AI Powered Customer Intelligence To Grow Your Business | Varun Sharma

Hey CX Nation,

In this week's episode of The CXChronicles Podcast #273, we welcomed Varun Sharma, Co-Founder & CEO of Enterpret based in New York, NY. 

Enterpret provides custom AI to transform how you understand customers – from feedback chaos into clear, confident action. Harness superintelligence that feels like intuition, so your product and CX leaders never miss a signal.

The Enterpret platform supercharges teams via advanced LLMs to help brands like Notion, The Farmer's Dog, and Perplexity build better products and experiences.

In this episode, Varun and Adrian chat through the Four CX Pillars: Team, Tools, Process & Feedback. Plus share some of the ideas that his team think through on a daily basis to build world class customer experiences.

**Episode #273 Highlight Reel:**

1. On a mission to connect product leaders with their customers
2. Pioneering customer intelligence with AI
3. Understand your customers wants & needs
4. Creating actionable reporting to lift your CX & EX
5. VOC support to help grow your business

Click here to learn more about Varun Sharma

Click here to learn more about Enterpret

Huge thanks to Varun for coming on The CXChronicles Podcast and featuring his work and efforts in pushing the customer experience & contact center space into the future. 

For all of our Apple & Spotify podcast listener friends, make sure you are following CXC & please leave a 5 star review so we can find new members of the "CX Nation". 

You know what would be even better?

Go tell your friends or teammates about CXC's custom content, strategic partner solutions (Hubspot, Intercom, & Freshworks) & On-Demand services & invite them to join the CX Nation, a community of 15K+ customer focused business leaders!

Want to see how your customer experience compares to the world's top-performing customer focused companies? 

Check out the CXC Healthzone, an intelligence platform that shares benchmarks & insights for how companies across the world are tackling The Four CX Pillars: Team, Tools, Process & Feedback & how they are building an AI-powered foundation for the future. 

Thanks to all of you for being apart of the "CX Nation" and helping customer focused business leaders across the world make happiness a habit!

Transcript

speaker-0 (00:02.274)

Alright guys, thanks so much for listening to another episode of the CX Chronicles podcast. I'm your host, Adrian Brady-Cesana. Super excited for today's show guys. We have an awesome guest today joining us. Varun Sharma, co-founder and CEO of Enterpret based out of New York City. Varun say hello to the CX Nation, my friend.

Hello everybody, excited to have you today.

So number one, I loved the conversation that we had a few days back. You were building a super duper cool company with Enterpret. You guys have a super duper cool story. I'm excited, man. I'm excited to kind of dig in today and have you kind of walk through how you and your team have thought about the four CX pillars as you've built and scaled Enterpret. And then I'm really excited to pick your brain as we chatted about last week. Some of the things that you guys are doing with AI. And maybe set a different way, Varun.

Some of the ways that Enterpret is helping to literally lead the navigation or the migration over these next thousand days, man, whereas you and I are joking, like, customer-focused business leaders across the world are going to need a lot of help with where we're about to go. So why don't you take a couple of minutes, man, set the stage, spend a minute or two introducing yourself and talking about some of the stepping stones that kind of got you into this place that you're in today as being the CEO of Enterpret.

Yeah, sure. Happy to and it's a pleasure to be here, Adrian. So I'm Varun, co-founder and CEO of Enterpret. We've been running the company for like last five years. Prior to that, my background is in high-tech customer success over the previous decade before I started Enterpret. So I graduated from, my master's from Columbia University in New York. Was studying like operations research, which is like applied math and management science. After that, my first job out of college was at LinkedIn.

speaker-1 (01:49.698)

where I was a pricing analyst basically first for pricing ad campaigns, but soon realized I like talking to people more in business more than Excel sheets. So I figured a way out to become a customer success account manager basically. Really loved that, was very good at the right of the bat. And so saw, did that for like three and half years overall at LinkedIn.

south at that large scale. And then what would really be part of helping build a company from the ground up. So move from New York to San Francisco, join a startup of that time called Amplitude Analytics, which is a product analytics company, which helps you understand user behavior by tracking user actions, basically. So when I joined, there were like 15 people in the company.

I was in the first enterprise CSM, did like hundreds of deployments of analytics basically with like product teams all over the world. Incredible learning experience, saw the company scale from like 15 to like 300 employees during my four years there. you know, and now they're a publicly listed company, they're doing over like 350 million ARR, so they've done really well. So finished four years there, then joined this AI company called Scale AI, which is an AI infra company. So I was doing some AI customer success.

for like teams like Facebook and Airbnb and likes with that. So, and that was a very fascinating experience is getting to see how these like larger tech companies are beginning to use AI to build a better product basically. And this is 2019, like no one was talking about AI that proactively, like it was almost considered like a negative signal.

Absolutely. I was just going to say this and I was going to call out the timeline because you guys were way ahead of this. Now it's up front and center. Five, six years ago, only the smartest and the brightest and the best of the visionaries knew where the hell this thing was going to go.

speaker-1 (03:30.112)

Yeah, and then as it just happens sometimes in life, my younger brother is also an AI researcher by research. So he was doing research on how to make computers understand bilingual speakers. So when people talking like Spanglish or English, they mix multiple languages, like computers would trip up on the grammatical syntax of that. But he had published research papers, had like 100 plus citations as a primary author. And then he was a leading backend engineer at Uber.

So he was also sharing here the research papers coming out have been pretty crazy in the last 12 months. And I was like, yeah, I'm also seeing some like very interesting things with AI, how these teams are using AI to build a better product experience using AI, which I was pretty fascinated to see. So we got the conviction this technology is going to go through some rapid evolution next few years. But, but, then it led me to the, I had a lot of like deep conviction, a prior problem statement from, from amplitude, which is like, everyone knows what is happening in the product, but why something's happening, one really knows.

So if you take the 10 smartest engineers at any smart company and you ask them, hey, what is our week for retention? Everyone will say, oh, it's 27.88%. Why is it 27%, why is it not 40 %? Oh, I don't know. Like, you know, I saw a tweet. I think there's a problem here. Then people start guessing. But what I realized is like in today's world, like your customers are telling you unprompted exactly what they like, what they don't like about your product, about your service. And that data is sitting in 10 different silos rotting away.

So if you could bring it all together in real time, unify the data and use AI to extract insights and structure the data and make sure that the right insight is presented to the right person in the right channel at the right time, man, you could unlock so much customer love and customer centricity in how products are built and how services provided. So that was basically the opportunity that we saw. So I quit my job, he quit his job. We started in 2020, March 1, 2020, I landed back in India to start Enterpret.

And then I moved back to New York again. so it's been like super fun just building this product while the technology has gone through incredible evolution. But we had deep condition the problem statement and it's been great to see the tech catching up to the vision in the last five years.

speaker-0 (05:39.384)

That's amazing, man. Number one, thank you for sharing that. Number two, super cool that you get to build this thing with your brother. It's super cool that the, you're getting, even though you guys before this were doing different things and you're on different paths. Pretty cool that it kind of came back together. And you and I talked about this the other day, man, like when I was a practitioner, when I was a CX practitioner and before, you know, starting CX Chronicles about five years ago, like,

I joked with you, some of my very best friends in every one of my companies that I worked at in New York City, or even when I came back here to Buffalo, New York and was working at EC Auctions, oftentimes it was our data analytics or it was our informatics leaders because it was number one, trying to understand data at a different type of level. Number two, was trying to understand which variables or which signals of customer feedback, or even of me. And you were talking about this the other day, like,

even before you start the damn customer journey, all the way back up at the very beginning of a customer journey where you're marketing and you're selling, right? You're trying to create awareness or consideration before you can convert. Guys, there is so much fricking information sitting in our CRM and our tickets, even in our Slack room, there's some really cool things that you can do with Slack because that's, if you don't know how to manage that well, that can become a black hole. But point being is like, you guys saw this world of like,

All right, wait a minute. Companies are literally putting this information everywhere, right? How do we pull it together? How do we start to extract actionable insights that we can share across the business, man? And so super duper cool, man. Super duper cool.

I mean, I totally agree with you. Let me ask you this. When was the last time you opened a support ticket for a product?

speaker-1 (07:25.933)

Yeah.

Never. No one is sorry. Meaning sorry that because of what I do, um, because what we do in this type of world, I just often don't because I already understand the entire, I already understand what they're doing. And then oftentimes you can, people like us, we can literally look at it. We see what taking the solution they're using. We could tell how they're triaging. We could tell sort of how they, so like I, I, it's, unless I absolutely have to, it's a high severity or complex thing that I have to touch the company. I don't it's, already know that it's.

Yeah, so, but what...

And you're right. think most people are like that, including myself, right? Like we, like a lot of people will go through the pain of like silent frustration for a long period of time. And it's only when something is like so deeply frustrating, yet you care enough about the product that you want to let them know. It matters enough that you want to open a support ticket, right? Then you open that up, right? You take your time to really explain the problem and solve it. So,

like any support ticket is a tip of an iceberg and that is just one channel. So and the fact that most companies at scale, they get like 50,000, 100,000 support tickets a month. Then they have like hundreds of like, know, thousands of reviews. They tens of thousands of app reviews. They get tons of comments on Twitter and Reddit. Like people go out of their way to tell you what they're liking, not liking, you know, because it matters more to them. Otherwise, like 95 % of people

speaker-1 (08:53.614)

They're feeling the same thing, but they don't even, they don't have time. So people who care to that extent, like they are telling you like what they don't like to not let, utilize the data to its fullest. Oh, what a sin it is, you know, in my opinion, right? It's like that is so like such a waste. know, we are this opportunity to change that. So I think like, because it is not only those 50,000 tickets, those 50,000 tickets often per month.

They don't even do it. They don't even submit it. They don't even tell you.

speaker-1 (09:21.218)

They represent the pain of like 5 million people. It's just like 4 million 9.8 they don't even bother to open it to people, but they suffer the frustration.

It's so true. It's so, true. So the interpret platform, you guys, um, you, you essentially supercharged teams. we were talking about this today, and I don't to, don't want to put words about, um, you supercharged teams, the advanced LLMs to help that some of the brands you guys work with are pretty incredible. You got Notion, the farmer's dog, perplexity, perplexity is a big one, man, right? Um, to build better products and experiences. What you're getting to right now, meaning taking all of these words, this feedback, the sentiment, the specifics of

of keywords or key phrases or key target themes, right? Can you spend a minute just kind of talking about, because the reality is even for some of our listeners that are ahead of some of this stuff, spend a minute kind of talking about in the early days of a turper, how did you guys start thinking about how to maybe build or leverage some of the LLMs that were out there? then, Vareena, I remember from our conversation last week, like you were telling me about just even the way that you can start to customize these LLMs and some of the key

logic that goes into it by way of each company, especially if you start to leverage those tickets, leverage those sales calls, leverage those support calls, leverage the feedback, all the signals that you keep talking about. How did you, like what was some of the early days or the early thoughts around how you guys were gonna do that? And then if you can at a high level, sort of what does that look like today in a simple type of way for our listeners who maybe don't even understand what some of that stuff looks like?

Yeah, I mean, it's been super fun building interpret. We don't matter to the problem statement, not the technology. It just so happens that the only way to solve this problem is using AI. So what interpret does at the high level is, know, we, our vision is to build the operating system of for customer centricity for all companies in the world. And we do that by unifying.

speaker-1 (11:16.142)

helping you unify, understand, and act on all of your customer feedback across all sources in real time. So we pull in your data, customer feedback data from like support calls, sales calls, support tickets, surveys, Slack, Reddit, Twitter, wherever any sort of...

SurveyMonkey, wherever your customers are talking to you in any format, and we can bring that data in real time. We use LLMs. We also take in your business context, like your health center documentation, your product change log, to understand your business language, your context, and you structure it all together into a customer knowledge graph, basically, in real time, and apply a five-level taxonomy feedback classification on that across all channels, and then categorize all feedback with your business context. That's the real mode, basically. Now, when we started, we were...

We started before even GPT-3 had come out in beta. We were just building in-house models. It was a drag. It was very painful, very difficult. But we're doing the best we could. And then as the technology got better, so two months before, ChartGPT was released. think I don't remember the exact date on this, but ChartGPT was released on November 30th, 2022, if I remember correctly. There was a paper that came out which led to ChartGPT, which was on instruction fine tuning by OpenAI.

So we would keep reading this research paper to see what else is coming out. And when that paper came out, we experimented with the, to see like what we could do. And that was first time we saw like a glimpse that we could actually fine tune these LLMs basically to really for our specific use case. So that was really insightful, exciting. So then we re-architected the entire platform over the next nine months using LLMs, leveraging LLMs for very specific jobs, basically. So we had a different LLMs which would understand

Twitter and Reddit threads to like summarize and get the context basically, because it's like, it's a comment to a comment to a comment. We are a different LLM that understand a 45 minute sales call and understand what feedback to extract from that, how to really analyze long transcript, a different LLM to understand the context of in a survey, half the feedback is the question itself and half is the answer. Things like all of these nuances, right? Like it's not just like putting a wrapper, but like really understanding.

speaker-1 (13:22.158)

How is feedback presenting itself in every channel, basically? We train an LLM for each of those tasks, right? And then we use different LLM for summarization, a different LLM for categorization, and different LLM for extracting insights. So we build this architecture using 23 or different fine-tuned models in production, and then really start clicking, basically. So customers finding value, and that's when we really start seeing a PMF getting stronger and stronger, and the revenue grow accelerated. And then that led us to like, you know,

very successful series A like last year that we raised. And then what we did is like we thought like we could, the technology keeps evolving. So we did another re-architecture basically in the last 12 months just to make ourselves more future proof basically. And now it's like state of the art models like we use GPT-5 for inferencing like the models, the text generation is done by Claude Sonnet 4.5. So the depth of insight is unparalleled today.

that we are able to generate and able to answer any difficult question about customer feedback, we can ask it, we'll get the answer. So I think we've been able and so I think honestly in today's era, like there is no mode that's permanent, right? You just got to keep building, keep improving your product, your core features. That's what we've done like over the last couple of years. So and you keep experimenting, like as better models come out, like how can we modify architecture to incorporate what is state of the art to continue showing that we can.

deliver the best, most specific, insightful answers for product and CX teams.

I love it. And the only reason why I thank you for doing that, was an excellent, the only reason why I do this is I'm still, and you know this because you guys are building this stuff and you've been building it for half of a decade now, but it's wild to me how many extremely intelligent, successful, just super experienced leaders, and in all fairness.

speaker-0 (15:13.198)

Many of them, it's because they've got the blessing and the fortune. They've got product teams and engineering teams that they look at as having to think about some of stuff. they don't understand the importance and the criticality of where these language models are really truly going and how rapidly, number one. Number two, guys, you maybe have heard me say this in the last couple of episodes. I know I've been saying this a lot in conversations with some of our clients and our prospects at CX Chronicles. I almost look at it like,

Everyone knows good information in, good information out, shit information in, bad reporting out. Everyone knows that logic. with some of this LLM advancement and what you're kind of bringing us through, it's almost like fuel, man. It's almost like when you pull up to the gas station, you're putting 93 into your Porsche, you're putting 89 into your pickup truck, and then you're putting 87 into your Ford Taurus, right? So there's these different qualities or levels of fuel.

And honestly guys, the way that some of this AI logic and these AI outputs is gonna work is the more time you spend creating the actual language and then the prompt logic, like Varun is walking us through, the better the quality. Now the last thing that I just wanna pick on real quick that you threw in at the end, Varun, do you think it's almost like the best companies on planet Earth in future?

They're really going to be the ones that continue to refine that better than anyone. You just said that it becomes a daily task, Then we got to start every day. We're trying to get a little bit better, a little bit tighter, a little bit sharper. We're reprompting. So like, I love that. I love that type of thought. One really quick thing before I forget. You mentioned building on OpenAI, Jai Pateek. Guys, I don't know how many people have stopped and kind of just taken a look at this, but like OpenAI's revenue growth is rapid, right? It's absolutely rapid.

In 20, with 2024 revenue is going to be estimated to be almost $4 billion 2025. There's reports that are suggesting that they might hit 13 billion. And then, and then like guys, like within the next 500 days, they might hit 20 billion. long did it take Salesforce to hit 20 billion?

speaker-1 (17:16.705)

Yep.

speaker-1 (17:29.078)

Yeah.

Like this is fastest growth we've seen on earth.

I think. Yeah, yeah, exactly. mean, we are currently in unprecedented times. I think there have been many levels of excitement and analogies given. I mean, I'll just share an example. When we were considering starting Enterpret, like my brother, who's my co-founder, he convinced me on the technology bet, basically, and I convinced him on the problems in the bet. But what he told me then is like, he said, I think this could be as big as like going from on-prem to cloud.

basically, right? Like this could be the AWS moment, basically, when AWS came and it changed how everything became on the cloud. And that's a huge statement to say from someone who's in the field. And then I think what actually transpired next few years seemed like, oh, actually, this might be as big as the invention and acceleration of internet. What is the impact of that? But actually, this might be...

as big and as important as the invention of electricity.

speaker-0 (18:34.062)

I was, dude, this is what I use. I always use the electricity one, bro.

It's like how important was that in human evolution, venture capital city and the industrial evolution and all that happened in the world and economic growth that I saw all over the world because of that. So we are in an unprecedented time. think, know, 50 years from now, history lessons will be taught about this decade, like what rapid acceleration growth we had. So the growth revenue, the growth rate of OpenAI and Anthropic

does not surprise me. Like it is very well deserved. They are as fundamental to global economic growth now as electricity wasn't doing its time. So I think it will continue to grow. I'm very, very incredibly bullish on the potential of it. And I don't expect it to go down anytime soon.

Totally agree. One last thing I want to add on this. A couple months ago we were at customer contact week in Las Vegas and guys, I can't, it was one of the keynote executives that was delivering a speech. I'm gonna have to put this in the show notes because I don't remember off the top of my head, but it was CEO of FedEx and he basically was talking about his comparison room was he basically said,

Everyone can say whatever they want about what AI is like electricity or internet or whatever they would the spread to the spreadsheet Whatever they want to say. He said I'm gonna take it. I'm gonna take it even further. He's gonna say it's almost like It's almost like think about it. He was think about when the comic killed the dinosaurs. He said when the comet hit Life on earth had to change entirely as we knew it right all that and his thought his his message to the whole crowd was like

speaker-0 (20:23.82)

I'm going to leave you with this, and I know it's simple, but his thinking was like, for those of you who aren't making investments, educating and preparing and getting ready for where this is going to go, you have two options. You either get on board right this fricking second, and you start moving forward with the times, or you're just going become extinct. And I thought that that was such a good one because it's like me and you know that. There's still so many companies out there that

Maybe AI will affect or impact their industries or their businesses, maybe not as rapidly as say SaaS or technology. Maybe it's more of a physical, like a physical or a tangible or a traditional type of, but it's going to eventually. so anyway, so let's jump into the first pillar.

Yeah, I think maybe I could just end the last point. think that is a very graphic analogy, but it makes a lot of sense and a good wake up call. think when we were going back to previous question, when we built Antebrite 2.0 and the React recently, we we incorporated a couple of our core beliefs into that re-architecture as well. So one of the core beliefs was like that models will continue to get like faster, better, cheaper, you know, for the foreseeable future. Like even just today,

Gemini 3.0 was launched by Google, which seems to be like now the state of the art model and meets like other models in the world. now Gemini has the best model, Google has the best model in Gemini 3.0, now OpenAI will launch something else, Anthropo will launch something else, so on and so forth. This will continue. And then what we've done basically is like, it is now very easy for us to swap in a newer model into any part of architecture anytime we want. So that was the re-architecture. Hey, the models will keep getting faster and cheaper.

As soon as a better model comes out, we should be able to upgrade and add in any, improve any part of the product super easily. So, and so I think when people build using LLM's AI, like, you know, it's one good thing to do is like really think about some of the core beliefs that you want, bet you want to make like this one, and then build with that in mind. Like how, like the goal is to like future proof basically as much as possible in how you build and not just build like for this given moment.

speaker-0 (22:31.278)

So couldn't agree more and I'm glad you brought up Gemini. So here at CXE we're built on primarily Google and HubSpot. I'm glad you brought it Because this kind of backs up some of the stuff we're talking about about electricity or freaking dinosaurs, right? Or the asteroid. Gemini, that's another one. Google already has over 650 million monthly active users.

Yep.

In the summer guys, had 400 million. And then in Q1, they had 350 million. So the growth here is just unprecedented. And I think that the reality is, I keep saying on the last, almost this entire year's worth of episodes, that essentially tomorrow's leaders are going to be the people right this second.

Who are those users? They're gonna be the first wave of professionals or SMEs, subject matter experts. I don't care if you're talking about technology, accounting, finance, law, operating, know, the list goes on. This growth that me you were talking about, Varun, and this engagement of this type of ridiculous hockey.

Guys, these are the future leaders I keep talking about in these episodes where they're gonna be not only the first people that are using, understanding, comprehending, mastering, Because Verun is gonna be 20, Pareto's always gonna prevail, there's gonna be 20 % of all those users that literally become the builders of tomorrow. And then we're go back to your electricity really quick. And before we jump into the pose, I promise we'll jump into the pose, but this is awesome, and I knew you were gonna bring this. When electricity came, think about all of the industries that got built on top of it, right? Before electricity was here,

speaker-0 (24:13.312)

light bulbs, transistors, wiring, all of, I keep joking when you're generically calling it, the axes, the picks and the shovels, right? That different thing. All of the hardware, all the things. Come on, man, this is why it's like, this is gonna get fun and you guys are in a perfect time and place for it. So with that, let's jump in. Let's talk about the team at Atterpid. I'd love for you to spend a couple of minutes kind of talking about the first pillar of team of, how have you and the team at Atterpid, number one, like,

When you and your brother went out to build this thing, how did you kind of think about what were some of the first people you needed to bring under the team? Have you kind of thought about team building? And then today, now that you guys are growing and now that you've really been able to, you've been hitting your stride over the last year with working with all these incredible customers, how do you guys, how do you think about Team VrooN or how have you placed some of your team bets or how do you sort of think about what roles or what players you want to put on the pitch every single solitary day as you grow and you scale interpret into the future?

Yeah, I mean, we could probably spend three hours talking about this. I'm very, very passionate about this topic. But I think what I'll say is like,

I think at the end of the day, my belief is like, at the end of the day, a company is nothing but a group of people working together. Everything else is like software for rent, right? Like I rent open AS models. Someone else is renting my software that we're trying to sell, right? So it is just like I rent Slack, I rent whatever. We just like pay for subscriptions and tools and all of these things, but a company is just a group of people working together who believe an idea is worth chasing, right? A problem is worth solving. So that's like a very...

extremely simplistic and view, but you've taken on company building. So just to make that this is the most important thing. So currently the team is about like 50 people. We have a team of like 30 odd in the engineering design and product org and then team of like 20 for like operations and go to market. Split between, you know, Bangalore in India and then New York and some people in Denver and Seattle as well. And one person in Chicago. So the, the,

speaker-1 (26:18.47)

the belief you've taken basically is like on the engineering side of EPD side, like just general belief is like, we try to keep the team as small as possible. So like, you just chase like, like, like a big team is a vanity metric. Like the number of people who work for your company is a vanity, it means nothing, right? Like we just saw recently like this company called Gama, which is a AIS slide, creating tool. They reached a hundred million dollars in revenue profitably with just 50 employees, right? Like a 2 million revenue per employee.

incredible. That's about to become the norm. know about freaking time,

That's going to continue happening.

Yeah, it will start happening. The small teams will make like this incredible growth in the age of AI. I that's one the big things that will get unlocked. And so I think we took a similar piece, like, what is the minimum amount of team that we need to continue fueling the passion internally, making sure that growth, but also like, you know, we can continue investing in R &D and just growing the business. So that's a premise that we take basically like, and so like hiring is like accepting defeat. Every time you make a hire, you sort of accept defeat.

we can't get the job done, we need one more person for help. I like that. If you take a very humbling view to hiring, it gives you lot of respect and nuance, like, what kind of help do I actually need? And what level do I need? What scenario do I need? So that's the perspective that we take. And how we've done on EPD side, I think we've tried to balance both...

speaker-1 (27:50.67)

our internship program turned out to be very successful. We hired a very smart college interns, college kids and interns. And then like some of them turned out to be incredibly like incredible full-time employees right off the bat who were trained by us, groomed by us and showed their potential to an internship. And on the other spectrum, we've hired like very good senior engineers as well from other companies who wanted to be part of something bigger than themselves. So I think we've tried to balance that on the engineering side of house. And on the go-to-market side of the house, I...

I have a thing for like smart generalists basically. I think in an early stage startup, you just have to do so many different things. like our entire customer success team is built up of like ex management consultants basically. They're very smart. All of them use cloud code, they use cursor. They're very innovative. They build internal tools. Like they're very AI native.

So they don't follow a typical old school customer's playbook. They just solve customer problems any way they can, you know, using AI. And then I have a lot of smart generalists, right? So every time we have to take up a project, like they can, you know, they can understand ambiguous problems and solve it. So on the go-to-market side, like hire like smart generalists early on. And on the EPD side, like balance, like hungry, scrappy, early career, like engineers, interns out of college.

with very sophisticated polished engineers, leaders who are trying to have a bigger impact.

love it. So wait, you just gave me a bunch and I want to unpack a couple things. Number one, love the comment. Let's start with the CS side because guys, don't kill me for saying this, but I've been kind of calling it out for a while. CS, as y'all know, is definitely going to be one of the first areas impacted by the Pareto rule.

speaker-0 (29:44.718)

But I would argue, I would gladly place a bet for a beer, my friend, that probably in less than a thousand days, you're gonna see most CS teams and organizations at businesses and companies across the world at every single size level, SMB, mid-market, enterprise, they're already showing us, they're already doing it, guys. You're gonna see tomorrow's CS teams and leaders

they're literally gonna do, there's gonna be 20 % of those professional people that are gonna do everything. And then the other 80 % of the people that can't do what you just said where it's like leveraging, so number one, just becoming far more analytical, leveraging automations and workflows, getting exceptional at signaling and essentially analytical or customer portfolio health signaling, right? So being able to just generally see what's happening on a portfolio or in a pipeline and then knowing how to act upon what type of

activity comes about it, not with doing it, but with prompting it and or with create. Yeah, they gotta do it in the, you're going to do it in the, in the beginning. Then you're going to immediately learn how to prompt it and how to automate it. Cause you're not going to need a 50 people on a CSM team anymore. You're probably going to have 10 people like what Varun just said, who can literally manage the entire scope of what today's 50 people do on top of whatever your growth goals are. that's number one. Number two, it, what you were talking about with your engineers and your devs. And then I have a question too, but

It's funny when you get, when you look at the interpret LinkedIn corporate profile of it's really interesting to see on your team that's growing. guys are, you're, you're, you're, you're, you're almost to 70, 70 people approaching a hundred people on the team. And you're right. That's a, I'm not, that's vanity metric. Totally agree with you, you guys are the right folks. And what's interesting on your, your, people outside of your team today, what they've studied, right? Cause you can all go into the people, the people view of LinkedIn and kind of see what, where are people, what they've studied, what is their thing. You got a silver run on your team.

Here's the top things that people on your team have studied. Computer science, computation science, information technology, economics and mathematics. Fucking love it because think about it. Tomorrow's companies and then future leading companies. Guess what guys? That's going to become super duper common. And then to Verun's point, you're still going to need people that are awesome with humans and connecting and

speaker-0 (32:06.09)

I still believe last mile room last miles not to be done by a still think last mile with sales last mile with support last mile success still think you're going to need some pretty incredible people to sort of finish things off and bring it across line. But like it's super cool to see that before we jump off to my question for you is what's been like over the last couple of years. But now you've got this you're a global team. got a huge you got a huge part of your team over in India. You got people across North America New York City San Francisco Chicago.

What's that been like man? It's gotta be pretty damn cool. It's gotta be pretty cool to have literally different cultures, different countries. What's it been like to kinda grow that type of a team?

It's been super super fun. I'm of like nerded out about creating and scaling great cultures. while it's been super fun, I also keep thinking about and experimenting with how can we scale up culture basically and make sure we don't slow down or we still operate as one team across all these different cities. So.

I enjoy a lot of it, but I don't take any of it for granted. Basically, maintain some healthy level of paranoia to make sure that the internal culture remains like only gets better even if the people increase or like the time zones increase and so on and so forth. And I think like to your comment about like the degrees part, thank you for doing your research, Adrian. I would say like, think all of that is very important. I would say like, you know,

there is something to be said about like deep rational thinking and like other like arts degrees also be incredibly powerful in the future as well. just to give you an example, like our sales engineering leader, like she, know, her name is Kelly. She leads all of our like enterprise pilots and evaluations. Like, you know, she studied like political science in undergrad and she's like our technical like pre-sales leader. Like,

speaker-1 (34:05.602)

But because she studied that in government, think, and she started a career as an SDR. So she was like not even technical. So she just started by like in sales, like with a political science background, and then she started over time to become more and more technical. But one thing about her is like when, why she's in this role is like, not only is she like technical enough to like lead these like enterprise pilot deployments, but she can also understand the business context really well. know, so she can see the risks coming because of us like, and she...

See you.

She can understand the psychology of a customer well, she can build very good relationships. She's a very good communicator. She can write very good emails and documentation as needed because she spent like, know, so I think I would say like, I think we're in this era where if you started studying just computer science and economics, that's great. You will have to build like the other side of the house as a skill set. You'll have to learn the other skill set to become more complete and comprehensive. And if you started studying the arts,

you over time, you'd probably need to like learn some of these like vibe coding tools and at least understand technical concepts to become more, you know, more complete in your impact can have over time.

Absolutely agree, Vroon, absolutely agree. Let's jump to the second pillar of tools. Spend a minute or two talking about, and I'll give you two paths to run down here, Vroon. Number one, I just would love to understand what are some of the tools that you and the team invested in?

speaker-0 (35:32.364)

to build interpret. So said another way, like what were maybe some of the primary tools that you guys knew you were gonna build into your tech stack when you started to scale interpret. That's one path. And then maybe spend a little bit of time talking about how interpret actually connects and works with the customer's tools. Cause you guys have all these incredible customers. Again, I'm gonna like, you if you look at,

If you look at the Enterpret website, you guys are working with some world leading Canva and Perplexity and Notion, as I mentioned earlier, Fanatics. There's all these really, really, Monday. Spend a couple minutes kind of talking about tools, Spend a couple minutes kind of talking about what it's been like kind of managing all that, leveraging all that, making some of those big investment decisions on the tech side over the last five years as you've been building the company.

Yeah, we've kept a beginner's mindset, experimental mindset to get the best tools that we can, but also be open to upgrading as needed. So we're on AWS stack as much as possible. So we are behind a virtual private cloud in AWS. We use a lot of AWS services, but we also use models from OpenAI. And through AWS, we use a lot of the end-stopping models as well.

We have Snowflake as our data cloud house provider. We've been doing some experimentation with some other platform as well. But I think the idea is your tech stack should basically represent the problems that you care about solving the most and should be represent what is needed to solve those problems. So for us, being a, you

Like for example, like, you know, we are an analytics platform. So we need to be able to like store data very flexibly. We need the queries to be very fast and performant. So like just you should know what you're trying to solve for and get like the best solution that fits those needs. So that's how internally we built a solution to say, what can, what will allow us to build the best product for our customers and which vendor can get us to that, you know, a reasonable price, I would say. And then internally, yeah, please go ahead.

speaker-0 (37:41.644)

So, yeah, exactly. Before you go on with that, so awesome. Super awesome. So like, number one, dude, that's one of the best answers that I've gotten in terms, in all these episodes around certainly how I think about it and then certainly how I think, how we both know tomorrow's future leaders and tomorrow's future executives need to think about it because...

We just watched the biggest, we just watched a 20 year bull run of SaaS technology going from a, this is why, by the way, this is why I called it some of the open AI and some of the Google growth for us. We just, guys, we just watched 20 years of SaaS growing from these interesting numbers, these, you know, these, uh, you know, uh, a hundred billion, 200 billion over 20 years.

up to today where we are at 1.27 trillion, still has a compound annual growth rate of 15 to 20%. That's about to stop, by the way, for when you I both know that because like, everything that's gonna come now is going to literally force utilization. It meaning it's going to force companies to look at, okay, we've got five tools, 10 tools, 20 tools. You and I both know big companies ruin and then because you guys have been digging in and.

You're plugging in the inter- you're plugging all the wires into the interpret magical box, if you will. Some of these big companies, man, they got a hundred different pieces of SaaS technology. None of them are connected. Most of them are underutilized. None of them are pulling back to an aggregate place in which people can disseminate and then essentially socialize the most important things that you're seeing. And then the other thing is this, like, most of those said tools

they already have a bunch of different incredible AI resources that are being built into them. Said in other ways, like, things are definitely gonna change quickly. And let me just say it this way, like, best companies on planet earth, so on that 1.27 trillion, best companies on planet earth, man, see utilization rates 30%. So like, the reason why I keep bringing this stuff up, guys, is like, between sales, engineering,

speaker-0 (39:52.056)

Customer success, definitely customer support. It's already there. You guys know it. You're already doing this every single day with your team. But like, it's gonna be impossible for CEOs and CFOs to not start auditing and digging in to their tax decks. Because guys, that's a big damn number for us. Once you get north of 100 people, 500, 1000 people in a company, start doing your multiplication of every seat count of Salesforce, HubSpot, Zendesk, Intercom.

It goes on and that's before you even get into some of your cloud stuff. And my point here is like the spend for the utilization will not be acceptable in the next thousand days, especially as customer focused business leaders, executives and founders across the world continue to get acclimated and continue to get educated and continue to get more comfortable with what the new expectations are going to look like for teams carrying the new weight, I don't know if you have any thoughts on that or any feedback, but like I can't.

get my brain to stop thinking about this and I keep every episode of it poking people like, what do you think about this? And you know, how does it shift? Do we see the beginning of like, SaaS finally starting to maybe see almost a downward trend because AI is going to be essentially replacing some of these things that let's call it what it is guys, sub part tools that did a thing, but some people liked them and then they buy them and then SaaS was always good at customers buying something.

setting them in contracts and then oftentimes people forgetting about it and then over time not even realizing that the company was under utilizing. So I don't know like any thoughts or or just feedback or immediate reactions to that whole thought.

Yeah, I mean, there's been a lot of innovation and experimentation happening on pricing models in the age of AI. You know, we obviously seed-based pricing was...

speaker-1 (41:41.036)

the default for most SaaS products for the longest time. And you're absolutely right, like there's tremendous underutilization happening. And then there was software built to like maximize the utilization of the underutilized SaaS, which is also like seed based something like that. And then I think there's been consumption based pricing basically of like, which I think, know, like companies like Snowflake charge on. And now we're seeing like some outcome based pricing as well, which is, you know, paper outcome. For example, these like AI chatbots basically they charge like,

$1 plus per closed ticket basically.

Pay per resolution, not pay for all the other stuff.

Yeah, exactly. So like, I think there's a lot of experimentation happening on that side. I also agree with you. think the per seat model will become rarer and rarer because consumption is not like equal, right? It's not like basically like you have the Pareto principle where like 20 % will consume 80 % of the volume. And if in AI, the underlying, if all software is going to be have some sort of core AI component to it, these AI model providers charge on consumption. They charge by number of tokens you're consuming.

So like your pricing will have to reflect that at some level, right? Like for example, right? So I think like we will see a lot of evolution of pricing models. And I think my hunch is like, I think more than seed-based pricing and even more than outcome-based pricing, consumption-based pricing will become more more dominant basically. Because the challenge with outcome-based pricing is like there's not even alignment if the outcome was achieved. Just because a ticket was resolved does not mean it was done properly.

speaker-0 (43:15.022)

100 %

But consumption is like more predictable. So I think like we'll see a lot of experimentation still, like I think like consumption based pricing will become the new seed based pricing in a few years time.

I agree. had, actually, and I didn't even mention this the other day when me and you were catching up on stuff. we had Pasquale De Mayo who is, vice president of customer experiences at AWS. So MSO web services. And he did an incredible episode, two 62 guys, Pasquale De Mayo, AWS. He did an incredible job of explaining how part of the absolutely ridiculous hockey stick growth that AWS had. Number one, let's call it what it is. And he does a good job of telling the story.

They built the thing that worked for Amazon. So that's number one, meaning like because of as Amazon became a behemoth, they knew that there was going to be the need for them building something so that they didn't have to continue to pay all these outside partners and vendors to be able to do the things that they knew that they could build themselves. And then they knew where their growth was going. That's number one. Number two, though, once they built it and it was working and they started to bring it to market and offer it to others.

they knew that it was going to be a consumption model from day one. And so even to this day, it's really interesting. Like AWS startups, so guys, for some of our founders, if you haven't done this already or reach out to us and I'll connect you with them, like AWS, for example, they will literally work with small businesses and they will almost do education, familiarization, help with like, said generically, pulling you through adoption because they don't know which one of us is going to fucking blow up. They don't know which one of us is going to become the next behemoth or the next world leader. And then

speaker-0 (44:52.204)

from a consumption model like Bruce was talking about, that feels just like such a better way of, if there's any shot of retaining and continuing to grow at the clips I said earlier, 1.27 trillion growing at 15%, maybe 15 % goes away. Maybe SaaS technology becomes more of a traditional type of Calgary where it's like three to 5%, right? Something that is rapid, but maybe that's what happens because then they're gonna have to flip to whether it's consumption model or paper resolution model.

Benioff and Salesforce are clearly betting on paper resolution. You bring up a good point where I would bet on Benioff too, but like every business has a different type of way of technically resolving things. So I don't totally buy that you can slap a tool in and then do that without either like an interpret or whether you have a really, really intelligent team coming in and thinking about.

what equals a resolution across the entire journey. So it can get complex, I love it, man. Varun, let's jump to the third pillar process. Just spend like a minute or two talking about, as the interpret team's grown, how'd you guys have to kind of think about wrangling process? How did you store tribal information? How did you empower your team to really be able to document and chronicle all the things that were happening as you guys were growing and scaling and learning about what your customers needed and then what your team, your employees,

we're learning on a daily basis as we're serving those customers.

Yeah, think so, great question. So what I would say that is like, think a fundamental premise is like, know, the process is for the people, the people are not here for the process, right? So the process exists to serve the people, to make things more efficient, faster, and ensure that we don't get in our own way of success, essentially. So it should be a living breathing thing. It should be like defined.

speaker-1 (46:43.776)

not rigid, should be like context-based basically over what we're trying to optimize for with some room for flexibility. And we should evaluate that every few months to see how we can update things as we evolve. Internally, for us, we run interpret on interpret for customer feedback. we provide support over Slack. So all of our shared Slack channels with customers, our team channel, all of our gong calls, all of our emails, everything, all of our Salesforce notes.

everything pipes into Enterpret. So it's the source of customer intelligence. And that data is consumed by go-to-market teams. That data is used by CS teams, which is a part of go-to-market, and our engineers, PMs, designers, and myself. So that is our single source of truth for any customer feedback that we want to use. Other tools like we use analytics, we use post-hoc and amplitude for analytics, and just like user data level.

And that's something really interesting is like we're now moving into this era of like MCP servers where like, you you can do that. So what we do is like we have an interpret MCP server and we have a postdoc MCP server. So we can connect all those sources very, very easily together. So we can say like, hey, for people who did not, you just ask question in natural language, like from postdoc MCP, like get me the list of people who created a chart but did not save it in the last one month. And then you can also then ask interpret MCP like for those lists of people.

find out what if they've given any feedback about like, know, chart creation process. And you can just marry Quant and Qual very easily, right? So that's why I say, let's keep your process nimble, like look at the tools that solve the problem the fastest. But yeah, internally like, you know, we try to keep like the systems of truth and knowledge very, very small and similar. So we use linear for project management, we use Slack, we have interpret for customer feedback, and then we, you use post-hoc MCP and interpret MCP together for like Quant versus Qual joins.

And then, know, for go-to-market, have like Salesforce and HubSpot and all these tools.

speaker-0 (48:43.886)

Fantastic. Guys, this is one of the biggest, well, if we're in just rip through it there, is one of the biggest things that we do for our customers at CXE, living playbooks. It sounds so stupid, it's so simple, but living, breathing playbooks.

How is the teamwork? How is the team built and orchestrated? What are the major tools that people need to understand across the different teams in your business? Where does the process lie? Where does the information, the tribal knowledge, the FAQs, all the internal information and intelligence that people across all teams need to understand? And then lastly, feedback. And let's jump into the fourth, the pillar of feedback. But how are you guys kind of managing, Brun, how are you thinking about...

Customer feedback, so what are some of the ways you guys are leveraging and acting upon your customer feedback today at Enterpret? And then maybe spend a minute talking about, you kind of just hit it right here with some of the things you guys are doing with your tools, but spend a minute as you're talking about the ways you guys manage your customer feedback and then some of the ways that you leverage your employee feedback.

Yeah, mean, employee feedback is we tend to make it like very informal and direct basically what people who have feedback and then you can like structurally share that and you know, things like that. But we don't put that in Entapet like Entapet is only for like customer feedback analysis. Customer feedback, think, you like I said, like we, most of it is unprompted, it's tracked in Entapet from like sales calls, like Slack messages. We have an intercom bot on our website from that.

any in-app service that we run, all of that. So all of the data, user research calls, pulls into interpret. And then we have these workflows built in. So we have an agent which can detect any bugs, based on emerging feedback. can agent that can detect any potential escalations. So those trigger basically to make sure that we're building a good product, providing good service. We have agents set tracking those things. And then we have weekly summaries being created by interpret.

speaker-1 (50:42.286)

We have a weekly summary for each big feature area that is sent into the right Slack channel. So on Monday morning, every team gets like a weekly digest of like, here's what happened to the product last week. I get an executive summary. So we build these workflows just to ensure that the awareness and alignment on voice of the customer is there 24 seven basically. And anything that is urgent, whether it's a bug or it's an escalation, they have agents tracking and monitoring that. And that is a reflection of a value that

our customers also get from Interfere and how they use it as well, which is to build a better product and provide a better service to your customers.

I love it for him for this is incredible man number one super appreciative for you coming on the show and sharing the story I think I told you I told you last week but like this companies like interpreter companies that I think are the ones that I would bet on for the future all damn day guys I think the other thing too is that companies like interpret and what Verun and his team are doing again I just I over the last seven probably 20-ish episodes for and I keep talking about this notion of like

the next thousand days, who are companies that are either going to be navigators or are going to help with the migration of the customer focused business leaders, going back to the dinosaur example, that are going to actually survive extinction. And they're going to be the men and the women that lead tomorrow's teams, build tomorrow's companies and leverage AI and leverage this wild shift in technology into the future. Before I let you go, my friend, where can people...

find out more about Enterpret and then where can people keep in touch with either you or your team if they want to learn more about how Enterpret could be a game changer for the business.

speaker-1 (52:20.622)

Yeah, for sure. So you can find us on our website with interpret.com. It's a pun on interpret, which is we help you interpret your feedbacks. It's like ENTERPERET.com. And then you can find me on LinkedIn. Just search for like Varun interpret. should be able to find me, feel free to connect me on LinkedIn, send me a message, happy to chat. And I'll reach out to a website and we have a sign up form. We'll be happy to show you what we're building here.

Awesome, it's been my absolute pleasure, brother. I can't wait to see what you guys do in the future. And I look forward to our next conversation, my friend.

It's been super fun, Adrian. Thank you so much for having me.