53

keeping people's data private. And you mentioned it, it's on my LinkedIn, so I

10:58

'll say it. But I'm

10:59

an advisor for Jasper AI, which is the content side of the house. And I just

11:05

went to their

11:06

company kickoff and did a fireside chat. And we're looking at using that at Git

11:12

Lab. But I won't

11:14

even let my team touch it until we have the legal, we understand what's

11:19

happening, what models are

11:20

being leveraged because particularly as a public company, that's a giant risk.

11:25

If anything that

11:26

we're putting in is going to be shared. And so for me, that's the number one

11:31

thing. I'm on a board

11:32

of a company too. It's not it's not Martec, but same thing they were using one

11:35

of our competitive

11:36

products. And I said, have you read the terms on that? Are you sure you

11:40

understand what's happening?

11:42

That's the first thing. The second thing is I worry when new hype comes out, it

11:48

's AI right now.

11:49

It's going to be tool here tool here tool here tool here tool here like I want

11:52

consistency. And I

11:54

told the team there in our own courage, you offer this to give me one place to

11:58

go, whether I'm a CMO,

11:59

whether I'm a day to day user of the product or platform that you're selling,

12:04

put the analytics

12:05

in there do everything in one place and drive people to that don't be just

12:09

another tool,

12:10

because there's going to be 50 million of them by the time we get out of, you

12:15

know, this whole

12:16

hype cycle, be the one that everybody comes to everybody can get the

12:21

information they need from.

12:22

Makes perfect sense. You have a DevSecOps platform, so everything you're saying

12:28

resonates with me.

12:29

One of the things that we think about a lot, and we share your sentiment with

12:36

regards to governance,

12:37

data privacy security, because I like to say that as it relates to AI,

12:41

especially in the enterprise,

12:42

we're at step three of a 10k or, you know, 100 yards into a marathon, right? It

12:49

's very new.

12:50

It's nascent. Obviously, it's exciting. Obviously, it's powerful. And obviously

12:55

, you're not going to,

12:56

you can't afford not to invest here. You know, I think people feel that. But at

13:02

the same time,

13:03

we see companies forming AI councils. And I think that's really smart. I think

13:07

that they

13:08

have a centralized place to sort of ask the right questions about if we're

13:11

going to onboard a vendor,

13:13

you want to make sure that they're responsible in their AI development and

13:16

their practices and

13:17

their governance. This is something we talk about a lot here. Do you think that

13:23

marketing teams,

13:24

I mean, engineering teams probably have this top of mind, I think. How about

13:28

marketing teams? Do

13:29

you think marketing teams have the same rigor in terms of the process to adopt

13:34

AI technologies

13:35

into their business? I think it's the responsibility of the company to ingrain

13:40

that, for sure. And

13:42

I will say, you know, we did that from the very beginning at GitLab. We had,

13:46

you know, the company

13:46

message, we repeat it. We're not allowed to use certain things like chat TPT on

13:51

our work computers.

13:52

Do I still think some people are probably? Do I think it's engineers? No. So,

13:58

so do I think we need

13:59

to continue repeating that message? Yes. Because the fact of the matter is you

14:03

and I are probably

14:04

more technical than some marketers out there. So like when you start talking

14:09

about LLM and

14:09

hallucinations and IP being shared, I think it's harder sometimes for them to

14:15

conceptualize what

14:16

that means. And so it's continuing to repeat that message and continue to help

14:21

them understand at

14:22

the right level why this could be a risk. Translate's another personal life too

14:28

, by the way. It drives

14:30

me nuts when I see people on social filling out these like surveys and putting

14:33

their picture in

14:34

and I'm like, oh, there's just capturing all of your data. This is why my

14:37

parents are not allowed on

14:39

on certain social sites. They will participate in those. Yeah, 100%. You know,

14:46

I think that the

14:47

onus does fall on the company to, you know, when we sell products into

14:52

companies like we're very

14:54

proactive at, you know, these are the questions you should be asking. And here

14:59

are the answers to

14:59

those questions. You know, we really do feel like you need to lead, be privacy

15:04

first. That's,

15:05

we think that's very important. And I think a lot of us, we've got a lot of

15:07

sales force DNA in here

15:09

and at Salesforce, obviously, we, you know, we really need to convince everyone

15:12

that your data

15:13

was okay in the cloud. You didn't need to sort on premise. So privacy first is

15:17

a, I think it's a,

15:18

it's an important motion here. When you, you've got a pretty big team at GitLab

15:23

. And, you know,

15:24

when you think about, you know, not only the entire company, but you're just

15:27

your entire marketing

15:28

team, like what kind of process do you have for evaluating and introducing new

15:35

tooling and new

15:36

technologies across the various marketing teams, your average company?

15:40

Yeah. I mean, I have, I have a VP leading each functional area. And that does,

15:45

it is a large

15:46

team because sales development, it's different in different companies. It does

15:50

report under me

15:51

in marketing and they're the ones that actually discover that VP was the one

15:56

that came to me and

15:57

said, it's time to replace what we're using. We want to use qualified. And so

16:00

we call them the

16:02

directly responsible individual, the DRI who will be leveraging the tech the

16:07

most is the one for

16:08

to go and seek out and strategically figure out where it fits. Then we have a

16:12

marketing ops team

16:13

who then takes it and figures out how's it going to fit into our stack.

16:17

Think the boring stuff like licensing, we're still paying on this. What are we

16:21

going to do

16:21

about that? And then I have that person as the responsible individual right now

16:26

of the privacy

16:27

side and saying, okay, let's take two double clicks, go talk to legal, make

16:31

sure that we understand

16:32

if this is going to keep our information, whatever we're using for it safe and

16:36

secure. And so it's

16:38

sort of functionally, you find what you need to, and I am pushing the AI thing

16:42

right now, I think

16:43

at GitLab, we're thinking a lot about AI in our platform. I don't know that we

16:48

're thinking

16:48

enough, at least in marketing about AI to improve our business practices and

16:53

our processes and

16:54

those types of things. Yeah, absolutely. And it's kind of interesting because

16:59

you see both

16:59

side of the coin, right? You're building, shipping, selling, supporting

17:04

products that are leveraging

17:05

AI and you're having those conversations at the same time, you're consuming new

17:09

AI products as well.

17:10

You're obviously a perfect person for GitLab, given that you've had, you have a

17:16

very strong

17:17

technical foundation, you have a lot of experience as an end user of that

17:22

product literally.

17:24

When you think about GitLab and how you're operating, do you think that the

17:29

engineering team

17:31

and the marketing team are thinking about AI in similar ways or do you think

17:34

there's,

17:35

they think about those things differently? So what we did when the hype cycle

17:40

started

17:41

earlier this year, that's when it really started taking off is we formed a

17:43

cross-functional group

17:45

that includes engineering, product managers, marketing, and a few people from

17:49

the field,

17:49

like we have field CTOs that are in front of customers all day. They met every

17:54

day for a while

17:55

and now we have the forums, like the Slack channels. And so I think the answer

18:00

in this case is yes,

18:02

not always, but when it came to AI, because this was such an important focus,

18:06

we really fostered

18:09

that interaction. We're asking the field to bring us back the clear feedback,

18:13

okay, you just

18:14

demoed this or the customer just tried this in beta, did the messaging land for

18:19

them and is the

18:20

product working and are we messaging it the way that it should be working? And

18:23

so in this case,

18:24

yes, because of the way that we've done it, not always though when it comes to

18:28

engineers

18:29

and marketing. We speak marketing and they are deeply technical and bridging

18:33

that gap,

18:34

critically important. That is actually, that's exactly what I was going to ask

18:39

you next,

18:39

given that you've made this journey, you've made this transition and knowing

18:44

how much you had to

18:44

learn along the way, how do you bridge those two worlds at GitLab? In my case,

18:50

because I have that

18:51

background, I try to lead this by example. And a lot of the team that I've

18:55

brought in have been at

18:56

other technical companies, but I have a code word for them, which is not a code

19:00

word at all,

19:00

which is this is too much marketing speak. And so it's at some company at Sales

19:05

force, it was probably

19:06

great, honestly, because you were selling to go to market leaders and that

19:11

works. That's why they

19:12

use the celebrities for all of their promotions and everything, but it doesn't

19:16

work in DevSecOps.

19:17

And so I personally try to lead and then the team follows the lead of this

19:23

maybe too much

19:23

marketing speak. We also are not afraid to float something by our CISO or our C

19:27

TO or our product

19:28

leader and say, hey, does this land for you? Because you are the people we're

19:32

selling to at

19:33

other companies. And so we have that collaboration as well. Yeah, absolutely.

19:38

Actually, I feel like

19:38

I could talk to you about this all day. You have a fascinating journey. We're

19:43

big fans and we really

19:44

appreciate you joining us today and sharing some of your insights. So thank you

19:47

very much. Yeah,

19:48

appreciate the time. Enjoy the conversation. Thanks.

19:50

[Music]

20:15

[Music]

20:39

[Music]