Change How Your Team Works With Dialpad AI

0:00

[MUSIC]

0:05

Hello everyone and welcome to Go to Market AI,

0:07

the future of your Go to Market tech stack.

0:09

I'm your host Sarah McConnell.

0:11

In these days, it seems like every company has AI.

0:13

But on this show, we want to go a level deeper so you can see

0:16

first-hand how businesses are using AI to solve your business challenges.

0:20

We're going deep into the use cases and getting

0:22

live demos of the latest and greatest in AI technology.

0:25

Today, I'm so excited to be joined by Jim Payne,

0:28

director of product marketing at Dialpad.

0:30

Jim, welcome to the show.

0:31

>> Yeah, thanks for having me.

0:33

>> First of all, I would love to hear a little bit more about who is Dialpad,

0:36

what do you guys do and who are you helping on the market?

0:39

>> Yeah, absolutely. Dialpad is a technology company focused

0:43

on artificial intelligence and communications.

0:45

That could mean anything from our ordinary knowledge worker who's

0:48

just doing everyday collaboration as well as helping the contact center as a

0:52

big one.

0:52

If you work in customer service,

0:54

there are a lot of really great artificial intelligence applications there,

0:57

that we're working on as well as revenue intelligence for sales folks,

1:01

a lot of amazing tools there as well.

1:03

Helping all of those personas in a very cool way.

1:06

>> Awesome. I would love to jump into the demo and see how your AI

1:10

functionality actually works live.

1:13

>> Yeah, absolutely.

1:15

We'll go ahead and just start here.

1:16

This is the Dialpad application.

1:18

This is home base for everything.

1:20

You'll see a lot of things that might seem ordinary on the surface.

1:23

Make a phone call, send a message,

1:26

start a video meeting, something like that.

1:28

Now, in any other world, these are just mere tools,

1:31

but for us, they become data inputs because when you're using Dialpad

1:35

for these different things, for communications, for collaboration, video, etc.

1:39

We're actually able to derive a lot of good insights,

1:42

provide automation in different places,

1:43

provide a lot of assistance as well.

1:45

I'll show you a few things, how that works.

1:47

I'll go ahead and make a phone call directly into this

1:50

and make sure it works just fine.

1:52

We'll show you a few things.

1:53

It might look a little bit odd just because I'm calling myself,

1:59

which is a bit strange, but we see a call coming in right out of the gates.

2:05

Now, we see a transcription engine, FHIRA.

2:08

The transcription is going to look strange just because, again,

2:10

I'm talking to myself, but this is where a lot of that goodness starts.

2:15

It is just in this transcription.

2:17

So let's say we have a customer who's calling in

2:19

and I'll go ahead and ask a couple of questions here.

2:21

You'll see some points of assistance that are very, very powerful.

2:24

So let's say I'd like to learn a little bit more about the pricing

2:28

on your support products.

2:29

So the customer asks that right away, we see this really cool assist FHIR,

2:36

because the artificial intelligence is listening,

2:38

and it's going to pull from unstructured data,

2:40

or you can program in all these really amazing real-time assist cards,

2:44

things like that.

2:45

So as an agent or as a salesperson,

2:47

those answers are always going to be surfaced to me right out of the gates.

2:51

No matter what, which is really cool,

2:53

because I've been a call center agent before,

2:54

I've worked in sales before.

2:56

Oftentimes, maybe a customer asks you a question,

2:59

you don't know the answer.

3:00

Your only recourse is really to stand up,

3:02

ask the people around you,

3:04

maybe furiously search through a knowledge base

3:06

that you have access to and hope you can find it or something like that,

3:09

or if you're at sales, you might be asking the people around you

3:11

to see if you can find an SE or something like that.

3:14

It's not very effective, and we know that times,

3:15

those deals, things like that, it's just not a great deal.

3:18

So the artificial intelligence is going to surface that information

3:21

to you right when you need it, so you can say it in real time,

3:24

whether it's pricing information, information,

3:28

you name it.

3:29

So a really, really powerful tool in that sense.

3:32

It's also going to do things like give you competitive intelligence.

3:35

Like for example, a customer might say,

3:36

"Hi, I'm also looking at GONG.

3:39

What are the differences between dial pad and GONG?"

3:41

Okay, that might be one.

3:45

And now look, oh, automatically,

3:46

the artificial intelligence says,

3:47

"Hey, I'm going to search your knowledge base for a GONG battle card,

3:50

and I'm going to tell you the differences right now."

3:52

So now I can speak about it intelligently,

3:54

as opposed to going to talk to someone in marketing or competitive intelligence

3:59

and finding out what the key differences are, that kind of thing.

4:02

It's a much more powerful way to do these things.

4:05

So again, this is all powered by the artificial intelligence.

4:09

This is in production today.

4:10

It gives you some great talk tracks in that sense.

4:12

It also puts a human in the middle, which we really like.

4:15

So if I'm a sales rep, I can say, "Hey, this was helpful.

4:18

This was not going to help train the models."

4:20

It's also going to put that AI parenting concept that we have in place.

4:24

That becomes very important,

4:26

because then whoever's managing your knowledge base

4:28

is going to get automatically flagged to say,

4:30

"This is not helpful. This is helpful.

4:33

Here are some things you might want to change."

4:35

It's also going to automatically identify redundancies, things like that,

4:39

so that way your knowledge base is always comprehensive and update.

4:42

Yeah, this is incredible.

4:44

Just right off the bat.

4:45

I think we hear a lot about AI being able to help us scale,

4:48

or I hate this term, but do more with less.

4:51

And that was kind of the narrative I feel like when it first,

4:53

you know, had this explosion last year of AI.

4:56

And it's always really cool to see it come to fruition in products

4:59

and how that looks like in practice.

5:01

And I think this is such a key, important use case,

5:03

which is being able to answer questions on the fly.

5:06

To your point of like, I think about our sales team,

5:08

as something comes up in a sales call that they really don't.

5:10

What are you going to do during mid-call?

5:12

Are you going to like frantically slack people

5:13

and try to ask questions?

5:14

It just creates a not great buying experience for your prospect.

5:18

And being able to pick that up and suggest answers is just,

5:21

I feel like one of my favorite use cases of AI and helping teams scale.

5:24

So this is really cool.

5:26

Thank you.

5:27

Yeah, it's very practical.

5:29

And you know, even outside of the real-time things,

5:31

you see stuff like, oh, I don't know, you have people walk off the job

5:35

or someone retires and maybe they've worked there a long time

5:37

and they have all this institutional knowledge in their minds.

5:40

How do you get that from them?

5:41

And how do you make it accessible to another person?

5:44

This kind of negates that, right?

5:46

It also makes training a lot easier,

5:48

whereas opposed to, oh, I don't know, trying to make sure every person

5:51

remembers every fact,

5:52

every piece of information, they don't have too many more.

5:55

Because the artificial intelligence is going to actually take that out of their

5:58

control.

5:58

So now they don't need to remember that you just focus on talking.

6:01

You teach them how to use the AI.

6:02

It's onboarding easier.

6:04

I mean, it just makes everything easier when you create that knowledge center

6:07

and then just use AI to actually service it to people.

6:09

So then outside of that, let's even look at another thing.

6:13

Let's say security, that's always going to be front and center

6:17

for just about everything and everybody.

6:19

Let's go ahead and see what happens when I talk about my credit card

6:22

information here.

6:23

So my credit card number is 1111-2222-3334-4444.

6:33

So we got the credit card number.

6:38

Now automatically, we see that the artificial intelligence is

6:40

redacting that information right out of the case.

6:42

So it's a front level of security, which is really nice to actually keep people

6:47

's private

6:47

information private.

6:49

It helps mitigate fraud, things like that.

6:51

It also makes things more efficient on the model training side that we have,

6:54

because it's already gone through a level of scrubbing automatically.

6:57

And then it's going to go through manuredity later.

6:59

So we see that as well, which is very, very powerful.

7:02

That's incredible.

7:03

So you as the end user, if you went in and looked at this transcript,

7:06

that's already redacted.

7:07

You don't have to worry about that.

7:09

Yeah, that's an incredible safety feature.

7:11

Yeah, a really nice safety feature, because you have admins,

7:13

you have a lot of people touching transcripts, things like that.

7:16

We don't want that information in there.

7:17

PII is PII for a reason.

7:20

So we've got to get it out.

7:21

So we get it out automatically so that we show up in the transcript, which is

7:24

nice.

7:25

So there's some of the live assistant features that we wanted to show you.

7:29

Let me go ahead and exit out of that call.

7:31

We can disposition the call, which is very, very straightforward.

7:34

You know, we'll go ahead and just put nothing in there and complete it.

7:38

Yeah.

7:39

Now, I want to show you some of the back end analytics.

7:42

OK, so a lot of this stuff is only as good as the back end analytics.

7:45

We can we can do a lot of cool things.

7:47

But the whole goal for artificial intelligence is really that it creates more

7:50

of a circle of

7:51

improvement as opposed to just a point solution for us.

7:55

So if we go even into the analytics, one of the cool things that we're doing

7:58

from this

7:59

is automatically inferring sentiment.

8:00

And you can see that right here.

8:01

A lot of companies are doing customer satisfaction scoring, but they do it

8:06

through manual surveys,

8:07

things like that.

8:07

Let me ask you, Sarah, do you ever stick around for two to three minutes to do

8:12

those customer

8:12

satisfaction surveys after you're on the call with your internet provider or

8:16

someone like that?

8:17

Can confidence say I've never done it one time?

8:18

Yes, exactly.

8:20

Nobody does, right?

8:21

So now all of a sudden, if you're working in customer experience or something

8:24

like that,

8:24

you're flying blind.

8:25

You don't even know where to train people.

8:27

Maybe your customers are happy.

8:28

You may not even know it.

8:29

Or you're doing a good job in certain areas and you don't know what those best

8:33

practices are.

8:33

So you're really just guessing.

8:35

So for us, because we're listening to every single interaction, we're able to

8:39

automatically

8:40

apply customer satisfaction source to every single thing.

8:43

So that's macro data, which is really nice.

8:45

You can say, here's how we're actually doing.

8:47

Then also you get all this interesting micro data.

8:50

So these are moments in time where you can say, hey, here is a moment of

8:53

negative sentiment

8:54

that the customer had.

8:54

Here is a moment of positive sentiment.

8:56

Here's where the call went wrong.

8:58

You can evaluate it over time.

8:59

Now you're able to take action as a whole.

9:02

So for example, you could just filter in the analytics and say, hey,

9:05

what call categories?

9:07

Because we're getting automatic call categorization from the artificial

9:10

intelligence,

9:11

which we see right there.

9:12

So we could say, oh, it's taking unstructured data and giving it structure like

9:15

, hey,

9:16

these calls are about billing, they're about support, etc, etc.

9:19

And now I can filter as an administrator and say, where are my worst customer

9:23

satisfaction

9:23

scores coming from?

9:24

Am I like, oh, it's all about billing.

9:27

So now I can work on our billing procedures as a business process.

9:30

I can also work on tailoring training to our reps to make sure that they know

9:33

exactly

9:34

how to handle billing issues, etc, etc.

9:37

You can also look at individual reps or agents and say, this agent is

9:42

struggling,

9:43

this agent is not here.

9:44

We just need to work with them in whatever way necessary.

9:47

So it gives you a lot of really good, rich data.

9:50

You also get these nice recaps and summaries and action items from calls,

9:54

another time saver for sales folks, or if you're working from the road,

9:58

things like that, you get all those recaps and action items automatically from

10:02

that transcript.

10:03

So this is all powered by Dialpad GPT.

10:05

That is our bespoke large language model that we've launched a while back.

10:08

It works very, very well to provide these things, which is very nice.

10:12

So you get all of that from a single call.

10:14

Yeah, the sentiment score, I think, is so interesting because I do,

10:17

you know, I'm obviously on the marketing side, but from a customer success side

10:22

I know it can be really hard to get that feedback, especially when a lot of the

10:25

times when you're

10:26

asking for feedback, it's the customers who are happy.

10:29

And so it kind of creates this echo chamber.

10:30

We've heard that's a problem.

10:31

Like we know we're doing really well and we get that echo chamber as positive,

10:35

but getting those, finding your gaps and finding out where there might be some

10:38

issues

10:39

and being able to pinpoint those, be able to pull that just out of transcripts

10:42

and having AI

10:42

help you figure out is it a specific area or to your point is it a specific

10:46

individual that's

10:48

really struggling.

10:49

I can see where this would be really, I'm thinking of our customer success team

10:52

and I'm seeing with this have a ton of value.

10:54

Yeah, for customers success for sales, it helps you forecast, you know, you

10:58

could,

10:59

we can use the same sentiment in that sales use case where we can evaluate

11:03

likely the buy,

11:04

tell if you forecast sales, you can also have all these interesting moments in

11:08

time where you

11:08

look at a seller who's underperforming and you can look at the same thing and

11:12

just say, oh,

11:13

you know, Sarah's not closing any business.

11:15

Let's take a look.

11:15

Let's take a look at like what she's saying, what she's doing, you know, is she

11:20

screwing up?

11:20

Is she not? Is it just bad luck?

11:22

Does she just need, maybe she just doesn't know about the materials we have or

11:25

the sales

11:26

tool or things like that.

11:26

So it helps you answer it.

11:28

It's like you're solving the right problem now as opposed to sometimes you don

11:32

't know what the

11:32

right problem is and AI helps you know what that is.

11:35

And now you can go solve it in a more actionable way.

11:38

We also see this for QA, you know, when I was a, when I was a call center agent

11:43

we had scheduled QA days, which is like a scorecard where your supervisor sits

11:46

behind you.

11:47

They listen to all your calls.

11:48

Super manual, super terrible.

11:51

Some companies handle this, especially when you're in a headlight regulated

11:54

industry,

11:55

where you need to follow certain QA criteria to be in compliance with some

12:00

government

12:00

regulation or something like that.

12:01

Very, very difficult.

12:03

They have supervisors.

12:04

We see customers have teams of five, six people who their whole job is just to

12:08

listen to call

12:08

recordings.

12:09

And they just listen to them on, you know, two or three X speed and they just

12:12

check

12:12

all boxes just to make sure that everything is correct.

12:14

We take that all out of your hands.

12:15

It's so easy to configure.

12:16

You can just say, hey, here are the things we need to make sure are happening.

12:19

We need to make sure that our agents or our sellers are saying the right things

12:24

perfect.

12:25

And the AI is going to go ahead and evaluate every single call against those

12:28

criteria that you set

12:29

so you can evaluate conformity, things like that.

12:32

So a very, very powerful tool in that sense.

12:34

I mean, it's just going to automate and automate and then give you better

12:37

insights.

12:38

The better insights you have, the better training you have,

12:40

people get better and more analytics, more insights,

12:43

and you just continue in this sort of circle and everybody gets better.

12:47

The company gets better.

12:48

I really do believe that companies who aren't adopting these kinds of practices

12:51

with AI

12:52

are going to fall off.

12:54

And we see that even historically, right?

12:55

How many bowling alleys have you been to recently, Sarah, that have manual pin

12:59

setters?

13:01

You know, that was the origins of bowling.

13:03

It was a person sitting there.

13:04

You knocked the pins over and then they had some people and they would come all

13:07

up for you.

13:07

When was the last time you saw one of those?

13:09

Never.

13:09

Ever.

13:11

They fall off.

13:12

You know, we even remember back when e-commerce was coming around.

13:15

The way I did that, I had to order something if I couldn't find it in the store

13:18

as I had to call.

13:18

And there was a person who answered a phone and I had to give them an

13:21

information,

13:21

you know, what I wanted, part numbers, and then someone had to go manually

13:24

fulfill it, right?

13:25

To fill out a card, mail it in, you know, to a fulfillment center.

13:29

That doesn't exist anymore because anybody who thought that e-commerce was just

13:33

like,

13:33

you know, that'll do that.

13:35

There's just a flash in the pan, right?

13:37

Like those companies are gone now, you know, because everything is e-commerce.

13:40

This is the future and companies who operate this way will set themselves apart

13:44

And I believe that in my DNA.

13:46

I cannot agree more.

13:48

And another thing I want to call it here that is not specific to the AI

13:51

functionality,

13:51

but just to your guys' UX is I do, I really love it.

13:54

I think anytime I see these AI-powered products where just even that scorecard

13:59

on the right-hand

14:00

side, how easy it is to digest that information, the quick 100% complete, 50%

14:03

complete,

14:04

you've got the logo there that tells you this is helping you from DialPads AI.

14:08

That's always just such a useful part of having AI in products now is something

14:12

I'm hoping we'll see more and more products do is really think about like how

14:15

the UX looks

14:16

and how they're surfacing AI and telling end users that they're benefiting it

14:20

from

14:21

inner products. So anyways, this really stood out to me in DialPads product

14:23

that I just really

14:23

love how it's set up and designed.

14:25

Thank you so much. Yeah, we pride ourselves on that.

14:28

The accessibility is so important. And the fact that it's one solution is

14:32

really nice.

14:33

It's not a bunch of point solutions cobbled together.

14:35

And it's all working in real time, which is a big deal as well.

14:37

You'll see a lot of point solutions going to tack on to your telephony or your

14:42

collaboration

14:42

or have you and they'll give you some decent insights on the back end.

14:45

Because this is an entire stack that's fully owned by us,

14:49

we get to actually listen to every single call path as opposed to just a blob

14:52

of sound

14:53

and do all the identification on the back end.

14:55

And now we get to provide all that stuff in real time, which is really, really,

14:59

you get all the same analytics, which are good, but then the real time

15:01

assistance too.

15:02

Yeah, that's amazing. Cool. Well, that's everything I wanted to show.

15:07

It's all very straightforward. These are some of the more practical ways where

15:09

you

15:10

guys using AI today. I mean, we could show different stuff all day long.

15:14

And I would invite everybody to see more just reach out to us.

15:17

We can do a detailed custom demonstration as well.

15:19

Absolutely. Jim, thank you so much for taking us through that demo.

15:22

If you're ready, I would love to transition into our lightning round Q&A

15:25

and ask you a couple questions.

15:26

Of course. Yeah, go for it.

15:28

So the first one is how long has Dialpad been building AI into your product?

15:31

Yeah, this is, I mean, it's not new for us. I remember even when OpenAI and

15:38

chat GBT started to have their moment, our CTO kept saying,

15:42

we're going to see about 100 new AI companies pop up,

15:45

nowhere. And sure enough, that was the case.

15:47

We've been doing this since 2016, so that was when we really started doing the

15:53

real time

15:53

transcription, which is where so much of this is coming from.

15:56

So it's been a long time. I see you're building in other things like semantic

16:01

search with AI,

16:02

the automatic speech recognition with AI.

16:05

The most recent big infrastructure piece was Dialpad GBT, which we talked about

16:10

which has really just helped you scale in a more economical way,

16:12

because it's a bespoke LOM designed for business communications.

16:16

You know, as opposed to being trained on the open internet, it's just trained

16:20

to do this stuff.

16:21

It's not going to give you a recipe for Apple Pie, you know, because that's not

16:24

what it's for.

16:25

Yeah. But it's going to do all this in a very accurate way.

16:27

That's also very cost-effective and scalable.

16:30

Amazing. And is what you show today, is that available for your customers right

16:33

now?

16:33

It is. Everything I showed today is generally available.

16:36

You can access it. It's all there.

16:39

Amazing. And speaking of customers, who are some of the customers that are

16:42

benefiting from Dialpad AI?

16:44

Yeah. Oh, so many. I think more than 98%. I believe our customers are using AI

16:50

in some form

16:51

or another with Dialpad. We have lots of large customers, big and small.

16:55

The LA Chargers are a big one. They're a heavy user of artificial intelligence

17:01

inside

17:01

Dialpad, which is really, really cool. T-Mobile is a great partner of ours.

17:06

They resell.

17:07

They actually resell Dialpad, which is really nice. So all of their customers

17:12

are getting a lot

17:13

of really great AI things as well. Car gurus is another one that kind of comes

17:18

to mind.

17:18

They obviously have a very large customer service team that comes with

17:23

everything so driven through

17:25

our website and obviously buying a car is kind of a different experience with

17:28

them and more innovative.

17:29

So they've really adopted that innovative mindset to say, "Hey, how can we be

17:32

efficient with this?

17:33

How can we use artificial intelligence to dialpad to be more effective?"

17:36

Those are some very fun logos. And then the last question is, what is next on

17:40

your AI roadmap at

17:41

Dialpad? Yeah, so many cool things coming out next. We're working on a lot of

17:45

polished. I mean,

17:46

I really believe we've only scratched the surface of what we can do here. You

17:50

know, you're talking about

17:51

cooler things like right capability into CRM. So it's like, oh, all of a sudden

17:56

, note taking for

17:58

sales reps become completely irrelevant, which is really neat. We've got some

18:01

next-gen chat

18:02

bot stuff coming out, which is really exciting too. I've had so many bad chat

18:06

bot experiences,

18:08

you know, where you're like, all it does is just searches the help center and

18:11

it's like,

18:12

"I don't know, talk to a person." And you're like, "Oh, come on." And then you

18:14

got to talk to a person

18:15

and give them all that context again. Because we're working across every

18:18

channel, we're already

18:19

going to preserve all that context. And we have some awesome digital self-

18:22

service products and

18:23

chatbots available. We're going to start folding in generative AI into those,

18:27

which is going to be

18:28

really, really neat and exciting. There's some next-gen agent assist stuff

18:32

coming out too,

18:33

which I'm excited about. You saw the first generation here. The next gen is

18:37

going to be even better

18:38

where it'll start actually writing talk tracks for you as opposed to just

18:42

finding information

18:43

and giving it to you, which is going to be really neat also. I mean, the world

18:48

is always

18:48

in a sense. So we've got a lot of cool stuff coming out shortly and even more

18:54

in the near future in

18:54

the distant future rather. Amazing. I can't wait to see what Daupah does next

18:58

because even the demo

18:58

that you showed today, I thought was really incredible. So excited to see where

19:01

you guys take

19:02

this in the future. But, Jim, thank you so much for joining us on the show

19:04

today. I really appreciate

19:05

you taking the time and showing us this incredible demo. So thank you so much.

19:10

Thank you.

19:10

[Music]