Soham Mazumdar, WisdomAI CEO on Turning Data into Instant AI Insights

We're live. Sorry. Hi, Soham. Welcome to Adventure with Grace. Hey, Grace. This is awesome. Super excited to be here.

I'm so excited to have another, you know, U of I alumni to be on our show. And maybe we could start with your background. Your background is super impressive. You're the co-founder of Rubrik, and now you co-founded your... well i think this is the third company right because previous to that company that you sold to facebook and then and then you started rubric which is like a public company now and then uh the successful

co-founding alumni including the founder of clean and other people who are like you know and now he started with some ai maybe we could talk about like what were some core lessons that you've learned early on in your career kind of shaping today Yeah. Yeah, you're right. This is like innings number three. They were all like very, very different, you know, like very different spaces, very different kinds of startups. Yeah.

There are some learnings and there are things that I had to unlearn along the way as well. My first startup, I actually was an engineer at Google for about six years. I started very early.

around the time of the google ipo but my first startup was like a great lesson in customers don't just come to your product just because you built it you know that was like one of my In Google, it's very easy to rack up hundreds of millions of customers trying to use your product because they get free product placement.

as a little link on the google page that simply does not happen in a startup so the first one was like my first time understanding just how hard it is to get customers and learned a lot about uh you know i was trying to sell something in the local business space and just like how to scale a business there it's very hard very different from how you would do consumer or how you do

b2b so there was a little bit of a founder market fit issue that i also discovered um the next one which was much more exciting was rubric i think that's the

been the highlight of my career. I think, again, I mean, the great thing about Rubrik, and you mentioned it in some ways, was the team that we were able to build.

incredibly high caliber folks across the board and in fact one of the things that uh has happened from rubric has been that there are a whole bunch of new entrepreneurs who have come about uh you know gleaned in one uh this was to me obviously and there are like about eight or nine others who are all doing uh quite well so again like incredibly high talent density um we went after a very

particular angle in terms of particular end of the enterprise that we were trying to service. It was like not the shiny, sexy AI kind of enterprise it was much about how do you protect people's data it sounds kind of boring but what is really cool is that the it is one of the only stories that i have been able to tell my parents and they just immediately got it

It's one of those things that you can very easily grasp and learn a lot about how do you build a large team? How do you go to market? How do you hire? How do you get great talent? a lot of different learnings which i think most learnings i would be what say that what kind of like translated really well when i was uh starting with to me i just realized that the quality of the founding team makes a huge huge huge impact

in terms of what you'll be able to do downstream. So that's probably the most consistent learning that I have been able to carry through.

i wonder so like i think the first learning is like super crucial because like i think especially coming from an engineering background like i think people are very tend to gravitate towards like let's just build something like with our existing technology just for the sake of like building stuff but i think like to sell to like customer it's like a completely different stack of skill switch um

When it comes to, like, you know, building a team at, like, Rubrik back then, like, how do you guys kind of, like, ensemble the team? And, you know, I guess, like, also, like, yeah, like, how do you kind of, like... what do you think make it so successful that I already IPO'd? And I have so many questions for you as like, you know, if I, if I co-founded a company and IPO'd, I'm not saying I would be like sitting in Florida, like sipping my refrigerator or anything, but.

i would consider like doing investing or like you know wrapping up stuff like what motivates you to like start a second company or third company technically uh yeah i think uh okay so on uh on rubric and getting a great team together so i give a lot of credit to like uh our ceo my co-founder um you should kind of hear him talk sometimes he talks about like maximal thinking

which is really about how do you build a team where everybody does their best work, nobody feels constrained. You imagine the impossible and you're always kind of chasing for it. That was the early culture of rubric. And that was just very, very, very stimulating for all the folks who came on board. And that's something that is really attractive to people.

as well that you know have a culture which is really high achievement oriented no artificial limits you know you you go as far as you can go i think that is that is something that i feel every great company needs to truly be successful. In terms of like my own personal journey, you know, like, look, I am a builder at heart, you know, like I started off as at Google, I'm a tinkerer, I love to just keep

doing things. Now the challenge is that when the first time I moved away from Google, it turned into something that I would best describe as a toy project. There was no real reason why it should have turned into a business.

uh i have learned since then but this basic passion for building has always remained right so if if not build i wouldn't really know what to do so that was one one aspect of it and and you know like honestly timing matters a lot right i mean i think We live in this like, you know, I don't know, in my entire career, like I have never experienced a moment like this where you can just see the world change in front of you or like the potential is just like so ridiculous that.

As a builder, it's almost like a siren calling that you have to go and build and do something new. And this does not happen easily. I think that was something that was certainly...

what drew me into this, just the opportunity space that AI had opened up. I wonder, maybe we could start talking about like with some AI, because it is kind of like the next generation BI tool, like, or, you know. I wonder how do you define, I guess, like, you know, getting insights from your own company and how is it like different than like last generation of?

I guess like business intelligence or like however you design. Yeah. So we are like trying our best to not use the word business intelligence, but since I've mentioned it, let's go along with it. so you know we're kind of calling ourselves as the ai data analyst and what that means that we empower data teams in enterprises to train, think of it as a digital twin, like an AI data analyst, which is, again, fully grounded in the specifics of the data and the specific nuances of the business.

and then allows you to get insights from the data. That's what we're doing. In terms of the kinds of things that you do, this will tell you how we are different from what it used to be. So imagine like a CEO. The CEO's question is, okay, how am I going to do great in this upcoming quarter?

you know like a chief marketing officer is thinking how did i be allocating my like big marketing spend i don't like these fairly meaty questions that require your business and turns out could get to this answer is it's you know it's a complex question i think something is blocking your uh microphone really

yeah wait so i think i hear the sounds that are kind of like how about now oh now it's perfect yeah okay i think i was like moving my hand around my laptop that's probably what happened i'll keep that'll be better thank you Yeah, so we're building the AI data analyst. It's going to be 100% grounded in your business data. Our ultimate goal, and this will tell you why it's different from how it's been done so far.

is that we are here to empower the business leader or the operator who is making day-to-day decisions so this is not like a trained data analyst this is about everybody else it could very well be the ceo wanting to plan out their quarter or like uh somebody trying to optimize their marketing spend these are business professional business users these are not uh you know data scientists data analysts so how do you empower

the operators to make decisions that's the overall goal that we're trying to go for uh and and this requires uh you know a few things uh first of all again every business has tremendous amount of nuance and in some ways the data analyst is the one who knows how to wrangle the data knows how to wrangle the low level uh nitty gritties of how to extract information from all the data that they kind of work with

Essentially, the goal is how do you make AI do all of this work in a manner that you can fully trust? Unlike other things, you go to ChatGPT, you ask the question, it doesn't matter how accurate the answer is or you can more or less evaluate what's going on you don't like it you ask another prompt you kind of fine-tune it but here like when you're asking about you know optimize my campaign

you there is really no room for hallucination so again how do you solve this problem through which you would make day-to-day decisions to be done by non-technical folks with insane levels of trust and accuracy

so that's the problem that we're trying to solve and if you can do this this is pretty much the holy grail it's just that the previous generation technologies have always been about empowering analysts to do the work and prepare reports and those reports would be what the other posters come from

consume. So we're completely doing a paradigm shift. Instead of like folks preparing reports, we are essentially giving direct access to questions, direct access to answers. So that's the paradigm shift that we're going after.

i wonder so like i guess like how does the i want to say how this background works but like i guess like in the past um version of uh analytics for different companies um how was it processed in the past like i know that like you know there's like you know i know that you guys kind of help um people process like you know their financial data and everything is like

oh great question uh i think this way i didn't answer your question the last question properly so so the previous version was there are folks who are trained analysts had tools that allowed them to create reports reports would be like you know imagine three or four charts put together so when you get a question about how to optimize my campaigns you would

somebody would crunch the numbers a week later they would come back and say here is this like dashboard and through this dashboard will tell you what we think you should know then when you have a new question that is not kind of answered there already uh you're essentially stuck

and then you'll get the new answer another week later so that that was the status quo what we are essentially saying is that this this gap of like taking seven days to get an answer is something that we want to eliminate and we want to optimize for uh and this is where again like if the if there is an ai you know like digital uh data analyst who is trained to

do the analysis, really understands how the litigating nuances exist, then you can reduce this time to answer to a matter of seconds or minutes, as opposed to days, which is the current status quo. And this whole opportunity to train AI to understand your data was simply non-existent two years back. And hence the previous generation of technologies were built the way they were. So that's the big.

technology shift that we are capitalizing on um i wonder so like when it comes to like analyzing the data i guess like maybe if uh from the business perspective like how much of the money are currently spending on this particular i guess like i don't know if it's like under the forecasting or just

business operation and like i guess like what does the go-to market kind of look like is that like you know you chat with a coo and they'd be like hey you are like the seattle um you know um yeah question so uh so first things first we are definitely targeting enterprises so we do speak with cfo's ceos lots of cxos

the great news is that every one of them like immediately loves our product again i'm obviously a little bit biased here maybe i love it as much but i feel like they do because they can connect with it because i me as a business leader myself in the past I have always needed insights from data and it's hard and it takes a long time. So if you say I'm going to get instant answers, that's a big deal. And I'm going to get instant answers. It's going to be proactive.

he's doing work behind the scenes every monday morning i know exactly what i should be looking at this is magical right this is great uh however the interesting dynamic here is The next question after feeling the magic is, can I trust this answer? You know, like I'm going to make a big business decision based on this. Is this reliable? And like, you know, previously there was this person I used to, you know.

talk to when I was to fully trust them, now I'm trusting this AI, how do I believe that this is correct? So inevitably, then our conversation shifts to working with the data team because ultimately they are the guardians of all the, you know. They have all the tribal knowledge. They have all the low-level knowledge. So we inevitably have to get the data team super excited about becoming like the AI model builders.

they are basically the train they are training the ai they are maintaining the ai they are responsible for the upkeep of it so how do we give them the confidence that they have transferred their they have trained the AI sufficiently well to end up being their representatives in a live setting. So our go-to-market involves selling to cios and cdos because they ultimately are the decision makers for us even though the ceos and the ceos get very excited um in terms of like

that does that answer your question uh grace yeah yeah for sure um i wonder so like how do you think about like the um from the business perspective to like i mean obviously everyone wants to kind of like find from the business perspective uh how uh maybe we could talk about like you know how does it integrate it into their own

database and then like what are some challenges let's say you know if i use uh you know if i like just dump all my data i mean this is a extreme and non-technical example but like if i just dump all my data into chat gpt and then it kind of like

I asked it, like, hey, Chagipathy, give me something. Or like, you know, the Microsoft Copilot, I just asked, like, you know, on this Excel sheet, like, gave me, you know, what's a, I guess, like, a pattern or something. Like, I guess there's, like, some sort of, like... general like direction on like what i should kind of like look at my business but how how do you think about like the challenges of like these other tools too great question i think uh again

we are a post chat gpt company right so it's just been two years and so very much inspired by chat gpt so what you described that hey i have this excel file or like you know i have some pdf report i upload it into chat gpt ChatGPT would do some analysis and tell me where to go. Just take this mechanism of how you can work with your own data.

And think about like, what would it take to make it work in the context of an organization? So first things first, an organization's data itself is sort of scattered all over the place, right? They might be using like, think of, you know, like a Databricks or a data warehouse for some amounts of data. They might have like some, you know, transactional database.

They have SaaS tools where some of the other data that is scattered. It's hard to condense a business data into the form factor of a single spreadsheet. It is inevitably across multiple different silos. These silos don't quite speak to each other. These are truly independent disjointed silos. In order for us to do the work we do, one of the things that we need to crack is the ability to work with these enterprise systems. We connect directly into these enterprise systems.

and query them as appropriate. Inevitably, enterprise systems are hugely messy. You want every enterprise system to ultimately look like a single spreadsheet. The reality is, if you looked at a single spreadsheet as one table, I'm talking about thousands of tables that have overlaps and there are redundancies that are things that are deprecated, things that have different versions.

The math to do the revenue calculation exists in the head of this one person. If they write the query, they know what to do, but then good luck trying to imagine an AI kind of understanding it. So all of this like fragmentation, the messiness, this is sort of what we are, the whole goal of like our onboarding, any customer is connect to the data, but then sort of

You can call it reverse engineer, you can call it train. Understand the mess that exists and then provide an interface through which you can interact with this data.

so it's kind of like the same principle that you described with chat gpt except that how do you do this at enterprise scale and then we require a lot of reliability like a lot of like uh repeatability you know like it should not happen that you know a common scenario like you know we are designing this product where let's say hundreds and thousands of people would be asking questions you want to make sure that each person gets you know they don't understand the low level nuances at all

and yet you want them to get identical answers for the same question so that's like another like level of work we need to do which is about how do you make sure that you can do this in a scalable manner where each one

it's not each one for themselves right there's a lot of like commonalities between questions and you should get you know two people asking about revenue should not get two different answers because if that happens there is a problem so anyway so those are like some of the challenges we need to solve

wonder, you know, I guess like one of the things that I think, you know, maybe if you want to share a showcase or like a demo or something, we can actually click present if you want to share a demo. Oh gosh, really? okay uh i did not know that was that happened uh live stream uh wait where did i present or invite So it's kind of like a Zoom-ish function. Okay. This thing. The plus button? Yeah. Okay. Share screen.

share screen okay okay give me a second to okay let's stay here but give me a second to Do you want me to add to stage or do you want to open? I don't know. Can you see something right now? Yeah, I see a good afternoon.

you can see good afternoon okay yeah no i think this is your email this is your email or this is the dashboard yeah this is like the this is the landing page it looks a bit like gpt the yeah weird part is that it's kind of uh connected to uh the difference of course with chat gpd is that i'm connected to all of these enterprise sources um yeah a whole bunch of things um and the you know in a basic form what you can do is you can ask questions of it can i add to stage uh yeah please do yeah yeah

oh by the way hi at word so edward said nice insight and like a compliment basically the audience oh okay so i'll go like super quick because i'm always like a little bit worried about unrehearsed demos on live stream um Yeah, so let's look at this, right? So the basic product, right, to start here, looks very much like ChatGPT. Obviously, the difference is that, you know, we are connected to all of these different enterprise sources.

uh and we can in fact like connect to multiple of them all at once and kind of mix and match uh the the basic you know the very basic thing what you would do here is that when you last question get answers uh so you know let's just see what's here so this is like a basic sales finance uh data set you can you can ask about revenue you can ask about uh conversion rate win rate all sorts of things

So let me start with one basic question. Show me revenue trajectory quarterly and only keep the items bigger than 10,000. Wow, love that. There. Let's close this. Okay, Grace, you didn't tell me about a live demo, but. I know, I know. Sorry. I feel like you did such a great live demo at our conference. It's like you're prepared. Yeah. In a basic form, what you can do with the product is you ask a question and you get answers.

where it gets interesting is that you know you can by the way these like little dots here allow you to further edit the answer if you want it but the next thing that you could do here is to say i see a peak in q4 2024 can i do a deep dive into it all right so now i'll turn on our deep analysis Let's see what happens. Again, you see the difference here. The first question was a very clear-cut question. I asked for a very specific number and did some analysis.

the second one is a little bit more complex because it's like you know this is like a not an not a easy question to answer because i want to explain help it explain to me why there is a peak in 2024. so it'll kind of come up with a plan and then i'll like say okay let's start off like this agentic exploration and it'll now kick off like this multi-pronged analysis to figure out an answer for me

This will take some time, so I'll not spend too much time here. Behind the scenes, the data is coming from BigQuery, which is kind of like a data warehouse where you can put a lot of your data. So there is a lot of complexity in the data itself, which we are kind of utilizing. to answer some of these questions accurately. But the ultimate goal is that this section about how the data is laid out is This is behind the scenes. This is not something that folks are supposed to see.

The main final output is this interface through which you can get your answers and so on. This is doing a very long exploration and will take a while to complete the whole analysis and come back with a point of view. including exact summary, key findings, what are some places where we should be digging deeper into? Are there some anomalies? A whole bunch of analysis into my quarter has been done.

and all i needed to do was to just speak at it and just did it right and this would easily be like a few weeks of work to come back with something and i can just keep going i can ask for further drill downs all of that so um that's a very basic version of what happens in the product um the there are other things uh you know we have uh you know you can like share this information you know you can um you know

You can prepare dashboards. You can schedule these. We have the ability to do work behind the scenes. This is like a new feature we launched around proactive agents where you can just say, um you know like i have my i'll show you my proactive agent um this is like my daily monitoring agent through which i try to monitor our product performance so look at like you know

overall usage and tickets that are being created and then if there is like ever a drop in the overall uh usage then give me like a notification and then no don't just give me a notification do some advanced analysis for me as well right so essentially help me narrow down into where i need to go and see the uh you know see the

help me kind of do that next level of analysis so this is kind of like also happening behind the scenes and every uh every morning i get like a report on like what's interesting if there is something interesting And there are other interesting capabilities. We do have the ability to put the data into... dashboards and so on so again like it so it's a very like visual product in many ways um so there's a lot you can do here and i can just say hey create i want to share analysis of 2024.

create a dashboard and send it to somebody. You can just say that much and it will just happen. Those are other interesting aspects of the product.

yeah for sure i wonder so like i guess like one of the questions is like how do you price such a complex product and um what about like the memory and context layer of this so for example each different company have their own like um governance um documents and then there is also like different type of access to uh you know I'm sure like if you're the CTO or like the CISO or like CIO they can monitor it very closely but I guess like

for the you know for i guess like for like a actual employee if i want to be like yeah i'm gonna generate like a graph for my sales team so i can report to the ceo like those type of like access and i guess like also the contacts layer and the memory layer like how do you kind of like uh so big scale of things our superpower uh

i don't even know how you came up with this question but this is like super spot on the the main value prop that we are building is essentially harnessing this enterprise-wide context this this is like this We are actually riffing on like terms for it. We called it a knowledge fabric once. Then we came up with universal context. The current hot name I'm going for is super memory. Because ultimately what you need is like this.

super memory or uber context which can be used to serve all the people in the organization so that they can rely on this like uber thing and this uber uber context is what is built by built by specialists or not just built I would say trained by specialists and it keeps improving through usage this is kind of a very very core uh you know IP sort of that we are building is how to

create this enterprise-wide context and continuously enhance it. Again, this is very important. Number two, Again, one of the big differences with like a chat GPT and so on is that when you're deploying at an enterprise, who gets to see what is non-negotiable, right? again like even if it's finance well then usually like you know two percent of the companies are even allowed to know what's happening with finance even if you're talking about sales

Imagine there is an organization with 1,000 people in the sales department. There is a whole pyramid of who gets to see what. There is the CRO who gets to see everything. But then if you look at different levels in the hierarchy, each person is getting to see the you know the what their team can see right so this is like again like from a you know technical

details about the product we absolutely you know very we take security very seriously this is like a very important property of the product you know we can handle hierarchical role based the ability to federate your identity as you're kind of going into these different systems. So there's just a bunch of like capabilities built in to make sure that there is no leakage of information from one person to another.

And again, all the while, you still want to make sure that there is an Uber context that is continuously being built so that this Uber context can benefit all the people who are consuming the product. Because again, the consumption model for us is very much based on usage. And the usage comes either in the form of number of people using it, alternatively, the number of actions that we're taking on it.

we absolutely want more people we this is very much not designed for you know this five person special game to use it is absolutely designed for like you know the thousand other people in the organization to use

So we need this ability to harness context and then allow lots of people to safely use it. I wonder how many people are your i guess like how do you define your icp so you mentioned like you know obviously on your website there's like you know dscope or which is like another speaker at our conference and then there is uh the father clean or like the glean team like father clean was also on our path

past like podcast guys and like how do you think about like you know i feel like those are like very um fast growing like startup even startup unicorns but like um for the you know for the thousand people team i feel like a lot of it were more towards like you know enterprise or you know fortune 500 like kind of traditional companies what's your go-to market strategy to capture these like

two different type of companies right like one is like the skill up and the one is like kind of like you know i don't know if you guys have like you know i'm saying like uh like a more i guess like a traditional

Great point. So I think the, what's a common theme? I mean, the common theme is, so again, so there is a way to look at companies that, you know, like as a business what you're doing and so on but then there is another layer which is okay as a data shop and i view every organization as something that has to deal with data

right so at that level where the commonalities are so i think our value prop is where data starts mattering now the question is when does data start mattering so if you're like a large organization data matters period um if you're a fast growing uh you know even like an e-commerce company that's moving fast data matters because they're just like extremely data if you have like complex supply chain it absolutely matters

if your sales are growing as fast as it is for clean data. It comes down to is data critical to your day-to-day decision making or not? which requires a certain velocity or it requires a certain scale so the places which does not make sense right so i would say small ultra small companies i actually think you know we are

I mean, they can probably use us, but it's not a great fit. In certain sense, there is a certain amount of scale that is required. I would describe as mid-market enterprise and above. is where our fit is most, not so much in really smaller organizations. So I wish I could just market to all YC companies and make business. I don't think that's the right GTM for us.

In terms of, since you're talking about GTM, one of the things that we are very keen to launch, and this is recognizing the... experiential aspect of the product you know like like when we talk about our product a lot of people want to kind of like you know touch and feel it it's like oh wow this happened no i just spoke to it and something happened

that's that's quite brings a lot of delight um and how do you kind of like make that more accessible to people so we are in fact like launching like a free trial you know version of our product this is something that we're building towards

which will just open it up to a lot of people to at least get a taste for it. But in terms of where we can earn revenue, I think there is a certain scale required. I would say mid to large enterprises where really... uh the the scale lies now the great news for us uh is that this this whole space is very i would say it's we are industry agnostic so again it could be

you know like patreon is like one of our uh you know patreon that's like a what would you call it it's a you know it's my very zeitgeisty uh you know a very digital native company uh and on the other hand we have like you know oil and gas and we have like ultra large tech like cisco so it's like a full spectrum the commonality is that data is important for sure um are they using

so when they're using the product so i guess like a patreon's data and uh like a cisco's data it's probably very different from the format so like patreon for people who don't know it's kind of like um a tipping company for your favorite creators or it's kind of like creator subscription company i think maybe they will have like you know video or audio blah blah blah but although the business data is probably living on like excel or something but

but on the other hand there's like you know cisco that like the data could be like varies right maybe i don't know if it's like you know their marketing department or their sales department or like just the whole company uses wisdom but um how do you think about like the processing different type of data whether it's like you know like different yeah so i guess like yeah ultimately uh

So Patreon actually turns out as one of the more complex organizations. They use Databricks extensively. Cisco, we are talking about, I mean, again, using Snowflake and so on. So ultimately, I think... even though the businesses are doing very different things the data has you know like data is tables and text and you know JSON and like

you know so at like the if you take one level deeper into like the technical side of what it means uh for data then it's there are like lots of commonalities right so ultimately it'll get into some

some sort of large database or it's going to remain in some file repository. That's the commonality. At the tech level, I think there are commonalities. I think where it gets very interesting is that the problems that you're solving are very different right so in one case patreon we're talking about you know growth and creators right or like you know cisco will cisco we're deployed within finance departments it comes down to you know like

managing payments to suppliers or something, or with the oil and gas customer, we are talking about manufacturing operations, right? So the business problem that you're solving is different.

but at the technology level there are a lot of commonalities the the key though is that this context that we are building is sort of like a marriage between this low level data and like high level business content the concepts right so that's kind of where the that's why like this like training like this uber context which is extremely extremely custom for each customer

is the crux of the problem that we are trying to solve. And this has to be done in the context of each business individually this is not some like magical uber context it's just like works across all businesses that just simply cannot work when it comes to analytics uh when it comes to like analytics um i guess like how do you set up your own company so like

um how many people are working on like the product itself and then how many people are working like you know selling it and then obviously you guys already landed all these like big logos like cisco or you know all these great companies on your website like how do you think about um I guess like both building out the product, but also like running the sales team as CEO, like how do you kind of spend your own time every day? Yeah. So yeah, my own time.

it's like already it's very chopped up uh well today started really well with this with this live stream yeah so as a company we're um so i would say first uh until like uh middle of the year uh it was like entirely like you know founder-led myself and my cook my other couple the two of us kind of focused a lot on gtm and then you know the rest of the organization was essentially engineering um so we started kind of building up the gtm team from middle of i mean like around june

So at this point, we have about eight-member overall GTM organization, including sales, marketing. I think you're close with also your... sound device again sorry like I think I can there's like a sorry about that about that noise yeah so I was saying that you know it was largely just engineering

And, you know, I'll count myself as overhead until middle of the year. But since then we have kind of like been scaling up our GTM team as well. And that's kind of reached about eight or nine people right now. and engineering is a little over 20. So the whole company is like 30 to 35 actually. There are like, I don't know the exact start joining dates.

so so i'm a little bit off on the exact numbers but that's the range we are in um that's actually very fast like growing since like you guys are in fairly like early stage like i guess like since um in the past like your role was more of like you know i guess on the uh chief architect um but although you're also the co-founder but like i guess like what

or some something that kind of like surprised you because and then your previous company that sold to facebook was like oh my god sorry um how do you think about i guess like transitioning yourself into the ceo role

And like, what were some things that you felt like you've learned from those experiences? Yeah, I was, I hope my team isn't seeing this though, like laugh at me. So. you know like as a as an engineer you think about hey somebody is using my product this is wonderful i'm so happy you're giving me some feedback that's great uh but as a sales guy i'm like well

i don't care whether i care you're using the product but i want you to pay money so so there's that so there is a difference between like uh you know people saying they're excited about your product or like trying it out

and giving you money for it um and this took me a little bit of time to kind of grok that difference uh um so that was like one one big thing that you know like it's uh like making people pay for your product is very different than making them feel excited about your product or giving it a spin. So we have, I think over time, I have gotten better at it, but more importantly, now we have folks within the team who are more specialized sellers.

kind of take us towards revenue quicker than i am personally capable of so that was like one big difference for me um again like i would say i i know that my my you know previous ceo people used to like obsess about uh messaging you know like this is like in a very precise form what the company does and i used to kind of never really pay much attention to it And I realized later on what critical role that exact positioning makes in the big scheme of things about the success of NE.

any venture because like you you need like a you know people need to immediately get what is it that you do and how it is going to connect to what they are looking for and if you cannot do that in a very effective manner. Even a very great technology doesn't quite go anywhere. So I think for me, it's been a lot of insane amounts of learning about.

you know how to position your product and like what does it mean to make people pay money so that's those are like in my own personal growth areas having spent most of my time uh as a builder in the past life uh speaking of paying money how do you price a product in the ai era uh so i think the Okay, so there is this aspirational thing that everybody always wants to do, which is like value-based pricing. Yeah. The challenge is, again, it's easier said than done, you know, like ultimately.

it's not enough to just say value-based pricing. You need to be able to show attribution. You need to be able to clearly quantify what value you're delivering. So this is like a journey for us. We are definitely

trying to get to that point. And I feel a year from now, I'll give you a different answer around how we price it. But today, there is a very more basic usage-centric model that we do, which either is uh you know based on number of users using it alternatively based on number of you know discrete actions that you are making um we are like trying out like a more value-based model in a few cases where

the exact business case that we are solving with analytics is more clearly established. So that's an area of active work for us. But currently, it's very much user and usage based.

um i was wondering in terms of i guess like uh one thing i wrote down was like attribution right so like how do you make sure like how do you even create these attribution yourself like so you know when you are selling to like a complex team so like it could be like marketing and sales all together you know for example for marketing maybe you run like a billboards ads and then you run these like you know

um other ads in like different category of things like how do you attribute the revenue to you know the marketing the sales or you know like all these kind of different things track attribution from like the ai tool so like obviously the ai probably you know help you generate like a better graph but how do you say that's like you know i'm sure it's like a lot faster than you know i go on excel or i go on like these like 40 different like data tooling to see the results but you know how do you

Not easy. I think there are a few ways to do it. It's like, okay, what's the alternative? How would you do it? If you had to answer, you know... this level of complexity questions and this volume, how would you do it any other way? The answer would... So you can easily translate it into saying that you're talking about... these many person years of work needed to kind of do this work and hence you should kind of pay all of that money um i think that's a so that that's a way uh i think the

Devil's advocate argument on that would be that, well, then it sounds like I should not ask too many questions because I'm creating more work. Maybe I was fine already. Why bother? So I think it's better to try to tie yourself to some clear, tangible business outcome. uh you know so there are business processes that we can easily tie into right so example you know there's a business review that is going on or like there is a you know

CFO has to close the quarter or like CFO has to go into an earnings call. CRO has to run their like weekly business operations. If you can tie into some of these operations that are, you know. very data heavy operations that every business needs to run to uh that's a way to tie in towards some like very very tangible piece of value this these are like some thoughts that we have um uh but again like this is like a it's a it's a it's not like uh

This is not like a customer success agent. Customer success agents are great. It's like you get tickets, but you anyway need to handle those tickets. It's very easy to quantify the number of tickets that you have deflected. you should just get paid for it in the case of analytics it's a little bit different uh and especially when you're trying to create like a category of like you know you have much more consumption than you would have done otherwise right

It becomes a little trickier. So I think it comes down to tying yourself to some business process or like some clear workflows that you need to do to run your business and essentially tie your revenue to those workflows.

i wonder like i guess like how do you think about uh like the fundraising milestones obviously you guys have raised from like some of the best investors and um you know to achieving i guess like the different uh set of benchmarks from you know c to a to b to eventually ipo like you know what are some lessons that you feel like you've learned from like the past experience to kind of set up

your own like milestone for each um um i i think uh i mean like we live do live in like a very we're like funding climate at a very peculiar stage in the technology cycle uh i would say right now there is an ai land grab that is going on right i think there is a you know and invest i mean there's a it's a little bit crothy on the investor side as well but what we have based on the money that we have raised and based on the

you know some of the proof points that we have with our customers even the market is just absolutely ripe for like disruption right now um i would say a year back people people when i say people i mean customers had a little bit of uh hesitation you know like oh this ai thing how good is can it be should i really invest in it or now i think this year there is no excuse it's it's almost like

like are you under a rock if you're not not like thinking about major changes in your organization there is something wrong with you right so i think people customers are essentially very well primed to make a leap

And we just need to kind of scale like crazy. That's really what the mandate is for us. I mean, I would love to go 10X next year if we can. And I just feel that the market has the... appetite for for solutions like us so again like not thinking about like a b c d milestone i'm simply thinking about how do we just like crazily grow our business uh what's your roadmap to like crazy leap to your business so uh i mean a few things uh again moving away from founder-led selling is a is one of those

leap moments. So very, very aggressively looking to grow on the GTM side. on the product side there are like some very specific unlocks i mean i talked about the trial free trial motion so that's i think going to play a big role um in terms of like enterprise data sources there are like a few very specific pointed pieces that we need to get right. I think that will kind of open up a lot of large enterprise opportunities for us.

I think on the marketing side, I would say largely we have remained under the radar. I mean, a big stream of things I think we could, you know, in terms of like putting our word out there. I think there is an opportunity to just 10x what we are doing today. So, I mean, it's like really honestly just on all cylinders is what we need to do. Love that.

well so i'm like uh i want to wrap with a one minute for you uh what's your favorite book what's my favorite book wow favorite book OK, this is a hard one. I'll say a few books I've liked. Again, I used to read books a long time back. There's a book by this guy called Richard Bach, Jonathan Livingstone Seagull. of it that's a very very powerful book uh it's like a very aspirational um that's one book that i found great um wait what is it called it's uh i hope i'm getting the right name right

Yeah, it's called Jonathan Livingston Seagull. Oh, that's a book's name? I thought that was the author's name. Yeah, I know, I know. It's the name of a seagull. Oh my god, what? what is it about it's this like so seagull is like a you know seagulls right it's not like a bird the bird that nobody likes it's like this uh you know it's this you know you just want it to be shooed away

But Jonathan Livingston, the seagull, wants to soar like an eagle. That's the story. This dude who nobody likes, but he really wants to step up. that's a book that has stayed with me for a while wow okay i'll totally check that out who uh who made the biggest impact in your career um career um i think uh i would say

Again, early on, a lot of credit to my mom. I think growing up, she was like a massive influence on me. But I would say Bipul Sinha, my CEO at Wistami, at Rubrik, he... you know i think definitely on the entrepreneurship side uh just learned a ridiculous amount of stuff from him so those i mean from a career perspective i would like give him a lot of credit love that uh who would you invite to your dinner party

Dinner party. Yeah, Grace, you're very welcome to join a dinner party. I don't know. Bill Gates? Oh, actually unique answer. I love that. I think like nowadays, like you know a lot of people just say Elon Musk because he is like always in the news but I feel like Bill Gates would be such a great guest because he also like accomplished like so much and then he said I feel like fairly low-key in the press yeah I think like for me like uh like from a

like vibing perspective i'll have a much easier time with it with yeah for sure uh who uh where can we find you outside of work uh i mean there is i think again like you know i have two young kids uh family is like a big part of my life i think uh outside of work that is probably quite all-consuming

yeah well so um thank you so much for coming to the show today it was such a master class for everything um analytics and i guess like a new way of like describing like your business outcome basically Cool. Awesome. It was amazing chatting with you, Chris. Thanks so much for having me on the show. Let me quickly.