How this CEO turned 25,000 hours of sales calls into a self-learning go-to-market engine

1. 0:00 – 2:36

   Introduction to Matt Britton

   1. MBMatt Britton0:00

      With my company, my sales team was consistently telling me that they just didn't know how to find anything. They didn't know how to find what customers were interested in.

   2. CVClaire Vo0:08

      You had a bunch of salespeople. They said, "I need more information to serve our customers better." You realized you had twenty-five thousand hours [chuckles] or something of recorded customer calls, which are the perfect source of truth for how customers wanna be interacted with. You're gonna show us a Zap now that takes a single recording and does a bunch of stuff.

   3. MBMatt Britton0:27

      So basically, I needed to figure out, how can I create a feed for Zapier so it knew the call ID of each new call as it occurred? So the first step is essentially a trigger where a new call comes in. It'll basically scrape the information from Gong, and one of the things Gong will give you is that call ID. So that appended to the URL essentially is all I needed to give Browse to be able to go to that URL and be able to essentially scrape the entire transcript. It wasn't connected. I had to kind of hack it together.

   4. CVClaire Vo1:01

      I love a CEO that knows how to build it. I love a CEO who knows that no problem is not solvable. [upbeat music] Welcome back to How I AI. I'm Claire Vo, product leader and AI obsessive, here on a mission to help you build better with these new tools. Today we have Matt Britton, CEO of Suzy. Now, normally we show two or three workflows, but today Matt's gonna show off the one mega workflow that rules it all at his company. He's gonna show you how to take a single asset and turn it into tons of go-to-market goodness, from emails to customers, enrich data sources, and even marketing assets that can be used to source more prospects that are gonna be successful with your product. Let's get to it. This episode is brought to you by Brex. If you're listening to this show, you already know AI is changing how we work in real, practical ways. Brex is bringing that same power to finance. Brex is the intelligent finance platform built for founders. With autonomous agents running in the background, your finance stack basically runs itself. Cards are issued, expenses are filed, and fraud is stopped in real time without you having to think about it. Add Brex's banking solution with a high-yield treasury account, and you've got a system that helps you spend smarter, move faster, and scale with confidence. One in three startups in the US already runs on Brex. You can, too, at brex.com/howiai.

2. 2:36 – 4:17

   Why Zapier became the backbone of Matt’s AI automations

   1. CVClaire Vo2:36

      Matt, thanks for coming on How I AI. I'm excited because, as I was saying before we started the show, we get vibe coders left and right, and I know we're gonna talk about some custom software that you built, but we just do not get enough on the go-to-market and marketing side of AI automation. So I'm really excited to show what you have to share. So really appreciate you joining today.

   2. MBMatt Britton3:00

      Excited to be here.

   3. CVClaire Vo3:01

      So you and I both really love Zapier, and I have to ask, even before the age of AI, was this a tool that you relied on? Why has that-- this specific software become kind of the backbone of so many of your AI-based automations?

   4. MBMatt Britton3:17

      So I've always been fairly technical, but I've never been a coder. Um, I sold the first ads ever on Facebook directly to Mark Zuckerberg and Eduardo Saverin in two thousand and five. Um, I bought some of the first Google keywords ever to exist, right, when I started my business, um, in two thousand and two, my first ad agency. Um, so I've al-always loved sort of ad tech and getting like... understanding how these tools work, but at the same time, never been an engineer. And as I've wanted to get more sophisticated in the, in the tools and solutions I've built for various companies that I've run, I've needed to not just use one tool, like AdWords, but multiple tools to stitch things together to be more efficient. And I was turned on to Zapier, like most other people, just through a Google search, and I think I wanted to connect, um, you know, leads that were coming in through my website to some type of flow, where it automatically emailed the person who signed up, and then I just kind of started to dive into it. But to your point, Claire, it wasn't until Zapier integrated AI when kind of my mind just became blown in terms of what's possible.

3. 4:17 – 9:02

   Identifying your core business problem

   1. CVClaire Vo4:17

      Okay, so you're gonna show us how you take a single asset, and I won't spoil what it is, and turn it into basically a full suite of activities across your marketing, sales, internal company work. So why don't you pull that up and show us what you built?

   2. MBMatt Britton4:32

      Before I pull it up, I guess I should say that I think it's all about figuring out what problem that you wanna solve. And I think with AI in general, people get so overwhelmed with just the amount of tools and the rate of change, that they just find themselves kind of playing around with all these tools, trying to get to the point where they feel like they're comfortable in understanding them, but at the same time, they're not really moving their business forward. And I think the reason that's the case is people don't ever take a step back and think, like, "What is the core problem that I need to solve for my business? Like, what's holding me back from growing faster than I want to? What, what's getting in my way, or what's an opportunity I know is there, but, you know, I'm not, I'm not able to take advantage of it?" And with my company, what I was hearing over and over again was, my sales team was consistently telling me that they just didn't know how to find anything. They didn't know how to find what customers were interested in. They didn't know how to find, um, how to speak to people of a certain industry or a certain title in terms of identifying use cases, so just so many unknowns. And so once I understood and put my finger on that, o-on that problem, I just became very sort of tunnel-visioned, and I was determined to figure out how I can build solutions that can aid my sales and customer success team to be more in the know.... So once you've actually identified the problem, the next step is figuring out what data can help you seize that opportunity. And in the case of, you know, understanding our customers and, and, and getting that information, it just so happens that since the pandemic, when our company went remote, we've been using this tool called Gong that's essentially attached to Zoom calls that records every single call that we have. So it says, "This call is being recorded for quality assurance purposes." And I always knew we had, obviously, Zoom, and I knew that we had Gong, but what I didn't know is that their transcripts were amazing, and that we actually had 25,000 hours of call transcripts that had been amassed over the last five years. And if you think about understanding information about your customers and your business, there's no better source of truth. So we have since built an entire operating system around this information, not just the historical information, but a variety of different workflows that happen with each new call that occurs. Because it's not just about understanding what's happened in the past, but it's also being able to be highly responsive to what's going on in the present. So the first thing I'm gonna show today is an automation that we have created based upon calls our t- our, our teams have, either our sales team or our customer success team. Um, so a- and essentially what happens is, as soon as that call is over, a series of events happen with that individual transcript. We also do things sort of at large with the aggregate transcripts, if that makes sense. But right now, I'm gonna show you what happens kind of like real time once a call is completed.

   3. CVClaire Vo7:26

      Great.

   4. MBMatt Britton7:26

      So-

   5. CVClaire Vo7:26

      So while you pull that up-

   6. MBMatt Britton7:27

      Yes

   7. CVClaire Vo7:27

      ... just to recap for our listeners, you had a bunch of salespeople. They said, "I don't know how to find the information that I need. I don't know how to generate the information I need. I need more information to serve our customers better." You realized you had, and I'm-- correct me if I'm wrong, 25,000 hours [chuckles] or something of-

   8. MBMatt Britton7:43

      That's correct

   9. CVClaire Vo7:44

      ... recorded customer calls, which are the perfect source of truth for-

   10. MBMatt Britton7:48

       Yeah

   11. CVClaire Vo7:49

       ... how customers wanna be interacted with, and you decided that was going to be the core context for a lot of these actions inside your company. And then you're gonna show us a Zap now that takes a single recording and does a bunch of stuff. I got a preview of this, and it does a lot of things.

   12. MBMatt Britton8:07

       I've tr- I tried to give AI to my engineering team to figure stuff like this out, and it just became overwhelming to them, even to integrate it in the product. And what's been helpful for me was first building things on my own, and I'm not technical enough to be able to build on top of our software product. So the tools like the one I'm gonna show you today was a great way for me to be able to dive into the power of AI because it was on the edges of the-- it wasn't the product. It was sort of on the edges of how we operate it. And through that, though, I became far more adept at AI, and now I'm very much involved in our product itself. So often, people struggle to find a way in, and there's lots of different ways in. One way is actually building something for yourself personally or building something for the marketing organization or somewhere else, and then through that process, you really start to get it, and then you can start to be more, um, you know, proficient in AI in, in much more substantive ways within the business.

4. 9:02 – 13:42

   How Matt built the initial trigger automation with Browse AI

   1. CVClaire Vo9:02

      Yeah, and I want all the other CEOs and executives watching this podcast to listen to exactly what you said because it is not sufficient to instruct your engineers to build-

   2. MBMatt Britton9:12

      No

   3. CVClaire Vo9:12

      ... AI.

   4. MBMatt Britton9:12

      You'll go nowhere. Yep.

   5. CVClaire Vo9:13

      No, you'll go nowhere, and I've said this a lot. This is a moment for actual hard skill building, um, in leaders, which is you actually have accessible skills to build in using AI, building AI tools, using these sort of like no-code versions of tools to really upskill yourself on the capability, and that's gonna make you a much more relevant leader, much more frugal-

   6. MBMatt Britton9:35

      Yeah, you're opening up the hood.

   7. CVClaire Vo9:36

      Yeah.

   8. MBMatt Britton9:36

      It's like, you think about it, if you bring your car in and you don't know anything about fixing a car, and they tell you it's $4,000 to fix a transmission-

   9. CVClaire Vo9:44

      [chuckles]

   10. MBMatt Britton9:44

       ... you're gonna say, "Okay," because you need your transmission fixed, right? But if you actually just open up the hood and you understand how transmi- transmission works-

   11. CVClaire Vo9:51

       Yep

   12. MBMatt Britton9:51

       ... even if you're not a mechanic, maybe f- you can say, "Well, it really shouldn't cost $4,000 to fix this."

   13. CVClaire Vo9:56

       Yeah.

   14. MBMatt Britton9:56

       "It really costs closer to $2,000," and I think that's sort of the same analogy when it comes to AI. So, um, so I-- so the f- the, the first step is building what I call a trigger automation, and this trigger automation essentially comes from a tool, uh, that we've created called, um, that we use called Browse.ai. So this is Browse.ai, and essentially what Browse.ai does is it runs like a, a script where essentially scrapes information from Gong calls. So what you see here is a URL string. You need a URL string in order to identify a, a call transcript as it comes in, and Gong didn't have an easy way to do this. So I basically started to bring up a bunch of, uh, call transcripts, and what I started to see is they all kind of start the same way, and they just end it with this call ID. So the only thing different from call to call was this call ID. So basically, I needed to figure out, how can I create a feed for Zapier so it knew the call ID of each new call as, as it kind of occurred? So the first step is essentially a trigger where a new call comes in, right? And then what happens is, when the new call comes in, what it will do is it'll basically scrape the information from Gong, and one of the things Gong will give you is that call ID. So I'm able to actually see the call ID. So if I click here, um, and I scroll over, you'll actually see that there's a call ID that I can identify here, um, which is right here. And so that appended to the URL essentially is all I needed to give Browse to be able to go to that URL and essentially scrape the call transcript. So it, it wasn't connected. I had to kind of hack it together. So if you'll see here, it basically knows what to run just based upon what's brought in.... and then it will go to this page, which, which I will show you here, which actually is where the transcript is, and it's able to essentially scrape the entire transcript. So this is the raw transcript that's coming from the Gong calls by Browse AI going to that Gong page and just getting this information. But I had to initiate it, so that first step essentially initiates the scrape, and then when the scrape is completed, it starts my next automation.

   15. CVClaire Vo12:19

       Yeah, and so just to call this out for folks that are trying to build their own thing, it's okay if your tool itself does not expose the data you want. In this age now, you can usually use another tool or an alternative-

   16. MBMatt Britton12:32

       There's always a way. Yeah

   17. CVClaire Vo12:33

       ... or an LLM to really pull the data you need out of, out of any system.

   18. MBMatt Britton12:38

       Yeah, and I could have given up, Claire, like, at that point. At that probably one step took me the longest, and if I never would've gotten past that step, I ne- And I think a lot of people would probably have given up at that step. But after I got over that hurdle, then everything else became so much easier, and it's really like an analogy for life, like, building something like this. And there are other stumbles I've had along the way in building things, but you just have to know that there's a way, and, and using... And because just 'cause a tool doesn't do it, doesn't mean it can't be done. And this, like, in, in the rear view mirror, seems obvious, and now if I had a similar challenge, I'd be able to do it right away. 'Cause what will happen is, every time you solve a problem such as that, the next time you need to build something, you'll have all these sort of ideas and, like, hacks, um, in your tool chest, so to speak. And then now I'm at a point where there's, like, nothing you can tell me to build that I wouldn't know how to build because I just know how all these little things can be solved for. And you learn coding along the way. Like, along the way, you learn what JSON means and all these things by having your hands on it a- and, and creating the automations.

5. 13:42 – 14:00

   The value of CEOs getting hands-on with building

   1. CVClaire Vo13:42

      A- a hundred percent. And what I was gonna say is, this is a CEO that I love. I love a CEO that knows how to build it. I love a CEO who knows that, like, no problem is not solvable. Um, and I think just even getting hands-on with some of these no-code tools and these AI tools just gives you a little bit more context to be bolder

6. 14:00 – 20:14

   Scraping and processing call transcripts

   1. CVClaire Vo14:00

      about what you build. Okay, so you have-

   2. MBMatt Britton14:01

      That's right

   3. CVClaire Vo14:02

      ... a transcript.

   4. MBMatt Britton14:02

      So the task is done, right? The, the, the call's done, and so this next trigger is triggered when Browse AI successfully scrapes a call transcript. And the first thing it'll do is obviously it'll trigger it, and you'll see here it'll give me the entire transcript of the call. Um, and that's basically... N- now it's like, okay, now I'm in business, right? Now I have everything I need. And there's a bunch of other stuff in that Gong call transcript that I use to do database lookups throughout that we'll kind of get into. I'll, I'll have a delay of about two minutes before pulling in data like this, just because I wanna make sure that all the data is brought in, the scrape is done, and I'm just... You- you're prone to errors, especially if you're running a lot of, um, tasks quickly, if you don't put in a delay. So always, like, a one or two-minute delay is a buffer just to let the system catch up so it doesn't break. So, so that, that's kind of self-explanatory. The next thing I do is I run a formatter, where I'm basically removing all the HTML from the transcript. So when you scrape sometimes it'll pull in the HTML, and I don't want that. I just actually want the actual text. So I run a formatter step, where I'm removing all that. Um, I'm pulling out anything I need to that might confuse, uh, the analysis, so I'm just essentially getting the raw text. Then what I do is I start to enrich the data with other information besides just the Gong transcript. 'Cause I had the Gong transcript, but one of the things I knew I wanted to build was, after the call is done, I wanted to be able to tell the salesperson that was on that call what transpired. I wanted to make it easy for them to write a follow-up email. I wanted to be able to identify who their supervisor was, right? But that wasn't directly pulled in through, through Gong. However, we have other data sources that essentially can connect that information. So we have a Google sheet here. For example, if you look up this ID, it connects the ID to the brand and the brand to the user, which is a whole separate workflow that we created, so it can kind of connect the dots. 'Cause when you're running an automation, you're not always gonna get the data from the trigger. Sometimes you have to round it out, and the way you round it out is using things like lookups on Google Sheets. So you're pulling everything in. It's almost like you're going down a path, you're on a hiking trail, and you wanna be able to pull the supplies you need along the way before you get to the destination. And when I started, I had a backpack, but the backpack didn't have water in it, and now I have water, right? 'Cause I grabbed it from here. And you're kind of going along a journey, and I personally w- one reason why I love Zapier versus other tools is the way my mind thinks is in a very sequential framework, where there's other platforms like N8N, you know, or Botpress, where it's basically like, looks like almost like an octopus, how it's branching out. I just have a hard time thinking that way. Now, over time, I've had to, because I'm basically describing the difference between automation and agents. 'Cause agents are not deterministic. Agents have different ways, and my brain has struggled with understanding agents, and I'm finally getting there. But basically, the progression I see people having to take in AI is you start with using AI as a tool. You know, "ChatGPT, give me a recipe for lasagna." Then it's, okay, automations, which we're talking about now, and then you get into the world of agents, where it's not just always going from step one to two to three. It might go from step one to three to eight, based on what you're trying to accomplish.

   5. CVClaire Vo17:16

      Well, one tip for you, or one tip for the listeners here I found, is we'll go through this whole thing, and a good exercise I found is taking an, a sequential, step-based automation and trying to use, for example, Zapier agents, and just describe that automation in natural-

   6. MBMatt Britton17:34

      Yeah

   7. CVClaire Vo17:34

      ... language and steps, and see how-

   8. MBMatt Britton17:36

      Yeah

   9. CVClaire Vo17:36

      ... close you can get. Even that replication across modalities can be a good way to, to test your exercise. So-

   10. MBMatt Britton17:41

       Yeah, just test it out, 100%.

   11. CVClaire Vo17:43

       Yep.

   12. MBMatt Britton17:43

       Yeah, and, and so looking up the information, so it's able to basically grab the information, and then after I feel like I've had all the information, the next thing I'm gonna do is this is where I'm starting to pull in the LLMs. And an important part here is, first and foremost, knowing what LLMs to use. And one thing I've had a hard time with is actually just keep-- We have so many automations now, and I think, uh, we can do a better job at the organizational design behind it. 'Cause what happens is, I built so many things, and I don't always do proper handoffs. So for example, here-... I- it should always say, "Use latest stable version," but it didn't, right? So, so I'm gonna change it now here live on the spot because I wanna be using the latest version. You also wanna make sure you're using the best model. I still think GPT-4 Turbo is probably a good model for this, but you could see in, in the platform in Zapier, there are multiple different versions that you can choose from based upon-- And obviously, they, they all eat up different amounts of coins, and, um, but it's pretty incredible in terms of all the models. Now, with GPT-5, it's supposed to be able to choose for you, but it's unclear to me how that works in the context of an API, and for some reason, still in Zapier, you're still able to choose. And, you know, I, I spend a lot of time testing. I'll go in the ChatGPT and test a sample input in a variety of different models to make sure-- It's like whatever's the best output the quickest is what I'll tend to use.

   13. CVClaire Vo19:03

       So you're still leaving it using classic GPT-4 Turbo, a good old classic, classic favorite.

   14. MBMatt Britton19:08

       Right.

   15. CVClaire Vo19:09

       AI is supposed to make work easier, but I've been there: weeks of setup, endless back and forth with engineering, and yet another tool the team never really adopts. That's why I use Zapier's AI orchestration platform. It connects with nearly eight thousand apps, so I can finally put AI to work without the drama, without the delays, and without pulling engineering in every time I wanna automate something. With Zapier, you can roll out AI-powered workflows that do real work across your whole company in days, not weeks. I use Zapier every single day. It automatically responds to leads with enriched, personalized data, it checks my calendar weekly and offers smarter ways to manage my time, and it even drafts emails for every new request that lands in my inbox. All of that running quietly in the background so I can focus on the work that matters. And Zapier's built for scale. With enterprise-grade security, compliance, and governance, it's trusted by teams at Dropbox, Airbnb, Opendoor, and thousands more. Go to try.zapier.com/howiai to learn more about how Zapier can bring the power of AI orchestration to your entire org.

7. 20:14 – 23:25

   Using LLMs to generate call summaries and sentiment scores

   1. CVClaire Vo20:14

      Let's talk a little bit about this, this prompt.

   2. MBMatt Britton20:16

      Yeah.

   3. CVClaire Vo20:16

      So tell me what first kind of summarization exercises you wanna do here.

   4. MBMatt Britton20:22

      Yeah, so basically here, and, and the reason I can use a model like GB-- like, like a GPT-4 is... And, and, you know, part of it, again, is just keeping up with continuing to update the models, but I don't fix things if they're not broken. So this particular Zap works perfectly for us, and it gives us everything we need, and we don't need more rigor in analysis here 'cause it's just some core things that we wanna identify. So I'd rather not spend the extra money and even go through it, but at a certain point, if it didn't work, I would look at models if it wasn't going fast enough. What's interesting is that older models tend to work faster and faster and faster over time, and they actually error out less, and, and sometimes the older models get updated as the new model is updated as well, so it's not like you're, you're driving an, an '84 Chevy, so to speak. So here, this is a key step. This is called Core Summary Generator, and what this is asked to do is analyze the customer success call transcript between Suzy and our client to gauge the health of customer relationships and identify improvement areas. Start summaries with the customer's company name, key participants, and then kind of going through it, ask for the key stakeholders, and then it gives a call overview. We should describe the call's purpose, the main topics discussed, and the outcome. Exclude small talk. And then I have a variety of different instructions: Assess the overall customer sentiment, noting any frustrations or concern; provide a sentiment score from one to ten, where ten reflects high satisfaction and one indicates potential discontinuation of our services. This is a key thing 'cause it allows us to quantify customer sentiment over time, and we actually benchmark this against actual churn. So-- And, and it's been highly predictive, um, in terms of the-- If you take the average sentiment score of customer calls over the past year, it's a huge predictor of if the customer is not just gonna churn or, or are they gonna upsell if they're really happy? Also, one great thing the customer successfully did on the call, kind of identify that, and what are some things that they actually could have d- done better? And then list the next steps. So this is basically just an overarching prompt where it'll take a, uh, a transcript, and it'll identify all this information for me, and then that content I can do a variety of different things with, but it's a huge part of the overall output.

   5. CVClaire Vo22:29

      So one of the things I wanna call out here as I was reading your prompt is, in an ideal world, all your best CSMs are doing this after every call in a perfect-

   6. MBMatt Britton22:40

      Exactly

   7. CVClaire Vo22:41

      ... way with great, you know, objective self-evaluation, all this kind of stuff. And the reality is, we're all so busy that-

   8. MBMatt Britton22:48

      Yeah

   9. CVClaire Vo22:48

      ... you know, your customer success folks or your sales folks are probably going meeting to meeting to meeting, and at the end of the day, trying to figure out their notes and put little things in. And I just think what's nice about this is you can make the customer success or account manager's job a lot easier and let them be exceptional at their job by automating some of the work that they, they would do. And so I think it's a really good hygiene step for anybody to think, you know, "After I'm coming out of a meeting, if I were to do my best job possible, what are the five things I would do coming out of each meeting?" And then just automate that for yourself, and then you know that every time you're gonna be doing

8. 23:25 – 26:17

   Creating a Slack channel for real-time call insights

   1. CVClaire Vo23:25

      that.

   2. MBMatt Britton23:26

      Yep. So there's a bunch of other steps I added, and I'm not gonna go through all of them because there's a ton of them, but basically, it, it looks up the user on Slack. So I understand-- the user meaning the, our, our employee on Slack. It identifies the people who aren't from our company so you can kind of exclude them. Um, it's able to find the user here. So I use Slack as a, as a lookup sometimes 'cause Sl- our company's entire directory's on Slack. So if I'm trying to get someone's email address in an automated fashion, and I have their name, I can actually use Slack as a lookup tool in Zap without even actually posting anything to Slack. So sometimes these tools actually can be used for other purposes that's not their core purpose. And then, basically, the fir-- one of the main things I do from this is I send a channel message. So basically, after the call is done, you can see new customer success caller as the account, the opportunity.... the leader of the call from our company, the date of the call, and it basically has that summary, um, that gets sent out. So we have a, we have a channel that are- that's a constant feed that I, obviously, the CEO, I'm very attuned to, and I'll, I'll share it right now, where basically every time a customer call is done, it just pops up on Slack. And I'm able to really... You know, we have 300 employees at our company, and I'm really able to get a sense of the kind of pulse of the company, what customers care about, just based upon looking at, at something like this. And it's, you know, that alone, if that was the only invention that came out of AI, it would be pretty incredible if you think about it. And this is just, like, one of many things that we do. So I'm gonna pull open Slack right now, and you can see here, this is a sample call, and it shows who the ka- uh, key stakeholders were, were, uh, what the call attempted to establish, what was the sentiment score, got an, an eight, right? Opportunities for upselling, feedback, and next steps. And it has basically a transcript here, and it's great for us to do. If a customer is not happy, right, if they, um, you know, for some reason, uh, score below a seven, we have a churn, uh, notification channel, um, where basically it's called Churn Early Warning System, where it'll tell us if a customer is not happy for whatever reason. And s- and, and we've had to modulate it, 'cause sometimes a client will say they're not ha- it, it'll say the client's not happy, but maybe they're just not happy with how their business is going. So it, it's, it's not always, like, a science. Um, and then in the channel, sometimes the rep will say, "Oh, no, they're fine, it's just this." But we have in many instances, and to your point earlier, like, sometimes the, the rep might not wanna tell anybody, right? Maybe it's a Friday afternoon, they just don't wanna deal with it, and then what happens is we end up forgetting about it, and then the customer churns three months later, and we're like: "Why didn't you just tell us?" We don't have to do that anymore. We don't have to ask somebody how that call went with Procter & Gamble. It's

9. 26:17 – 28:35

   Extracting keywords for Google Ads campaigns

   1. MBMatt Britton26:17

      just here.

   2. CVClaire Vo26:17

      Yep. Okay, great, so you take the transcript, you post all of them, so everybody in the company has access to customer calls and summaries, which is just great for general sentiment analysis, knowledge sharing, transparency. You take any ones where the sentiment analysis is low, and you put them in sort of like a warning area, churn alert channel, um, where I'm sure you're paying a little extra attention, so you can get ahead of potential churn risks, which as a B2B girl, I really, really love. And then, so that's, that's a little bit more like the account ops side of things-

   3. MBMatt Britton26:50

      Yeah

   4. CVClaire Vo26:50

      ... but then you take off a bunch of marketing-

   5. MBMatt Britton26:53

      Yeah, there's a bunch of other things.

   6. CVClaire Vo26:54

      Wow!

   7. MBMatt Britton26:54

      Yeah, so the, this next one, again, this is all part of the same automation, is another LLM call, where we're basically describing what Suzy does, and we're saying, "Analyze the key areas of interest stated in the transcript, and output a bunch of keywords that we should be buying in Google." So if customers are using words that we-- th- that are describing what they're interested in or what we sell, and we're not running Google keywords for them, we want to. So basically, these keywords will be mentioned, we extract them, and then we run an automation to add those keywords to our Google campaigns automatically.

   8. CVClaire Vo27:30

      So not only are you taking sort of... This is-- I love this one, so I want people to pay attention.

   9. MBMatt Britton27:35

      Yeah.

   10. CVClaire Vo27:35

       So not only are you taking the account-level specific, uh, context, but you're saying, "Our customers will tell us in their words what they're looking for, what problems they're trying to solve." These customer calls are a rich source of market insight, and so you're gonna use these customer calls to actually extract out market surface areas, keywords, places where you can put marketing dollars against, and then reach customers similar to the customers that you're successful with, which is a really nice closed-loop solution. Um, and again-

   11. MBMatt Britton28:07

       That, that's right

   12. CVClaire Vo28:08

       ... you know, we were talking about how this note summary is the, the way, in an ideal world, a customer success manager would provide notes. In an ideal organization, your, you know, paid search manager would be monitoring all these calls [chuckles] and doing all this for you.

   13. MBMatt Britton28:25

       Right.

   14. CVClaire Vo28:25

       But we don't live in ideal worlds, and people are busy.

   15. MBMatt Britton28:27

       No.

   16. CVClaire Vo28:27

       And so again, this is, is not only designing from the point of view of, like, what would a person do, but also, what would a team do?

   17. MBMatt Britton28:34

       That's

10. 28:35 – 29:48

    Building an AI coach for sales and customer success teams

    1. MBMatt Britton28:35

       right. The other thing we do is we've done a coach into this. So the next step essentially is called individual call feedback, and what this does is it actually creates a feedback note to the person on the call. So this just goes to the sales rep on the, on the, our sales, our customer success rep saying, "Here's what you did. Here's what you did right. Here's what you did wrong," and actually sends it to them right afterwards, so they understand how to get better, which is something that we would had to hire somebody to be on their back and tell them, which they know on their own. What's interesting is, like, the people that really wanna get better, this is-- AI is an incredible tool because they're gonna want this feedback. And the people who never really want to hear from anyone to begin with, they're not gonna wanna hear this, but y- they wouldn't have been good in either way. So that kind of goes to the point that, like, it's gonna make the good people that much better, right? And we add this to a data set. So we have a s- a feedback called data set, so we can actually see, are there trends? Like, is AI detecting that this person always cuts calls short, or they always interrupt the customer, or they don't talk about... And then when it comes time to reviewing them, it's all data-driven. It's not just myopic. If w- if managers change over, we have all this information, and the good ones want this information.

11. 29:48 – 35:25

    Creating a follow-up email writer for post-call communication

    1. CVClaire Vo29:48

       Yeah, what I was actually gonna reflect on is, you're talking about this from the point of view of the individual contributor, the CSM, the AE, but what I was thinking is, so much of AE and CSM performance is really gated on, do they have a good sales manager coach? Do they have a good SVP sales-

    2. MBMatt Britton30:04

       Yeah

    3. CVClaire Vo30:04

       ... that can actually provide them targeted coaching on all of their tools-

    4. MBMatt Britton30:07

       That's right

    5. CVClaire Vo30:07

       ... right when it's relevant.

    6. MBMatt Britton30:09

       Yeah.

    7. CVClaire Vo30:10

       And this sort of, like, evens the playing field. Your manager could be great, your manager could be terrible. In every call, [chuckles] you're gonna get-... kind of objective feedback on your performance, and so again, it helps up-level the performance across the organization.

    8. MBMatt Britton30:24

       Yeah, and it's democratized. You're right, 100%.

    9. CVClaire Vo30:26

       Yeah.

    10. MBMatt Britton30:26

        The other thing we realized, just going back to problem-solving, is we heard from our sales team and our customer team, "You know, it takes so much time for us to write a good follow-up email after the call." So now we added follow-up email writer, where essentially you write an email that they, we- that they would wanna send as a follow-up to the call, and actually just a- and, and designed very well, and it's sent to them for them to basically copy and paste it and send it. And it's just a way for them, so right after their call, they'll get the feedback in their inbox, and they'll get this email, and they can copy and paste and send or edit it. And, you know, we could have made this automated, but, you know, that's where the human in the loop matters, right? We don't-- What if there's, the context is wrong? What if they don't want to send the feedback right away, or if they want to copy somebody new? So that's why we have to have a human in the loop here. So the churn early warning detector basically sends through two different paths, and these paths essentially, um, kind of dictate who we should notify and who we shouldn't. So we've also now started to do much more marketing-driven things from this data, one of which is we start to create a database. This is called customer profile database, and what customer profile database does is it essentially structures data after each call with things like, what's the role of the customer? What product areas of Suzy are they most interested in? What business trends are they most interested in? And we have a structured database across all the calls, which gets fed into a GPT. So if a salesperson's going into a call with a brand manager of an automotive company, they can say: "What are the things that brand managers of automotive companies are most interested in, in terms of trends of interest in our product?" And it'll tell them right away, because the data in the aggregate is stored with a different tool that we deploy. So again, not only do we have the automated things that are happening, but we have this aggregate database that we unlock the value of on an ongoing basis.

    11. CVClaire Vo32:16

        Okay, I have to ask you a question. Again, as a B2B enterprise girl, are you using a CRM? Like, is this data going into Salesforce, or are you like-

    12. MBMatt Britton32:24

        Yes.

    13. CVClaire Vo32:25

        "Well, it can [chuckles] it can all go in Google Sheets, we don't care." I'm just curious-

    14. MBMatt Britton32:28

        Well, I mean, it-

    15. CVClaire Vo32:29

        Although from CRM

    16. MBMatt Britton32:29

        ... you know, it today it goes into Salesforce, but the q- you know, I think the reason Marc Benioff is leaning into Agentforce is for that reason, right? It's like, what's the point, right? So, like, theoretically, everything I'm building right now is a better v... What I just showed you, I believe, is a better version of Salesforce. And guess what? The salesperson doesn't have to enter a record. It's entered, and the manager's getting information, and they can chat with the data, and they can pull reports and aggregate. That's basically what Salesforce was built for, and, you know, a- from a meta standpoint, our company is facing the same thing with market research, where, like, we built this m- so we're all trying to figure out how to disrupt ourselves based upon what's happening. But you're right, I mean, that, and that's sort of the fundamental issue that exists today.

    17. CVClaire Vo33:11

        What I was reflecting on, though, is one of the challenges with Salesforce-- Well, you know, one of the reasons Salesforce did so well is because of the flexibility of implementing your own data schema and kind of editing-

    18. MBMatt Britton33:20

        Yeah, yeah, of course.

    19. CVClaire Vo33:21

        And one of the limitations is like, gosh, you have to go through your Salesforce admin to, like, set up anything and get it up-

    20. MBMatt Britton33:26

        Right

    21. CVClaire Vo33:27

        ... you know, and then, and then [crosstalk]

    22. MBMatt Britton33:28

        And all the charts and graphs weren't gray, and-

    23. CVClaire Vo33:30

        Yeah

    24. MBMatt Britton33:31

        ... no one really knew how to... I mean, you just sometimes wanna know, like, what's the status of the P&G account? It's what you wanna know, and it's just good luck getting that done. Well, right now, you could just literally just speak it or type it, and you get it, and that's kind of where we're all heading to.

    25. CVClaire Vo33:44

        Yeah, and then what you're showing is you could create these one-off, loosely structured Google Sheets, for example, for different various lookups. They don't have to be perfect. They don't have to be hardened in your CRM.

    26. MBMatt Britton33:55

        That's true.

    27. CVClaire Vo33:55

        But they're useful to your team, and I think that-

    28. MBMatt Britton33:57

        Well, it's structured. It's a structured database, which, and the, you know, I think, you know, for RAG, structured databases work much better, and this is a structured database, and that's really all you need. I think a key point here, it goes back to what I mentioned earlier, is you just have to find-- It, it's not about the tool, it's about the data. People are so focused on the application layer. It means nothing without the data. And to me, it's like th- this is the ultimate source of data, and this is the treasure trove, and this is people in the wild saying what they want. So I wanna build everything on top of this data. So that's why when, when we were prepping for today's interview, you're like, "We'll show a bunch of different things," and, and the way I look at it is differently. I'm gonna show you one thing that has many different tentacles based on the most important thing, which is what our customers are saying, and that's a different way of looking at it.

    29. CVClaire Vo34:43

        Yeah, and I want you to show one more sort of marketing use case off this-

    30. MBMatt Britton34:48

        Yeah
12. 35:25 – 37:51

    Generating redacted blog content from customer conversations

    1. CVClaire Vo35:25

       one-off things.

    2. MBMatt Britton35:25

       So this is the last one I'll show you, which is, this one was controversial at first, and it ha- it required massive testing to push it live, which is... So we speak to somebody, a sa- a financial service brand, and they talk-- Suzy is a market research company, right? So we compete with companies like Qualtrics and SurveyMonkey, et cetera. So we're gonna have a-- We had a call in, with a financial services company, and they wanna name a, a new product. Say, it's a new credit card or something. That's a use case that other financial services companies might wanna use us for. Now, obviously, we can't share that X bank is thinking about renaming something, so we-- but we wanna share that Suzy can now do this new use case. So what we did is we've done an automation, where it basically extracts any identifying information from the call. So basically, that includes the brand, the brand name, any specific strategy that the company had, anything that's identifying to them and-... we redact, and we have to test it over and over and over again to make sure that nothing could get through that could be a- 'cause we'll lose customers, and, and we breach confi- so we can't do any of that. But at the same time, if a salesperson just talked to a beverage company about, you know, testing packaging, they're very w- welcome. Their next call will say, "Yeah, I just spoke to another company about this." And that's kind of what we want to have a programmatic approach to. So what this does is it'll take those transcripts, and it'll write a blog post that fully redacts all that specified information but focuses just on the idea of what we talked about. It'll create a graphic, a headline. It'll even create a custom, um, CTA at the bottom, and it will- and, and it'll optimize it for SEO, and it publishes it on our blog. And it publishes it 21 days later, which is just something that we want to do to even make sure to the nth degree that any privacy or anything... So we, we p- but now we have 10,000 blog posts that are created on the calls that we're making [chuckles] without any human intervention. It just goes. It goes, and goes, and goes on every single use case that you can think of, and now we're running ads against these through Google Dynamic Search Ads. So, you know, and it- we're starting to get now... It takes a while to gain SEO traction with stuff like this. But even before that, now if someone searches for anything that Suzy has possibly talked to somebody about, we have a blog post up there, and we run ads against it.

13. 37:51 – 40:19

    How this approach changes team building and hiring priorities

    1. CVClaire Vo37:51

       This is, this is amazing. I love this. This gives me so many ideas, and what I like about this is it's taking, again, your richest source of insight about not just what-

    2. MBMatt Britton38:03

       That's it

    3. CVClaire Vo38:03

       ... a customer wants but what the marketer want, what the market wants, and creating assets that then you can use to go reach similar customers with similar problems. So again, like, your most successful customers are gonna look like your most successful customers, and so you wanna go find more, more of those folks. So again, to recap for everybody, [chuckles] a single Gong call generates a summary, a Slack post, a churn risk alert, a follow-up email, a coaching email to the CSM. Um, it enriches a bunch of data. It sends out off the automations. It identify keywords for you to bid on, and then it generates content for you to both bid on and send paid traffic to, but also generate to get organic traffic going off one call. So the other thing I wanna call out for people is, in this age of AI and automation, you can take a very simple asset and extract [chuckles] the, like, nth degree of value out of that asset, which I think is such a useful and helpful workflow for people. So Matt, we had-- this is a How I AI First. You have created such a big workflow that we have only shown one, and I think-

    4. MBMatt Britton39:17

       Yeah

    5. CVClaire Vo39:17

       ... that's enough. And we'll have people reach out to you. I know you have a couple other really interesting workflows, but we're gonna get ba- you back to building zaps and running this amazing team. Before I let you go, let me ask, um, two lightning round questions, and then we'll, we'll get you out of here. One is, you know, as I've been reflecting, this is a good reflection of how great individual contributors work or how great teams work. How has this changed how you think about building the shape of your team and your startup right now?

    6. MBMatt Britton39:45

       Yeah, I think it's far more individual contributors, far more people who wanna put their hands on keyboard, people who are willing to learn, um, people who are motivated and ambitious, that are, that are proactive at finding solutions. I think those are the people who are gonna drive the next great businesses, not order takers, not people who walk into work every day and wait to be told what to do. Because you could just, you could, you could just solve what I'm able to do if I tell AI what to do. So I don't need pe- more people to tell what to do. I need people who are gonna come up with new ideas and solutions and be proactive.

14. 40:19 – 42:53

    Matt’s prompting techniques and final thoughts

    1. CVClaire Vo40:19

       Yeah, what I say is this is the age of the super IC. Like, if you can be a-

    2. MBMatt Britton40:22

       Yeah

    3. CVClaire Vo40:22

       ... super IC-

    4. MBMatt Britton40:23

       Well said

    5. CVClaire Vo40:23

       ... you are gonna go so, so far. You know, second question: Who do you think should own this inside your team? I know you're building a lot of it, but is this a role? Is this everybody's job? Who do you think needs to be thinking about building these kinds of automations?

    6. MBMatt Britton40:38

       Well, I th- I think that you need, like, a couple of g- go-to-market orchestrators that are, are almost like general contractors that are looking at the blueprint of all the different automations. But then I think you need people who are owning the output of those automations, and so the marketing team should own the output of the blogs. If that's not working, they should go to the, you know, the automate- the automation team and say, "Well, this is breaking. How do we make it better?" Et cetera. I think that's the best design, but it does require definitely new roles in the organization.

    7. CVClaire Vo41:06

       Yeah, for sure. And then, of course, the last question, which is prompting techniques. When AI is not giving you what you want, what do you do? Maybe in ChatGPT, like, do you bribe? Are you an all-caps person? What do you do?

    8. MBMatt Britton41:21

       I, I have a framework where I first set what I'm trying to accomplish, and then I kind of set the framework for the prompt, like, almost like guardrails. Like, "Here's what not to do." Uh, and then, and then I think, for me, telling it what not to do is a great way of kind of eliminating the issues I see until I get close, and when I get it close to it, uh, it outputs something I actually want, then I refine what I want it to actually do, and I think that's generally how I go about it.

    9. CVClaire Vo41:49

       Okay, so you're doing guardrail prompting.

    10. MBMatt Britton41:52

        Yes.

    11. CVClaire Vo41:52

        "Do not do," in addition to-

    12. MBMatt Britton41:53

        Yes

    13. CVClaire Vo41:54

        ... "This is what I want you to accomplish." Well, Matt, this has been amazing. I love this. I'm actually gonna go steal a bunch of your ideas for my own-

    14. MBMatt Britton41:59

        Please do

    15. CVClaire Vo42:00

        ... enterprise pipeline. Where can we find you, and how can we be helpful?

    16. MBMatt Britton42:03

        Uh, you can find learn more about me at mattbritton.com. Um, I just, uh, rolled out a new book in May called Generation AI, so definitely check that out. It's about Generation Alpha and the AI generation, and then you can learn more about my company, Suzy, at suzy.com, S-U-Z-Y dot com.

    17. CVClaire Vo42:20

        Well, Matt, I really appreciate it. Thanks for the time.

    18. MBMatt Britton42:22

        Thanks so much, Claire. [upbeat music]

    19. CVClaire Vo42:25

        Thanks so much for watching. If you enjoyed the show, please like and subscribe here on YouTube, or even better, leave us a comment with your thoughts. You can also find this podcast on Apple Podcasts, Spotify, or your favorite podcast app. Please consider leaving us a rating and review, which will help others find the show. You can see all our episodes and learn more about the show at howiaipod.com. See you next time! [upbeat music]

Episode duration: 42:53