Umesh Sachdev & Rajan Anandan on the Making of Uniphore as a Conversational AI Company, Pivots, and Building in the US Market

SHOW NOTES:

• At the start: A last attempt at entrepreneurship [02:06] 
• The pivot that rewired growth [06:34] 
• How Uniphore zoomed in on the US market [09:48] 
• Uniphore’s roadmap to finding its ICP [14:41] 
• The gravity of being multi-threaded [20:00] 
• Leadership should be stage-appropriate [23:56] 
• Strong networks: The key to enterprise sales [37:36] 
• Professional services & the moat of extensibility [38:58] 
• The art of pricing well [41:21] 
• What really seals a sales deal? [45:42]

TRANSCRIPT

Rajan Anandan: Uniphore launched in 2008 as a conversational AI company. Can you imagine that? Sixteen years ago, conversational AI, there wasn’t much conversational AI happening then. So I think one [achievement] is just, you know, staying at it for a very long period of time. Tremendous grit, amazing perseverance, what does it take to really win in the US market. 

And as you’ll hear from Umesh (Sachdev), I think, the trajectory change that you saw once you moved was just extraordinary, and we’ll talk about that. So, Umesh has built a very good company. I’ve been very fortunate [to have seen it grow]. But, you know, it’s also been great this morning, you know, when we started, I said the rate limiting factor of any startup is the founder or the founders. How quickly do founders evolve? And I think personally, it’s been incredible to see the evolution of Umesh and his co-founder, Ravi (Saraogi). I mean, these guys evolved very fast, and that’s been really wonderful to see. 

Umesh, why don’t you start by describing what Uniphore is today, and then I’m going to take you back to the early days.

Umesh Sachdev: Uniphore, today, is a multimodal AI company. We’re probably one of the largest enterprise AI companies in the world. We service about 1,600 enterprises in 20 countries; within those companies, 750,000 active users. Those are employees of those companies that come in every day to do their job at this contact center or sales, whatever they’re doing. They use a co-pilot or a bot or some form of automation that Uniphore has provided. 

When Rajan talks about us as an enterprise AI company, we are a full stack provider. Everything that starts on top of a data lake, which comes in the form of converting data into knowledge. Providing models as a service, and then creating several applications with those AI apps, whether they’re co-pilots and things like that. Today, Uniphore does all of the above. We didn’t start [with] doing that 16 years ago. But today, for some of the biggest enterprises in the world, that’s what Uniphore does.

At the start: A last attempt at entrepreneurship [02:06]

Rajan: Got it. Great. 2008, you started the company. By 2014, you had $300,000 of ARR (annual recurring revenue). Very fast ramp, please note. Six years to get to $300,000 of ARR. And then you went from 300K to what, in four years, to about $8 million. And then in the last six years, you’ve gone from $8 million to $150+, and growing very fast. 

Umesh: That’s right.

Rajan: So let’s now go back to the early days, like, 2008, how did you guys get started on conversational AI? And then what happened in the first few years where you were, if I can call it, wandering around the IIT Madras campus. Talk to us about the early days.

Umesh: Yeah, so look, coming into 2008, Ravi and I, my co-founder, had spent two years wandering about, trying to do another startup, which was in the mobile location-based space. And that hadn’t gone anywhere. 

We were fresh undergraduates [in] computer science where peers of ours… 2006 and 2007… [those years] aren’t the same years as today. There was no venture capital of scale. There was no startup at scale. So we had taken the path that many weren’t taking. Our peers at the university hadn’t taken that. We had spent two years getting nowhere. 

And when it came to 2007 to 2008, we decided to do one last attempt at entrepreneurship in that phase of life. We said, “If this doesn’t work, we’ll go back and do something else, like our contemporaries are”. And that’s how we landed at IIT Madras. 

IIT Madras was one of the only known incubators at the time in the country. We met Professor Ashok Jhunjhunwala, who became our mentor. And we wanted to pick something that we knew still had one in 10 chance of succeeding, where we said, “Having spent two years trying to do the other business, let’s pick an idea that we’ll work on, which if it does succeed, it better make billions of people impacted, billions of dollars of impact and so on and so forth.” 

And so, it wasn’t the conversational AI, it wasn’t the technology that we fell in love with. The thesis was that 70% of India, which is about 700 million to 800 million people, would not use the internet like all of us know how to use it, which was also the thesis that Google picked up later with its NextBillion project. And we said that’s not because the internet would not be accessible to people in smaller towns or villages in India. It’ll be literacy, English language, and the browser being the interface. 

And so we said, “Okay, if we were to unlock this internet for 800 million people, by just letting them talk to the internet in a language of their choosing, and have the internet talk back, then guess what? That qualified to be that big, billion-impact idea.” And so, this was not a technology looking to solve a problem. We identified the problem and then we said, “Now how do we do this?” 

There was no Apple Siri when we were talking about this. There was no Google Assistant. The only known speech-to-text application was a company called Nuance, which is now owned by Microsoft. And that was the era in which we were wanting to do vernacular Indian language, NLP (natural language processing), speech-to-text, text-to-speech. 

Much later, our then mentor, Professor Jhunjhunwala said… he played on the fact that we were truly ignorant and our ignorance was so blissful. We had no idea what we were signing up for! So to your point of what was happening in the first few years, it was a lot of painstaking, getting one language after another. It was keyword recognition. It wasn’t as slick as the speech recognition as we know today. We think of conversational as, being this very accurate, you can say anything and it transcribes. Those weren’t the days [where that happened]. To get one word recognized in Hindi or Tamil would take us three months of training a vocabulary of those words. And so the way we started getting off the ground in terms of business is having proven the early prototypes. We started going to businesses in India, banks, hospitals, etc., and saying, “Look, there’s a part of the population in India that is not using your app. That you don’t have a bank branch or hospital to support them. Here’s a new way to access a new type of customer base. Use voice, use speech.” It wasn’t conversational [AI]. Nobody was buying AI from us then.

The pivot that rewired growth [06:34] 

And so that’s how we got started until one of those banks said, “Hey, listen, this is great. This pilot’s going well. But we have another team in the bank that has heard about you all and says they want to work with you. Are you okay if we introduce you to them?” And this was American Express. And their call center team wanted to talk to us because they realized that they can use speech recognition, etc. They had some initiative in the US at that time. Honestly, Ravi and I, our first reaction was, “This is mission drift.” We didn’t start the company to do call centers. Why would we do it? 

Rajan: Yeah, right. 

Umesh: The only reason we got into that space is, I had the foresight of asking the question, “How much would you pay us? And they said 200,000 of NRE dollars. Which even at that conversion rate in Indian Rupee was like a year’s salary in Uniphore. So we said, “Forget the mission drift! You’re gonna do it.” And that’s how we started in call centers. 

Rajan: That’s amazing. And what year was that? You started the company in 2008. 

Umesh: This was 2012 to 2013. 

Rajan: Okay. So about three to four years later. 

So then what was the next phase? So you get to 300K, then you get to a few million. Did you get a few million, just selling to like call centers? 

Umesh: Yes. So once we had decided or found the original product-market fit (PMF) in call centers, we were becoming a product company from a technology company that has a bespoke set of applications for the technology we had. We had started to productize the technology, build applications for call centers. That is the time the first institutional cheques, Indian angels, yourself, and several others, came into the company. 

And I remember a lot of people, once again, I feel old while saying this, when that investment took place, immediately after we had got a couple of million of some revenue and then Series A [from] Chiratae Ventures, which used to be IDG (Ventures), came in. And those investors had put a condition in the company and that condition was telling. What was the condition? They said, “Look, this is a DeepTech play. We haven’t seen too much DeepTech out of India, this is interesting. We’ve heard good things from our customers, but there is a business model flaw in the way you do business. And some of our learnings from Silicon Valley or America are selling to enterprises: there’s a new business model emerging. It’s called software as a service (SaaS). So we’ll only give you a series A if you convert your business to software as a service.” 

So up until 2014, we were selling on-prem (on-premises) licenses, maintenance dollars, etc. And once again, I was like, “Okay, how much is Series A? Three and a half million! Okay, I’ll convert to software as a service.” And so that was the next big pivot, which incidentally became our first near death experience because to take an on-prem business to SaaS, even at a couple of million dollars at scale, wasn’t trivial. 

But once the shift was complete, once the whole company got rewired, product, customer success, how we sell, etc,., then it was a very sharp growth. And I think, from that period on, we can say that the real Uniphore was born, which is about 2016.

How Uniphore zoomed in on the US market [09:48]

Rajan: Right. When did you know you had product-market fit? 

Umesh: I will say right around 2016. We knew contact centers, and there were, like, the number of customers, every conversation, we were hitting the right message, BPO (business process outsourcing), these big call center operators. 

Our original pitch was, “We’ve called into your call center, and we’ve heard the script that says every call is being recorded for quality and training purposes. How many quality auditors do you have in the company? And how many calls in a day can they listen to? What if I could tell you that you can now listen to 100% of the calls at less than the cost of those people.” That was product-market fit for us, which in today’s world is called speech analytics or international, interactional analytics or whatever. So we knew around 2015, 2016 that this message was working. We can now replicate. 

Rajan: So speech analytics for call centers, that was the [product]. Call centers were the ICP (ideal customer profile), speech analytics was the product. And then you knew, as soon as you pitched, that you had customers saying, “Yep, that makes sense.” That’s when you knew… 

Let’s now move to maybe the phase where you moved to the US right? Talk a little bit about what happened before and after you moved to the US. Why did you move to the U.S.?

Umesh: So, I will say a couple of years before I moved to the US, it was very clear in our minds that for us to be big, North America had to be a big market. A lot of our key customers, their headquarters were in North America. We were only picking up the pieces that the Indian counterparts were allowed to innovate in India, etc. But we never had the courage, never had the confidence. What does North America look like? One had visited Silicon Valley a few times like these, many trips like this. But we didn’t know what that looked like. 

The big pivot point happened in 2016, 2017 we were raising our Series B and John T Chambers, the then chairman of Cisco, former CEO of Cisco, met me in a conference. We got together. There was chemistry between us. And then he decided to invest in the company as part of the Series B round, and he joined the board. 

That was a big pivot point, because when I first discussed with him that, “John, I think North America should be the most important market for us. We have very little business coming out of North America. Should I hire somebody to run it? What should happen?” 

And then he says, “The answer’s obvious. You need to be here.” And basically, I remember him saying, “I’m your safety net. Just come. If it doesn’t work, I’m here.” And that’s why I picked Palo Alto to be our office. He lives 10 minutes from here. [We] stayed very close to him, used his network originally. So, him coming into the company was a big pivot point in giving us the confidence that ‘we need to do it’.

When I came here, the company had, give or take $8 million of ARR (annual recurring revenue). The last valuation of the round that we raised that he had come in at was $34, $35 million. But also, I had no more than six months of cash in the bank. 

Rajan: When you raised [series] B or before you raised the B? 

Umesh: This was after we raised [series] B. This was probably a year and a half after the [series] B [fund raising] that I was moving. My wife didn’t know that. I had convinced her, I had a young daughter, we moved, I didn’t tell her the small fact that the company has less than six months of cash left. I said, “It’ll be good. We’ll go, whatever.” 

Rajan: So when you moved, you had six months of cash. Okay. That’s interesting. 

Umesh: But by the way, Rajan, even back then it wasn’t the era where a lot of Indian founders were coming to the valley and succeeding like we have success stories today. So I’d met two people, and I’m glad I met those two people. One was a big success story of an Indian founder moving here and succeeding, and one was a story of what not to do.

The first one was Jaspreet (Singh), founder of Druva. And his story was, “Umesh, listen, when I moved, I put the house, I bet the house on this US move, we had $8 million or $9 million of rest of the world revenue. We shut it down. And I moved to the US and I came to the last six months of cash, and had this not worked, the company would not be around”. Druva is pre-IPO. We’ll hear good news about them at some point.

So, having those lessons from founders like me who tried this journey prior, having John Chambers as an asset, as a backer knowing what I knew about cash in the bank situation, knowing that we’ll have to start spending aggressively in the US to build the team, create a brand. The US is not a market for the faint of heart. This is a very different market. I still took the plunge, and I’m glad I did.

Uniphore’s roadmap to finding its ICP [14:41]

Rajan: Got it. So now let’s talk about … so when you moved, you had $8 million. Today you’re at $150 [million]. It’s been about six years. What really made the difference? So one, you moved, what did that enable you to do? By the way, we were talking about this, today we would recommend that founders move much earlier, like one or two million of revenues as soon as they know there’s some level of product-market fit. I want to get your thoughts on that later, but tell us what happened in those first few years after you moved. What worked, what didn’t work?

Umesh: So I will say for me, the first six, eight months, I have very vivid memories of those [months] because it was hard. I landed here knowing I have to start raising money. But I was also very clear, now that I made the pivot, this is a point of no return. We are going to be a US company. So I wanted to raise money from Silicon Valley-US investors as a company coming out of India, no brand recognition in the US. At that point, there were no customers in the US but I wanted to raise money in the US because I needed the brand value from the US VCs. 

Simultaneously, I was the only employee in the US, so I was the only sales guy, I was the only SE (sales executive), etc. And so I had to go acquire customers very quickly so that the VCs would give me credibility. And one of the things, one of the attractions of the US for me, was the talent. The talent who would take this company forward, the talent who had, in Silicon Valley, already seen six, seven, eight generations of enterprise companies succeeding, etc. 

And so, I was recruiting, I was selling, and I was raising money all at the same time. And I think by the first six, seven months in, we had raised our series C, $50 million. March Capital had led it; several others had joined. I had our first couple of very key hires. Our CMO (chief marketing officer) is still, [the person] who joined me back then. She’s still our CMO in the company. We had other executives who came in, who over time, went on to do other things. And just those three things happening, the talent coming in, a well known VC backing us in the US, and big brand customers becoming referenceable. 

Rajan: So basically, moving here, super critical. Raising capital from a US VC gives you the brand, the networks, etc. And then third is building the team. And then enterprise selling is… you’re absolutely right. Like, you could have also gone and sold to mid market. Maybe you could have sold to SMB (small and medium-sized business). Where did this whole… and you’ve been consistent with this, you’ve been so clear about your ICP. Where did that come from? We’re only going to sell enterprise, like where did this come from. And why? Was it hard to just stick with it? Because closing multi-million dollar deals is just not easy.

Umesh: Sixteen years is such a long time, that every mistake that you guys can collectively think of, I’ve probably made in this company. So while Rajan says I’ve been so true to my ICP, etc., there have been three times, not once, not twice, three times in this history of 16 years, where I’ve tried to move away from my ICP. 

Rajan: Oh! I didn’t know that. 

Umesh: We’ve been elephant hunters from the beginning. Our product-market fit has always been enterprise. Where we’ve got insights about what the customer wants has been an enterprise. Our products are built for enterprise. Our sales motion is built for enterprise. 

Three times, and all three times, a new CRO (chief revenue officer), a new RevOps (revenue operations) leader, a new somebody comes in and says, “Umesh, we need velocity. These sales cycles are too long. We need velocity. We need some run rate.” And not once, not twice, three times I gave in. I said, “You’re right! This is too hard. It takes me eight months to land a million dollar deal. If you can land a 100K deal in 30 days, please do it.” 

All three times it has never happened. We’ve sold 100K deals. It’s taken us eight months. And it’s taken us another eight months to implement. Same as the enterprise. That’s because of our product-market fit. Our product was never built for mid-market. CX (customer experience), the first space that we were selling in, now we sell many other things. Contact centers. Anyone who does contact center business, just take it from me right now, mid-market is a mirage. Don’t waste your time. 

Rajan: What do you mean? 

Umesh: There are 15 million contact center workers. 

Rajan: 15 million? In the US or globally? 

Umesh: Globally, 60% of them are with BPOs, 85% of the 15 million workers are not working in a 100-seat contact center. They’re working large contact centers. They’re working large contact centers, 1000 seats. The largest BPO in the world has 700,000 seats out of 15 million. The second largest is Teleperformance, 500,000 seats. Why would you waste your important time, valuable time, chasing a 100K deal in a market where the ICP just makes sense? 

If your product-market fit has occurred in the contact center, it would have occurred at the large enterprise. Said another way, that’s always been where we found success. And the sales cycle is similar. The cycle to convert ARR to revenue is similar. So our learning has been that our ICP is a large enterprise. And we are whale hunters. And we’ve got to structure ourselves to do that.

The gravity of being multi-threaded [20:00]

Rajan: Got it. And what advice, I think probably a third of the room is selling to large enterprises, what advice … I mean one, show up in person. It’s human to human. You’re an AI company, but humans sell to humans. What are some of the other pieces of advice you would give the founders in the room who are selling to enterprises? It takes time. Eight months. You gotta put up with it. What are the other pieces of advice?

Umesh: Look, the entire go-to-market motion for an enterprise is very different. And it took me a long time to accept this, because when you start reading the textbook on SaaS, there’s so many people in the [Silicon] Valley, in India, who now understand SaaS like there’s no tomorrow. Wall Street understands SaaS. Every analyst understands SaaS. But the vast majority of commonly available knowledge about SaaS metrics, KPIs (key performance indicator), are generally built for mid-market or SMB type go-to-market motions. A lot of that doesn’t work in the enterprise. 

Rajan: What doesn’t work? Give us some examples, bring it to life. 

Umesh: So, think about demand gen. In an enterprise, who are you selling to? Usually, it’s a lot of top-down selling, right? Even if it’s not a C-suite selling, it’s a VP or below selling, because those are the people who have budgets, who have the ability to buy. And to approach them, they’re not answering an inside sales call out of India. They’re not on LinkedIn trying to look for a solution, they’re not doing Google SEO (search engine optimization)-type search, right? So it starts with an A, B, everything approach, account-based approach. So you start by being so… And Ankur (Saigal) was the one, where we back in the Capillary [days] did this SaaS E (enterprise SaaS) initiative in India. And all of us for the first time were like, “What is SaaS E? Why does SaaS need a suffix?” Well it does, because it’s different. 

Rajan: SaaS E is for enterprise? 

Umesh: Yes, enterprise. So you start with meeting your customers where they are. At events, at steak dinners, at golf events, etc. So you create opportunities to spend intimate time with your customers because you need the ice breaking, you need the trust building, etc, while the sales process is on. Then you think about your sales teams, your AEs (account executives), etc. Where should they be based? Should they be based in low cost locations where it’s all remote selling? No. Because you can’t sell a large multi-million dollar deal to a large enterprise without having met them in person. Without having met several different players. You have to be multi-threaded. 

A big mistake that even within my team people make is they become single-threaded, “Oh, I have a fox, I have a champion. They promised me they’re gonna buy it this quarter.” Guess what happens? They never buy. Because you had not mapped the CFO (chief financial officer), you had not mapped procurement, you had not mapped CIO (chief information officer). Your champion was a business user who said, “I love your software.” But in the enterprise, [it’s] like selling to a triangle, you’re selling to the CXO (chief experience officer), simultaneously selling to the CIO. And the CFO, a lot of traditional sellers actively avoid CFO and CIO in the process, like, “Oh, they’re going to slow me down.” But guess what, if you do not map them, you cannot forecast the deal. 

So the type of sellers you need to have, need to have the muscle memory, need to have the instinct, need to want to get on a plane to go meet the customer every day. They should not be happy if they did not get on a plane up until the Wednesday of the week, etc. 

And then customer success, same thing. So the entire go-to-market needs to be wired to go after this multi million-dollar elephant if you’ve decided to be an enterprise company. Now, you still need metrics to track progress, for sure. But the velocity is very different from a mid-market play. The ACV (annual contract value) is very different from a mid-market play. Your lifetime value. CAC (customer acquisition cost) should never be even a topic of discussion if you’re selling to an enterprise. Because how much CAC can you spend to sell to an enterprise who gives you a lifetime value of $30 million.

Rajan: Oh, so you don’t measure CAC? 

Umesh: I mean, we do, but for the longest time, it’s irrelevant… If I spent $200,000 attracting a customer who has a lifetime value of $10 million, it’s a no brainer.

Leadership should be stage-appropriate [23:56]

Rajan: Got it. Okay, let’s maybe just shift gears a bit. So, talk a little bit about how you think about teams for stages. Talk a little bit about how you’ve thought about refresh, like the team refresh. 

Umesh: Yeah. So I’m trying to think, I don’t even have a good scientific answer to tell you what version I am at. But what I will tell you is philosophically a few years ago a switch flipped in my head when I became very comfortable with the fact that leadership teams and teams in general need to be seen as stage appropriate. When you’re in the 0 to 10 million journey, your line of sight is two years out and even if you break the speed of light record, in two years, you’re gonna be $30, $35, $40 million. So you’re bringing on people in your team with that vision. 

Well, guess what? You’re likely to be faster than you think to that outcome if the business is really humming. And then when you hit $20, $25 million, you’re like, “Look, my next stage is gonna be $60, $70, $100 million. I don’t think the people I got on board…” And then there’s this emotional tussle, “Oh they came. It was such a big process to bring them on. They’re finally settled. It feels like a great chemistry between me and whoever. But I also know at the back of my mind that they may not be able to scale to $100 million. But if I give them a lower role now, less than their title, it’ll hurt their ego.” So most CEOs, including myself, drag that decision for too long. 

So the learning is, every leadership position, up until a certain point, where the growth starts to stabilize etc., needs to be seen as stage appropriate CFO, CRO, CMOs. And one learning, I was giving out the C[-suite] titles very easily early on. If I were to redo this, I wouldn’t do it. 

Rajan: No. What would you do? 

Umesh: Start with VPs (vice president). VP of sales, VP of whatever. And that, by the way, is okay at the $10 million stage, at the $30 million stage. People understand. You’re not a company who can afford a real… 

Rajan: At what level would you bring  in a CXO? $50 million? $100 million? 

Umesh: I’d say $50 million. Beyond $50 million is when…

Rajan: How many heads of sales have you had?  

Umesh: Countless. 

Rajan: I know, but like how many? Because this is important, right? 

Umesh: More than 10. More than 10. For sure. 

Rajan: In the last five years since you moved to the US? Six years, how many heads of sales? 

Umesh: I would say, five to 10. 

Rajan: So, you’re churning the head of sales every year. 

Umesh: So here’s my rule about heads of sales in the US. A lot of Indian founders come and meet me about, “I’m trying to come to the US. I’m trying to set up a US GTM (go-to-market). What should I do?” And here’s a rule of thumb. I say, “First, hire your VP of sales. Don’t give them the CRO title. And be prepared. If they’re very successful for you, you’ll be firing them and changing them in a year.” And that’s counterintuitive to a lot of people. What they’re saying is, “If they’re really successful for me, why will I be rotating my VP of sales in a year?” And the answer is, “When you just land in the US… I was there six years ago, very fresh memory, you yourself as the founder have not figured out your go-to-market motion. So the chances are the VP of sales that you really need is not a sales leader. You need a mercenary who can run through walls. You have no brand value. Your product market fit is weak. You need somebody who can just take whatever you’re giving them.” And because of their Rolodex, or relationship, or their hustle, can land you the first two, three big logos. By the way, when you land those logos, now is the time you need to replicate the playbook and scale the sales team. That mercenary is no longer suitable for that. So it’s so counterintuitive, people don’t understand that [when] you get your VP of sales, don’t think about bringing somebody who’s led a hundred-people sales team—you’re not ready. And if they’re massively successful, then you need a hundred-people sales leader and be very comfortable with this process.

Rajan: I think this is important to keep in mind, this idea of a state specific, refreshing your leadership team. It’s hard, by the way, right, because all of you get very attached to the teams.

Umesh: By the way, I will say, if anyone here, like you said, is in the enterprise business, not mid-market, make no mistake, the head of sales of that company for now and forever, is the CEO. The lesson that John Chambers still teaches me is, “Till the day I retired [with] $50 billion of revenue, I was the head of sales, make no mistake.” 

I’ve seen Michael Dell do it. I was at lunch with him and Kris Gopalakrishnan, former CEO of Infosys; Kris is also an angel investor. At lunch, Michael Dell pulls his plate aside, pulls out a laptop and says, “Kris, have you seen my new laptop?” Michael Dell! All right. So for enterprise, the CEO has to be very comfortable in their skin that you’re the lead salesperson.

Rajan: So look, I mean, what’s happening? Like, how much of it is real? You’ve got $150 million of revenue. Is it all AI? What is the state of AI adoption in the enterprise today? And by the way, you’re now doing more than… you’re doing CX sales, you’re doing multiple functions, you’ve got this knowledge layer. But tell us a little bit about the state of AI adoption in the enterprise.

Umesh: So look, clearly, when the history book on AI will be written, it’ll be a pre-GPT (generative pre-trained transformer) era and a post-GPT era. The adoption of AI in enterprise has been happening for years, if not decades. It didn’t start when OpenAI made that, or Microsoft made a big investment in GPT and lit this thing on fire. The way the enterprises have adopted AI prior to GPT was that AI was part of a point solution in different business applications. Our sales team needs Gong. Gong has AI. The CX team needs Uniphore. Uniphore is AI. Marketing team needs Drift, which just got acquired by Salesloft. That is AI. So AI wasn’t this feature that was being bought; it was being bought as part of a business outcome application, as part of vertical SaaS. Post GPT, AI has shifted to become an infrastructure topic that the CIO now owns. That is a huge shift. 

And that’s really important for enterprise sellers to wrap your head around. Because AI is now a boardroom topic, it’s the CEO’s priority. And now that has given the license to the CIO to say, “All of you business leaders who bought all these SaaS applications, please stop. I control AI spend now.”

Why is that? Because the CIO is now trying to wrap their head around AI regulation, data privacy. Is it a third-party model? Should we train our own model, we have a lot of data, how are we going to do it? What’s the long term IP (intellectual property) of my business, etc? And the thinking emerging is, AI needs to become a horizontal model for the company, whether it’s one model or three or four different types of models. And then all applications will leverage these models that the CIO has decided so to do. Which means Gong or Uniphore or everyone else who built their own models to support those apps, that architecture is out of date at this moment, okay. 

In this new world, the state of adoption of those models is: I sense two emotions at the same time when I’m selling. I sense a panic, of ‘I need to make fast decisions because Wall Street’s expecting me to have the answers by the next earnings.’ But by the way, every 48 hours, I’m reading something new. So how can I make long term decisions? How can I go to my CEO and recommend a $300 million spend on Microsoft or Google or whatever when things could change?” And that’s why you’re seeing a bunch of pilots, a bunch of experimentation going on in the enterprise.

Rajan: Open versus closed. The question is, do you use Open AI or Gemini or do you leverage open-source build on that? What’s the framework to think about that? How do you guys think about it? What do you do?

Umesh: Well, I think in general, the answer is it depends on what you’re trying to achieve. What’s your moat, et cetera. We are an AI company. Our moat, so our belief is these models and we have our own model, by the way, for the application that we are good at, CX, sales, etc. We’ve trained our own multi-model. However, we believe firmly that in the next three years, these models are going to get commoditized. Open source, Open AI, Gemini, open source is catching up to GPT-5 really fast, so on and so forth. So the big moat in AI is not going to be in the models. It’s not going to be with data science, scientists monopoly, etc. It’s in data. Most of these models, open source or otherwise, are being trained on publicly available internet data. So anyone who has a strategy of having access to exclusive large quantity data that cannot be had, bought, or stolen. That is a real example. 

The reason Elon Musk, on every earnings call, comes out and says, “I’m lowering the price of my car, because I don’t care about margin per car, I’m an AI company.” What does he mean? He means the more number of drivers he has on Tesla cars in the world.

 Rajan: The more data he gets.

Umesh: The more data he gets. The more data he gets, because when I drive a Tesla, when you buy a Tesla on tesla dot com, there’s a box you’ve got to tick that you allow Tesla to use the data. And if you don’t tick that box, Tesla doesn’t sell you the car. 

 Rajan: Oh, is it? Wow. 

Umesh: So, that company is after exclusive data that only its network of cars is producing. No other car company ever will have that data. And Tesla has decided that’s the big moat. The real moat in AI is exclusive access to data. 

Rajan: And you think that, whether it’s functional applications like sales or marketing or CX (customer experience) or healthcare or retail or financial services, payments or fraud that there’s enough of a moat you can build on data as an AI company.

Umesh: 100%. 100%. Not only on that data, but because… Look, once again, a misconception in the enterprise is that GenAI (generative AI) is the silver bullet that’s gonna replace everything we use. No. GenAI to the contrary has to work with the current legacy IT stack. So the data of the domain, the data of that industry that you’re serving, plus how GenAI is being implemented with the current legacy ID stack, all those are unique data moats that general purpose GPT-4, Gemini [which are] out of the box… They’re not coming trained for those situations. Those are being sold with SIs (system integrator) like Accenture coming in and saying, “I’ll now fine tune this model. For an insurer with their IT stack”, and that’s like an 18-month project, which is why the pace of adoption in enterprise is low. 

So anyone who can still SaaSify that and say, “For insurance for this application, I have a data moat and therefore the model I produce, whether I use open source or my own, doesn’t matter. I shrink the time to value. I shrink the time you need from implementing an AI and going to an earnings meeting and say, ‘I have business outcomes.’” Anyone who can do that right now will still have a moat.

Rajan: Ok, cool. Let’s open it up. 

Audience: How do you get to know all the CIOs? Like, I also sell to the CIOs. So how do you..

Rajan: Yeah. How have you gotten to know every CIO in the country? That’s an interesting question. 

Audience: And then the second question is given that I’m not from this country also, but I’ve lived here for a long time, but like just the sales motion of being an engineer from background and from India, how do you overcome those challenges.

Umesh: Look as a stereotypical Indian computer science engineer, I was an introvert. I am an introvert, okay. But at some point, it was very clear to me by mentorship, by training, by exposure, that relationships, 50% of my sales happened because of a relationship, 50%. So $75 million out of 150 million. 

Rajan: Really?

Umesh: Yes, and when you recognize that superpower, when you recognize that’s important to you, then as founders, you … we learn to hustle to get the… and be resourceful to whatever you’re trying to solve. So when you decide, having those relationships is important, when I landed here, I was pulling on straws. I was going to IIT (Indian Institute of Technology) alumni and foundation people and saying, who do you know? I was going to John Chambers and saying, who do you know? And then whoever they were introducing me, I was saying, who do you know? That’s how you start. Today, of course, you have a team, you have sellers, people, Rolodex, you have investors, you’ve attended enough conferences, you’ve been on stages enough, you start investing in your own PR (public relations) at some point, to show up at the right conference, keynote speeches. All those are not tools to boost your ego, they’re tools to show up where the target individual that you’re trying to get to is paying attention to you. And you get very deliberate about it. And by the way, when you get their audience, then you got to be… you have to give something to them so that they come back for the next meeting. You have to add value, you have to say, how can I help? You have to have a listening ear of what’s your key problem. And then you can say, “Listen, I have some ways of helping you. I may not be ready to sell my product yet, but hey, if you want to talk to John Chambers, I can make that happen. As bartering whoever, if you want to meet Rajan Anandan, then I can make that happen.” And then I’ll call Rajan and say, please do me a favor. So you invest the time, and you be deliberate about building a network of CIOs. And then at some point, when you start selling, there’s referenceability, then people talk to each other. 

Strong networks: The key to enterprise sales [37:36]

Audience: We are co-founders of Dozer, which is an enterprise product for basically simplifying data movement and data access, real time data movement and data access. So you mentioned one of the first deals you landed was through entity Data. So do you feel, for a company that is just starting out and starting in enterprise, motion partnerships could be a good way to start? 

Umesh: Once again, I’m going to say it depends. But in your case, are you enterprise? Are you mid-market? What’s your… 

Audience: Enterprise. 

Umesh: Enterprise. Sometimes you build CI (customer intelligence) relationships by being through sh*t with them, tough times, you show up, you stabilize, and those relationships come out stronger on the other side. 

But for enterprise, for early stage, it’s a myth that, “Should I do partnership? Should I go direct?” There’s no strategic answer here. You take, you meet whoever is willing to meet you. That’s enterprise. If the CEO of a partner will meet you, go sit with them. If a CEO of an enterprise will meet you, go sit with them. At an early stage in enterprise, you’re not building a playbook that will get repeatable this way. You’re building relationships and networks, to the point that he said. You do it however you get it. 

Professional services & the moat of extensibility [38:58]   

Audience: And another question I have, I guess you have encountered [this] as well. Selling to enterprises, many times they have a lot of customization requirements. It’s very hard to replicate the exact same product in multiple enterprises. How do you cope with that in the beginning? 

Umesh: Just the way you cope with that when you’re much bigger also. First of all, if enterprise is your ICP they are not buying software off the shelf. No enterprise ever did. If they did, you’re being bought in a division of a department, not at the enterprise level. And you might strike half a million dollar deal but you’ll be churned out in a year. Because, somebody who bought you does not have the authority. The people who have the authority to buy enterprise-wide need custom. You take that into account into your product design itself. 

Extensibility is such a big moat that you can give yourself in the product. Knowing that customization will be in every customer that you sell. The bigger the deal, the bigger the customization. Professional services (PS) is a good word in enterprise. 

Rajan: Ah, say more about that. But you charge for it, right?

Umesh: Of course, because enterprise is custom. Because they need white-glove [services]. 

Rajan: But you charge for it. 

Umesh: Yes. Now, in the early years, we also subsidized it. But as soon as possible, just feel very confident to charge for it. Customers will be willing to pay for it as the scale comes in. What are the largest enterprise SaaS companies in the world? Salesforce, ServiceNow, Oracle. Not only do they have large PS (professional services) they also on top of PS, they  have consulting. They do transformation. They compete with Accenture for their own projects. All right? And that’s not a wrong design. That’s just right for the enterprise. That’s what your customer wants you to do. And if you’ve chosen enterprise as your ICP, you service them the way they want to be serviced. 

Be very clear on your ICP. I’ve made all those mistakes. Whoever will listen to me, I’ve been the guy who will show up, but if it’s not an enterprise, don’t waste your time. Because you’ll end up burning more cash behind an account like us if you’re built for enterprise. Because we don’t have the LTV (lifetime value) to give you. 

The art of pricing well [41:21]

Audience: Hey, so we’re building Toplyne. Our product lets you predict anything on your existing customers. So which free ones will pay you, which paying ones will pay more, and which paying ones might stop paying you. We sell into large SMB, early mid-market companies, that’s the ICP that we go after. As you were selling to enterprises in the US, how did you even get the gumption to figure out what you will charge them? So how did you figure out where you should start the pricing and how did you climb your ACVs (annual contract value) over time? 

Umesh: Yeah. So first, you didn’t ask me this, but the obvious thing is, if SMB and mid-market is your ICP, for the longest time, don’t even sell to enterprise. So the reverse advice to… 

Audience: I’m not trying to sell to enterprise.

Umesh: Very good. Pricing itself is a whole different universe. Product marketing needs to have muscle. So you do pricing, let me give you highlights, you do pricing by competitive information, you obviously take your cost into account. You usually are in a category where something similar exists, if not exactly the same, and you have pricing information through that. And then if you’re really slick about it, you think about the ROI (return on investment) that your software is delivering. And you come up with … so pricing itself is a big art, and it’s a lever that can increase or decrease your sales velocity. 

So pricing is a very important subject, but in a deal like that, of course we had never sold a deal of that size at that point, but we knew unit price, we had competitive information, and we knew if we were going at this price, we were still under what a larger competitor of ours was selling, and that was fine.

Rajan: But in general, we find that early stage companies, this happens to even later stage companies, they’re just completely underpricing because it’s easy to get the business. 

Audience: I have one more question. 

Rajan: Yeah, go ahead.

Audience: Enterprise is typically a top-down sales. Now, there are companies that are approaching, as I would say, bottom-up. 

Rajan: PLG (product-led growth). 

Audience: Yeah, but they’re clearly targeting enterprise. We are in the data space. I think about a product like Redpanda, for example, Kafka replacement or temporal workflow management, typically an enterprise product, but they have a PLG motion and I guess a top-down motion as well. I’m curious to see how you see this and how you think the two motions can be combined together.

Umesh: So in developer tools, in data tools, etc., PLG, bottoms-up works. But that might be the only isolated part of the enterprise. In no other part will PLG work, and even if it works, someday the diktat won’t come from the top. 

There is a tool in the enterprise that if you’re a startup selling to enterprise, you should be very off. And it’s called an ELA. What is an ELA? Superheroes, like Rajan, at Microsoft and Google come in and sell this enterprise licensing agreement. Okay. This is an all-you-can-eat buffet. Tens of millions of dollars get committed, with a quarterly penalty if the enterprise does not consume that ELA by a hyperscaler. 

And there comes a day where the CIO starts to call all the business owners, CRO, CMO, say, “How much money are you spending on those SaaS tools, by the way? What if I tell you that doesn’t come out of your budget anymore?” Because the CIO is under pressure to consume the ELA from the hyperscaler. The hyperscaler, you can say, may not have a competing product to you, but the hyperscaler has an app store, which is being thrown in as a tool to consume the ELA. 

And that’s why in an enterprise, it’s not enough to win a division of a department of the enterprise. I come back to” it’s a top-down sale. Developer tools and data might be the only exception, which historically has worked, where, because so many developers start to use a certain technology, the CI or CTO never come down hard on them and say, “replace everything”. But everywhere else, I cannot imagine a salesperson, a CX person, deciding to use their own software someday.

Rajan: And they won’t be allowed to, right? 

Umesh: They won’t be allowed to. 

Rajan: And especially now with AI problems. 

Umesh: InfoSec, data security, all that will come in.

What really seals a sales deal? [45:42]

Audience: What is the forcing function to close the deal? Is it just timing you stay in touch and you get subbed in at the right time? Or is that a forcing function to accelerate that pace? 

Umesh: So, at our upcoming sales kickoff … There’s this book by the way, that came out in 2023 when a lot of enterprise companies were missing their quarters, it’s called The Jolt Effect. 

Rajan: The Jolt Effect. 

Umesh: A very nice book, The Jolt Effect. 

Rajan: Ok.

Umesh: And basically the crux of the book is the biggest threat to a seller is not competition, is not whatever, it’s indecision. The customer is just not ready to make a decision. Your end of quarter means nothing to the customer. So there are lots of things that you create to create compelling events on the customer’s side. 

If nothing else works, you create FUD – fear, uncertainty, and doubt. Say, “If you don’t do this, you’ll be fired. If you don’t do this, your competitor’s gonna work with me,” you do all of the above, but the biggest thing you’re up against as a seller is indecision. And it’s a human emotion. 

The best thing you can do, but it’s also the hardest, is find a compelling event on the customer side. They have a budget event. They have to deliver a certain outcome by a certain date. If it’s a partner, their end customer is asking for your product now and they can’t say no, etc. But those are very hard to achieve. So at the last 48 hours of the quarter, it comes down to your sales force. And that’s how you get the customer to sign.

Rajan: I think the only thing I’d add is for early-stage companies, that sucking sound is really important, right. Because it’s an early-stage company, every customer will entertain all of you. You’re all very smart, pretty good storytellers. You’ve got some interesting products. But the real question is, is there a real need you’re trying to solve? If there’s no real need, they will just keep talking to you because they like meeting you. So I think you have to just be very careful. Okay, I’m just gonna ask one last question and we’ll wrap up. If there are other questions, can they reach out to you? 

Umesh:100%. 

Rajan: So it’s been 16 years. How many near death moments have you had? The company, not you.

Umesh: I would say, six or seven times legitimately. 

Rajan: Give us one or two. Give us two, bring it to life. 

Umesh: So two key ones: One, I mentioned, when we converted to SaaS, Series A happened, $3 million to $3.5 million, I don’t remember the exact amount, we were approaching $3 million of ARR (annual recurring revenue). So when somebody said, convert to SaaS as a condition of the round, I said that makes my life as a seller even easier. Instead of selling a $200,000, one crore deal, I’m selling EMI (equated monthly installment). Of course I’ll do it. And what I didn’t realize is it’s not an EMI. It’s a whole rewiring of the company, it’s how we structure the type of people. So that was a near death experience. Now it went down like this, but when it worked, it came up like this. So that was 2016.

Rajan: Yeah, that’s pretty good. 

Umesh: I will take that all day long. The second one was, this US move. It’s all glamorous now and all that and John Chambers, but I came with six months of cash, and I was recruiting for an expensive C-suite, VP, this, that, and the others, without mentioning this to anybody. And had the round not closed, had my customer not happened, given me the credibility, had the investor not moved in, another near death experience. So, am I used to living on the edge? You can say that, which is in my personal life and never take risks because I get all my thrills at Uniphore.

Rajan: There you go! Okay. Awesome. So tell us a little bit about, like, where does the confidence come from? How do you gear the machine to land the plane?

Umesh: Getting forecasting accuracy down right is one of the hardest things. And by the way, the most important things before you go public. Predictability is the most important thing. It’s not scale, it’s not whatever people think. You need to be of a certain size to go public. The minute you’re predictable, you can go public. It can be a small company, large company. 

There’s a lot that goes in. Pipeline, demand, conversion metrics, sizer deals, ACVs, you’re playing with all. Sales capacity, at some point it becomes a mathematical equation for the CFO, not the CRO. When you know it’s working, you start to play on sales capacity because you’ve got the art down to, “My reps are getting productive in so much time. It’s predictable, so I’m gonna add it before the first day of the financial year.” 

So in Q4 (fourth quarter), there’s a big ramp of hiring of salespeople. Sales, demand gen, marketing, et cetera. You start to build that process. But at the end of the day, as somebody who’s very involved in the sales process, my CFOs, the CROs, they do all of that, I focus on deals. I focus on what are those three $30 million opportunities that if my sales team does have a soft landing, then how do I come in and build that cushion? And so getting predictable forecasting for a year out is a very hard thing to do. 

But I think given where we are, we’re living the dream. AI is the one subject that there’s no dearth of spending right now. 

Rajan: Okay with that thank you. It’s great to have you. Congratulations on all the success. Also, thank you for being very transparent. I think hopefully you all took away, hopefully there were one, two, three, four, five things for all of you to take away. But Umesh, we wish you all the very best. 

Umesh: Thank you! 

Rajan: Let’s give a round of applause. 

This transcript has been edited for clarity.