Siddhartha Ahluwalia 0:45

Shivaku, welcome again on the Neon Show podcast.

So glad to have you back again.

Shivakumar Ganesan 0:48

Thank you so much for having me back here. You’ve called me for the second time. So I presume I say important things.

Siddhartha Ahluwalia 0:55

Yes, the first one went really well, right. And now I want to start, you know, by if you can give a recap of what Exotel is, what category you bring to our audience, and then we can maybe dive into the journey after COVID of Exotel first.

Shivakumar Ganesan 1:08

That makes sense because we’ve already recorded prior to COVID, I think in the previous podcast we’ve already covered. So we’re in the business of helping brands connect with their audience, their customers. So we deliver technologies using which companies can communicate, engage with their users, audience, customers.

So we are operating in the customer engagement, customer interaction space. So it sort of manifests itself in a few ways. So we are a MVNO, mobile virtual network operator.

We are also the largest call server in India, other than the telcos, and most likely in the top 10 in the world. We process almost about 2 billion calls a month. So if let’s say your delivery executive is calling, your driver is calling, sometimes the teacher and the student is talking, doctor and patient is talking, there is a very high chance that all of that is actually powered by Exotel.

So this is what we call as a platform business, sometimes CPaaS, communication platform as a service, sometimes cloud telephony, sometimes cloud voice. So these are all the different terminologies that we use to describe this tech. Then through an acquisition of a company called Ameyo, we also now got a contact center software, and we made this software integrate on top of the platform.

And then now we are serving a lot of enterprises using that contact center software. Now, just to put some perspective, people usually get confused about what a contact center software is. They think it’s a BPO.

We are not a BPO. We don’t provide the bodies or the people that actually chat or make these phone calls. Think of contact center as the OG software.

If you were to imagine, let’s say in the 60s or 70s, there was a time when phone call was the only way in which you could communicate with people remotely. And so contact center was a software that was built for managing phone calls. That’s where it started.

And nowadays people tend to think of CRM as something that everybody knows, but CRM actually sort of came up with the internet, and it is synonymous with email. But contact center software has actually existed even before that. So that’s what a contact center software is.

So it works with the CRM in many cases. Sometimes it’s the primary owner of voice. Sometimes we can also bring the voice into CRM through our APIs.

But these are the several ways in which contact centers and CRMs kind of work with each other. But suffice to say that all of this is in the CX space, customer experience space. We usually work on use cases like sales and onboarding, collections, service calls, customer service calls.

Then there is another segment called RM Connect, Relationship Manager Connect. You know, think of it as the person who will now, your broker who’s helping you trade or your savings bank account, relationship manager from the bank. So these are some of the other use cases.

I think these are the broad use cases that we operate in. Our AI journey also started through an acquisition of a company called Cogno AI, which used to be a leading brand in giving chatbots and a few other products in the BFSI space. So that was a starting point.

Of course, this is a pre-generative AI era. And then over time, we’ve also now built our own agentic stack. And now we offer voicebots, AI generated quality assurance, agent assist.

So these are some of the products that we have. But broadly, you can think of it as a set of all technologies in the customer interaction, customer engagement space.

Siddhartha Ahluwalia 4:52

So just to recap the journey, Exotel started roughly in 2011-2012 kind of time frame. 2012, you raised like 2.5 crores from Bloom and Mumbai Angels back then. And then the next eight years, there was zero funding for the company.

COVID hit, you were at 10 million then, and suddenly the revenue dropped to 5 million. I remember that. And then A91 invested in the company.

And I think that was a pivotal moment for the company once because also what happened is during the COVID, that the rise of calls through Exotel and the category became very prominent. And then there was, once the money started pouring in, you raised, I think, $100 million in a matter of two to three years. And then you also acquired two companies, Ameyo and Cogno AI during that period of time.

So describe the journey after COVID. You also shared that there was a bunch of layoffs during that period. And how was the journey coming till now?

What ballpark revenue range you are at today?

Shivakumar Ganesan 6:08

Absolutely. So what happened during COVID was, of course, because of the lockdowns, we actually had a lot of reliance on third-party logistics, e-commerce, cabs, food tech, all of that, of course, came to a grinding halt. So naturally, all the calls that these companies were making also stopped.

So our revenues became half. And like you rightly mentioned, we were a self-funded, profitable bootstrapped company up until then. So then that gives me no choice but to cut costs because I don’t have the money to run the show anymore.

So we had to almost let go of 30% to 35% of our employees at that point in time. It was just very painful because first of all, there is COVID. And then on top of it, you are now also asking people to go.

And of course, we tried as much as we can to keep them because times were uncertain. We don’t know how people will get a job, when they will get a job. So that was a really tough time.

But even during that time, I and my team were actually extremely busy because several governmental organizations wanted a number where they can now give information about COVID. Many NGOs wanted ways to disseminate information about where food is available, where work is available, and those kinds of things. And oftentimes, we actually had to do some of these things even without asking for money.

So those 21 days, those one and a half, two months actually were one of the busiest periods of my life because I was just helping. And this was my way of contributing back to the society. So that’s what we were doing and the team was doing.

So we were actually working very tirelessly. So then post that, what happened was then it became clear that everything’s going to be digital. So people who were buying from physical stores obviously now started buying online.

Payments were becoming digital. So naturally, remote communication became quite important. And so our SMS volumes, call volumes, all of those things started skyrocketing.

And that’s really when A91 sort of came in. And I would say that it’s a combination of right place at the right time. So I used to jokingly say that people usually ask me, why is it that you’ve been building this business for so long?

And did you not ever think of, I don’t know, just shutting shop and going home? So I tell them that actually, there are two issues. One, I don’t know how to quit.

I don’t know what that involves. Like what exactly is the process where you, okay, now today is the day when I’m going home. Like you are the one who started the company.

How can you go? So that’s one. I don’t know how to do that.

And secondly, I said that if you waited long enough, you can even see a supernova happening. And in a way, something like that happened for Exotel. So I think we were just waiting for our moment to arrive.

And I think that moment arrived then. And so that was one part. And the second part is, of course, we had to take advantage of it.

So like I was saying, we actually were working even harder through and during COVID. Even though our revenues were down, you know, we had to let go of people and things like that. But customers are actually coming after us and asking us for ourselves.

Our services became even more mainstay. Even very large banks immediately wanted to do work from home, immediately wanted to make calls remotely. And so all of that happened.

So I think that’s when A91 came in. So we did round one, round two. Actually, we did four rounds in a period of about 18 months.

Just in the year of 2021, I did two acquisitions and three rounds. And it was quite an exciting time. The story of Ameyo itself is also fascinating.

It actually did not happen during COVID. It actually started even before. So one of the largest use cases of a cloud-wise platform is contact centers applications.

And so I went to my team and I said, look, we need to now build this. Because now we are doing a few use cases. But this is actually one of the largest use cases.

And my team came back and gave me a two-year time frame for building out that product. Partly because we were self-funded. And so there was only so many engineers we had.

And so then I didn’t have the patience to wait for that long. So then I started speaking with Sachin Bhatia. They were actually thinking about how to now do a little bit of cloud voice.

They had internally started their own CPaaS equivalent and things like that. So I think it was sort of like, both of us sort of agreed that the world is moving towards the cloud. Contact center will move to the cloud.

For that, the voice platform also has to be there. And so the synergy was quite obvious. But the transaction itself was, of course, much harder than what we anticipated.

Both the companies didn’t really have, at that point in time, large backers with a lot of money. So we had to go and raise, which means we had to now convince somebody. But we hadn’t raised for eight years.

They hadn’t raised for almost four or five years. Growth rates were, if you remember last time, I also spoke about, you know, there are companies that say, go big or go home. This 40% or 35% growth rate is a difficult place to be.

Because it’s neither big, you’re not growing earth shatteringly well. But it’s not bad at all, right? Who doesn’t want 35% growth?

Go ahead.

Siddhartha Ahluwalia 11:34

So I just wanted to ask one more thing during this time. How much revenue once, you know, and you can share the process of this acquisition. How did it happen?

At what round? Like did it happen immediately after A91? Or how much revenue did Ameyo bring in?

Shivakumar Ganesan 11:49

Sure. So where I was going with this was, we were actually having these conversations even prior to COVID. And Sachin and I took a proposal to basically say, we will do an equity swap without actually raising any money.

All the founders were aligned. But unfortunately, we weren’t able to agree upon what’s the valuation at which we will do the swap. And after a little bit of back and forth, I realized, actually, both of us realized, you know, there’s no way this is happening without money on the table.

So then that conversation stopped. And then they went their way and we went our way. Then, of course, I went ahead and raised my first mini down with A91.

And that gave me a little bit of confidence that, okay, so now we have some backers. And then so I proposed it to our investors. They were pleased with the story.

And then by that time, what had happened was Ameyo had already started speaking to Route Mobile amongst other players. And Route Mobile had given them an offer and they were actually almost going to acquire. And in fact, the other co-founder in Ameyo was in favor of that deal.

And so I was just walking around in my balcony one day and I thought, should I call him? Should I not call him? So I called him up and I said, look, why don’t we restart the conversation?

And then he said, it’s too late. We are already, you know, almost going to sign the papers with Route Mobile. So then I was like, okay, what do I do now?

And then he also sort of jokingly said, hey, you’ve raised like 40 crores so far. Do you really think you have the money to acquire Ameyo? So then I was like, that’s right.

So what do we do? But I still didn’t stop. So I still pursued.

I called their investor. I called the other co-founder. Then I said, I’ll come to Delhi and I’ll catch up with you guys.

So I went there and I spoke to them. So fortunately for us and hard luck for Route Mobile, eventually somehow I grabbed the deal out of.

Siddhartha Ahluwalia 13:41

Did you make a better offer than Route Mobile?

Shivakumar Ganesan 13:45

We did make a better offer, not so much in terms of, so basically there were a lot of riders in the Route Mobile deal. So you had to make so much profit, you know, the payment was staggered and things like that. In our case, it was a very clean, here’s the money, take it and go.

Because in my case, actually I was looking for product synergies a lot. Whereas in Route Mobile’s case, I think it was more of a M&A inorganic strategy. So I think the 2 plus 2 greater than 4 is far higher for us than it was for Route Mobile at that point in time.

So I think that’s really what swayed Sachin also. Eventually we matched the offer. Maybe it was a better offer in terms of, you know, how quickly they got the money and how cleanly they got the money.

But the offer itself was sort of in that ballpark. I think that sort of swayed the deal. And last minute somehow we got the deal and life’s been glorious after that.

Siddhartha Ahluwalia 14:34

And how much, if you can share roughly, revenue did it add to Exotel after the M&A?

Shivakumar Ganesan 14:39

So I think they would have probably brought about 100 crores of revenue. And we were at about 120-ish crores of revenue at that point in time.

Siddhartha Ahluwalia 14:50

This is 6 months after A91 kind of.

Shivakumar Ganesan 14:52

Correct, correct, correct.

Siddhartha Ahluwalia 14:54

Until it bounced back to the original revenue.

Shivakumar Ganesan 14:56

Correct, correct, correct. So we went to almost 50 crores and then we came back to almost 120 crores. So that was the broad contours.

And so that happened. And so I was obviously speaking to a bunch of other investors at that point in time. So then we had to do a series C to fund this acquisition.

Then there was another series D. And then naturally I got more excited and I said, okay, next we have to bring in AI into the mix. And that’s how Cogno AI happened.

And then so that was series D. Of course, Cogno AI was a much smaller startup. But Ameyo was actually a very interesting case study.

In fact, I believe there has been a Harvard or some one of these MBA case studies that was written on, I have not seen it. And Avinash also had written a long post about this. So the interesting piece is that actually these two companies are almost similar in size in terms of revenue and profits.

And in fact, to tell you the truth, Ameyo actually had double the number of people compared to Exotel. So it was a very weird, is it a merger? Is it an acquisition?

Who’s buying? Who’s selling? Sachin had actually flipped over his stake.

So he’s actually our co-founder now.

Siddhartha Ahluwalia 16:05

Yeah. So he didn’t take cash out from the team. The rest of the co-founders took cash.

Shivakumar Ganesan 16:09

Correct. So it’s a very interesting deal. And I would say that after about three years, okay, so two interesting points.

My team originally had said they’ll take two years to build a contact center. Actually, we took three years to integrate Ameyo contact center into Exotel.

Siddhartha Ahluwalia 16:24

Was the revenue stayed intact during that journey, the Ameyo revenue?

Shivakumar Ganesan 16:27

So that’s another very interesting conversation. So in most acquisitions, what happens when an acquisition happens is that the revenue actually tanks because new people are coming in, processes are changing, there is uncertainty, culture is changing. And so naturally Ameyo’s revenue went down.

And then at some point we realized, oh shit, what are we doing here? We need to support our customers. And then I start focusing on it.

Everybody starts focusing on it. And then the revenue starts coming up. Same exact thing happened with Cogno.

So there was a point when all our customers were shouting, culture was changing. Then I had to start getting my hands dirty. Again, it came up.

What I didn’t anticipate was that the Exotel business also actually went through the same S-curve while we were doing all of this.

Siddhartha Ahluwalia 17:13

So you came back, came down from 120 crores again. Correct.

Shivakumar Ganesan 17:17

Because we were now so focused on trying to make contact center work, trying to make AI work. Our original business actually also sort of went. So I think there are three S-curves that sort of went.

Fortunately, they are spaced out a little bit. So it didn’t all happen right now. But right now, of course, all of them are in the uptrack.

Also because we’ve now sort of unified the culture and people and the systems are also integrated. So all of that’s happened. So right now it’s all sort of going on.

Siddhartha Ahluwalia 17:39

So 90% of the M&As don’t work out historically. Tell us, how did you make it work despite of these S-curves?

Shivakumar Ganesan 17:50

So, see, actually, I don’t know why these things don’t work. Maybe, I don’t know. But I’ll tell you that I was actually quite clear why.

So this was not like I’m going to add inorganic revenue or I’m going to acquire customers. None of that. Right.

So there’s actually very clear product synergies that I thought will add value to customers when we did this. Right. So I think and that is exactly what we ended up doing also.

So in that sense, the technology synergy was very, very strong.

Siddhartha Ahluwalia 18:17

Did you want to sell the Ameyo product to your existing customers?

Shivakumar Ganesan 18:21

Both ways. So, I mean, ultimately what we wanted to do was to be India’s first CCAS. There was no, I mean, there was no CCAS.

Technically, even today there is no CCAS. CCAS is contact center as a service. Think of it as SaaSifying contact center product.

So as I told you, contact center business existed even before SaaS came in.

Siddhartha Ahluwalia 18:41

Yeah.

Shivakumar Ganesan 18:41

So what is the equivalent of contact center software business in the SaaS world? That is CCAS. So obviously in the US, people are buying CCAS.

So in India, people had to still buy some wires, talk to telcos, put up some boxes. This was the way people are setting up contact centers and we wanted to change that. And it is happening.

And so in this sense, Exoto is actually a very unique company. So we have bits and pieces of our business competing with other companies. But on the whole, what Exotel offers, actually nobody else offers.

And this was by design. And this was a differentiation that I wanted to bring to the table for our customers. And so I would say that it hasn’t been easy though.

So there were lots of churns, culture mismatches, clashes, fights, people leaving. I actually making a number of mistakes without understanding why things are the way they are. I changed some processes here, some people there because it didn’t make sense to me.

So all of those mistakes were made. And I think some of this is responsible for those S-curves. But I’m happy to say that last quarter, we actually booked the highest ever contact center orders in the history of actually both the companies.

So certainly it’s working now.

Siddhartha Ahluwalia 19:56

Congratulations.

Shivakumar Ganesan 19:57

Thank you. So now having said all of this, I think I wasn’t, I hadn’t done any acquisition before this. So I only thought of what are all the things, the way it’s going to work.

I never sat and wrote down what are all the risks and how it’s going to affect us. But many of them came to affect us. So sitting here, I would say that acquisitions are not easy.

It’s very hard. Only because I had the founder energy and because it was my idea, my and Sachin’s idea, I think we gave it all to make it work. Otherwise, it’s actually non-trivial to make some of these things work.

So now I feel like unless there is a really good reason, I would probably not acquire randomly anymore because it’s a lot of work.

Siddhartha Ahluwalia 20:44

Can you tell us why it’s so hard to make it work?

Shivakumar Ganesan 20:48

So, I mean, ultimately, of course, people is a thing. But it’s not just that, you know, it’s not like I’m just saying that, you know, there are two sets of teams and then they have to come together. It’s not just that.

I think the amount of learning curve that everybody has to go through while continuing to support both the businesses and think about how the together world is going to happen into the future. All these things actually take a lot of time and old habits die hard. So Exotel at that point in time was a 10, 11-year-old company.

Ameyo was a 14, 15-year-old company, right? So you go and tell somebody, hey, you know, why are you doing this? They don’t know.

You are now going and telling them, okay, change it like this. Then they are like, no. So it’s very hard.

It’s very hard to change people and their ways. That’s the reason why I think it’s hard. What’s interesting is that our margin profile has actually improved over time.

So consequently, our profitability has also improved over time. And this has happened because of a little bit of a change in the mix of business that we are getting. So I think during and through COVID, we sold a lot of SMS and WhatsApp and that kind of stuff.

Some of these things actually aren’t that margin accretive. But now we are seeing a lot more AI, a lot more contact center. Of course, voice is our mainstay business.

Now this is a little bit more margin accretive. And so we may not have actually grown that rapidly when you talk in revenue terms. But actually, we have been giving almost 20% type CAGR growth in the last two years on the gross margins.

And so consequently, we are now again profitable.

Siddhartha Ahluwalia 22:15

And can you tell us more about what the core AI products are that you are offering to the customers?

Shivakumar Ganesan 22:22

So actually, there are several things that we are doing with AI. Usually, in most other companies, when they talk about AI, there are some internal use cases that they focus on. Naturally, they will get people to write code using agent, which we are also doing almost.

If I’m not wrong, about 20 to 25% of our code is now, even today is actually being written by AI. Yesterday, I got the happy news that for the first time in our history, my CTO told me whatever was supposed to release in October, they are going to release it in August. I’m like, how is that possible?

This has never happened. And of course, the answer is AI. So some of that is happening.

The second thing that most companies will also end up doing is enterprise search. Think of it as like a glean-ish. So how do we now get access to information about the company in the hands of all the employees?

So we are working on top of Gemini for that. So that’s happening. So these are the two internal facing things that we are doing like every other company.

But we are actually impacted by AI in a much larger way than other companies because we are in the CX space. As you know, customer service, contact centers, these are the places where AI’s impact is the largest and one of the first use cases that people are sort of going after.

Siddhartha Ahluwalia 23:36

People don’t want humans to make call. Now, AI agent can make end-to-end call because they can understand on the context.

Shivakumar Ganesan 23:42

That’s right.

Siddhartha Ahluwalia 23:43

What other human is saying, whatever language is saying.

Shivakumar Ganesan 23:45

Correct. Absolutely right. So we actually have to now redesign, rethink what our products are going to be in the future itself.

So naturally, we started with some of the basic things. So there’s a protocol agent assist, which now uses AI, crawls the knowledge base. And then as the conversations are happening, either on chat or voice, then there are pop-ups that now says, hey, here’s the answer to the question that the customer is asking, etc.

So that’s called agent assist. Then there is a product that we have called CQA, conversational quality analysis, which itself can be actually used in multiple ways. But one standard way in which people are using it is, has the agent now greeted the customer well, or has he or she now done these 10 things that there’s checklist that they’re supposed to do.

Earlier, people used to actually sit and listen to these calls, regulatory reasons, SOP reasons. And they used to cover about 2 to 5% of the calls, which itself was very expensive. Now, obviously, AI can now listen into all these calls simultaneously.

And then the implications are actually slowly, you can now start thinking what more you can do. So actually, voice of customer or customer feedback is no longer necessary. You can now run generative AI on all these calls.

And you will automatically now know. Earlier, you used to ask for CSAT and NPS. All of those things are not required.

You can actually pretty much build your product roadmap by just listening into these calls. And we can get AI to now build the product roadmap for you. The MIS is the control center that let’s say the executives want to see.

What is our customer saying? What is our customer thinking about our brand? All of that can actually be gleaned from the customer calls.

So there are several of these kinds of use cases that we are now powering using our CQA product. Then there is the voicebot, which is the one that everybody keeps talking about. So we have deployed one of our voicebot agents, collections agent.

About two, three months ago, I got to know that it has actually collected almost 3 crores of rupees in a month for one of the NBFCs. Completely automated. So just take a minute to just think about it, right?

Completely automated. No human is involved. There is a software that’s now helping collect 3 crores in a month.

I think that was just… When I heard it, I was first of all very happy. And secondly, it’s fascinating that it’s actually…

I mean, we only wrote the code. I know what it does, but it’s still hard to believe.

Siddhartha Ahluwalia 25:58

But the interesting thing is, the AI you are implementing for your customer is eating your own business.

Shivakumar Ganesan 26:06

Correct. So it’s a little bit… We are and we are not. I’ll tell you what I mean by that.

So it turns out that… Okay. So broadly, I think the total number of contacts and receipts will come down.

My estimation is that it will come down by about 80% over the course of three to five years. By the way, timeframes are actually anybody’s guess. In AI, things are changing every day.

So God knows. I’m just telling you my intuition right now. Now, what however happens is that…

So let’s say roughly a contact center agent is about 30,000 rupees. You’re paying that person a month. The company will now…

When they replace that seat with AI, they will want almost 70% of that to flow to through their profits. So it’s not like you can now offer AI and expect to actually charge 30,000 rupees a month because why will they now pay the same 30,000 rupees to you? They have to also have some benefit, right?

So this is what the world generally calls a salary tamp. Now, so let’s say 70% of that is retained by the company. Let’s say Unilever.

The remaining 30%, which is about 6 to 9,000 rupees is what is actually available for software or AI. So in this 9,000 rupees, I think almost 60% of it will go to foundational models, LLMs, think OpenAI, think Google, think cloud for hosting, all of that. So the remaining 40% is what will be available for applied AI companies such as ourselves.

So we are not a foundational research company, right? We are an applied AI company. And so roughly about 2,000, 2,400 rupees extra you’re getting out of that subsegment back to us.

So if you do the math, although the number of seats is reducing to almost 20%, because you’re getting this extra, let’s say roughly 2 rupees a minute per conversation on AI, it actually ends up becoming a larger deal than where you started. So you start with 100 seats, you go to 20 seats, but you do AI conversations, your bill will now increase from, let’s say you start with 1 lakh MRR, you’ll actually go to about 5 lakhs MRR. So we find ourselves in a really interesting situation.

It’s a very exciting situation because contact center seats are declining, but actually Exotel’s prospects become even brighter going forward.

Siddhartha Ahluwalia 28:26

Okay. And how much percentage of you believe that the current contact centers that we have will be completely replaced by AI in the future?

Shivakumar Ganesan 28:35

See, there’ll be these new processes that people are now starting off, which I think it is possible for people to imagine it completely using AI. But some of these older processes, I don’t ever think they will become 100% replaced by AI. You will have people who will be needed for complicated questions, for overseeing the bots, for teaching the bots, for compliance purposes, sometimes to take over a conversation when the bot is going rogue.

So my anticipation is that we are looking not more than 80% reduction over time. 20% I think will sort of continue to remain.

Siddhartha Ahluwalia 29:11

So for example, let’s assume 1 million people work in India in business outsourcing or business processing outsourcing units, right? BP0s as we call it. So you think only 20% of them will be left after some period in time?

Shivakumar Ganesan 29:26

That is correct.

Siddhartha Ahluwalia 29:27

So a lot of job reduction by AI?

Shivakumar Ganesan 29:29

A lot of job reduction by AI. I mean, that is somebody basically said that you will not lose your job to AI. You will lose your job to somebody who’s going to use AI.

And I think that’s kind of true. But I’m actually pretty excited about the future. It’s not at all.

Some amount of reskilling is required. Some amount of soul searching is required by everyone because a new tech is upon us. But suddenly, I don’t know if you’ve noticed, there is a new wave of engineers who are now being called for which is called wipe coders. And the good news is that wipe coders don’t actually have to be engineers. They don’t have to be computers.

You don’t need to know how a CPU works to be a wipe coder. So actually, all of us can be wipe coders. And then wipe coding is actually so much in demand right now.

You may have seen that Zuckerberg is paying $100 million packages to approach people from open AI and things like that. Of course, that’s for research. So there will be reskilling, there will be new opportunities that will open up.

So I anticipate that there’s going to be a massive demand in data annotation and data labeling services. You may have noticed that again, Zuckerberg paid almost $30 billion if I’m not wrong for a company called scaleAI. There’s going to be eval services.

How do I know if my bot is working fine or not? There’s going to be LLM ops, there’s going to be InfoSec on LLMs. So there’s going to be a whole new variety of software stack, AI stack that is going to get generated. And so different kinds of skills will be required.

So it’s all going to be exciting.

Siddhartha Ahluwalia 31:03

And how much do you think this global market is being captured by India in the things that you mentioned?

Shivakumar Ganesan 31:08

So I don’t know about global market being captured by India. Because as you know, there isn’t a lot of foundational AI work that’s happening in India today. Of course, that’s fast changing.

But I also think that it may be practically impossible for the global AI players to capture the India AI market as well.

Siddhartha Ahluwalia 31:28

Why

Shivakumar Ganesan 31:31

So see, if you roughly do the math, I told you that my anticipation is that we should be offering AI. In this case, I’m talking about voice AI at about 2 rupees a minute.

I think this is the rate at which the Indian businesses will be able to actually purchase. Now, while it is practically possible to offer a 2 rupees a minute even today, actually there is no zero margin left for the vendors at this rate. So now you’re expecting at least 40% margin on top of it.

But even at the current inference costs, I think that the savings in the West is so high that there will be there is actually no motivation for American companies to further reduce inference costs going forward. So I think they’re going to just forget about this cost problem and then they’ll move on to AGI and I don’t know whatever else that they’re focused on. So about a couple of weeks ago, I wrote on LinkedIn that I do believe that India’s time in AI is now.

I think it has arrived. It is arriving. And it is not the kind of AI work that we think it’s not.

So mostly people are obsessed with where is India’s LLM, where is India’s ChatGPT. This is how people are sort of thinking about it. But I think the real AI challenge in India is how do you deliver 2 rupees a minute in the case of voice AI.

And the same is applicable for images, same is applicable for chat, for cancer research, everything right. How do you offer 2 rupees a minute voice AI at 40% margin and the AI should be able to speak 30 languages simultaneously very well. I think this is a real challenge.

This is a uniquely Indian challenge. And this problem, I think India is going to have to solve for itself. And hence I believe India’s AI time is now.

Siddhartha Ahluwalia 33:16

I think global companies have distribution advantage. I’ll tell you a case study. Perplexity offered is one year perplexity free for all Airtel users.

And Airtel has 360 million users in India. But today, perplexity is the number one app on the app store ahead of open AI. So global players have that kind of distribution advantages.

Shivakumar Ganesan 33:39

No, actually, by the way, I’m only bringing in B2B perspective here. So I’m not tracking B2C side of AI that much. I don’t know what’s going to happen to search and things like that.

But on B2B side, actually, whatever I’m saying is going to be true. My sense is that the innovation is going to be keeping the quality of AI intact and then further dropping the inference costs and be able to do this in 30 more languages very well. I think this is the AI challenge for India and I think the B2B AI challenge for India, if you will.

Siddhartha Ahluwalia 34:14

And one interesting thing that I observe is that you have always been focused and been bullish about the India market. You kept on building for it. Right.

And now you see the result like 65 million dollars of ARR, which is completely opposite to what is getting preached in the Indian ecosystem that go to the West. Your first customer should come from the US.

How do you drill down this model to enter that India? You can build a 65 million ARR.

Shivakumar Ganesan 34:45

By the way, I don’t actually recommend that anybody else should do it because due to a weird set of circumstances, I ended up in this or Exotel ended up in this position. I’ll tell you what I mean. So the first one is actually I started roughly around the same time when Girish also started Freshworks.

So what is now, of course, being called as a Freshworks playbook is what I also had in mind when I started. Although at the back of my mind, when I started the company, I said, hey, I want to serve the Indian market. I want to serve the society around me.

These were the energies with which, of course, I started. And then in a very uniquely, unique Indian problem, of course, is calls. And so I sort of stumbled upon that problem and I asked myself if computing can move to the cloud, can communication not move to the cloud.

So these were the initial hypothesis questions that I was asking when I started. Now it so happens that the Indian telco architecture is actually quite different from the rest of the world. And so when we ended up solving for the India architecture, it’s not directly replicable to the rest of the world.

So A, it is different and B, there was no economies of scale in the architecture that we had solved for in the US or somewhere else. And also Indians talk a lot due to various reasons. So I would imagine that India probably makes per capita more phone calls than anybody else in the world for various reasons.

There is cheapest telco minute in the world, vernacular problems, language problems. There’s only so much English people know how to read and write. But I think these are all the reasons.

So for these reasons, Indians may end up making a lot many more phone calls. And so it’s a real problem here. Companies want to reach people on phone calls.

And so it’s a very uniquely Indian problem. This is not the case, for example, in Saudi Arabia or in the US where it is mostly customer service calls, inbound calls. There isn’t a lot of sales that happens on calls and things like that.

So the applicability of the solution is not that big as it is in India. So I think for these various reasons, it ended up playing out the way it did for us. I also think that if you’re building CRM, if you’re building SaaS, if you’re building software, you should go to the US.

In fact, one could also argue that if you’re building AI, you should go to the US as well. But our basis here, we have existing distribution, we have relationships. So when we are building AI, we are now saying, how do we build AI for India?

Just building on top of what we already have.

Siddhartha Ahluwalia 37:20

And why didn’t in Exotel journey, you ever thought of expanding to the US market or European market?

Shivakumar Ganesan 37:26

For the same two, so a couple of reasons. One, the architecture, our strength, our core strength is not necessarily replicable in the US. So the architecture in the US is different.

So some of the scale modes, the regulatory modes, the economies of scale that we had here, you can’t just take the software and put it there and then I mean, you can do it, but anybody else can also do it. And so and they are doing it almost about, there should be about a dozen companies in the white space in the US.

Siddhartha Ahluwalia 37:56

Twilio being the leader.

Shivakumar Ganesan 37:57

Twilio being the leader. And so I never really found a reason what exactly, in fact, the real thing is that I was actually quite impressed and inspired by Twilio. And so, in fact, as a tribute to Twilio, I hadn’t even told any of my colleagues this.

So I actually wore Twilio’s socks when I went for the 14th. So we celebrated our 14th foundation day a few weeks ago. So I was very inspired by Twilio’s work at that point in time.

So anyhow, so these are the reasons why we felt like there was no real angle. Now, one thing led to another, one thing led to another. So we just became larger and larger in India.

So now, every once in a while, I will now go to the US, spend a couple of weeks in the valley. Then there’ll be a couple of ideas, I’ll come back, but somehow it just never really takes off.

Siddhartha Ahluwalia 38:49

So, and what is the distribution of revenue between India and other geographies that Exotel is present in today?

Shivakumar Ganesan 38:54

80-20. So about 20% of our revenues comes from Middle East, Southeast Asia, Africa. So we focus a lot on emerging markets.

Our business in emerging markets is growing faster than our India business.

Siddhartha Ahluwalia 39:06

And what is your thought process on listing on India markets? Because certainly in the last three, four years, India has become one of the most attractive markets globally for listing. And a lot of companies are now flipping back, let’s say MoEngage is flipping back, CleverTap is flipping back, Capillary just flipped back and has already filed the DRHP.

Shivakumar Ganesan 39:26

So this is one of the things that actually played out to our, we were lucky that I always started in India. I always wanted to be in India, stay in India. So I don’t have to spend any money to flip back to India.

So we are fortunate that way. So no, it’s a great time to be in India. Markets are growing very well.

Money is flowing in. Liquidity is there. Backers are there.

Indian investors have understood what it means to support software businesses. So it’s a great time to list. I think that that time will come for Exotel also.

We are a very unique company, like I said, very hard to replicate. No one’s going to, you know, it’s not an easy business to just start off and then compete with Exotel. So I feel like companies like Exotel deserves to be public.

Siddhartha Ahluwalia 40:12

For retail investors to share the game.

Shivakumar Ganesan 40:14

Correct. It needs to be a long lasting institution. So I don’t think of it as much as what it means to me to go public as much as I just feel like Exotel is one of those assets that needs to, deserves to be public. So we will take it public.

So I think the growth story is still I think the Exotel’s growth story is, there is a lot of stuff to happen over time. I think we are in the first two to three years of, I don’t know, 20 year mega cycle of AI. So it’s imperative that Exotel executes AI really well, which hopefully we will.

And if that happens, 30% growth should happen. We should be seeing 30% growth for a long, long time. So I don’t, I don’t think that we need to hurry into taking the company public anytime soon.

In fact, I would even argue that on behalf of the retail investors is we have to be careful. We shouldn’t be taking companies that are in the pivot zone and then give it in the hands of retail investors, because if it doesn’t play out the way we think it is, then it’s not good for them. So I’m trying to be a little bit more careful about timing this.

And secondly, internally also we have to be ready. Our internal financial controls have to be ready. Our ability to behave like a long lasting institution, and that’s what, which is one of the things that I’m working on right now.

How does Exotel operate, even if Shivaku doesn’t land up, even if Sachin doesn’t land up, even if Eshwar doesn’t land up. So this is another question that I’m working on for the last almost a year or so, institutionalizing the company. So these are the things that creates long lasting institutions.

So I think it’ll take about two to three years for us to finish all of this.

Siddhartha Ahluwalia 41:49

And what did you mean by, you said we are in a 20 year AI cycle. Like why, what is AI cycle and what does this 20 year period mean?

Shivakumar Ganesan 41:59

So, I mean, I used 20 years a little bit loosely, but if you think about any transformative technology, and when I say transformation, what I mean is technologies that will, I don’t know, increase revenue, decrease cost, improve CX, reduce risk, you know, technologies that impact in multiple different ways simultaneously at the same time. These are very rare. Internet is one such.

Smartphones is one such. Laptop is one such. Of course, AI is one such.

So what is the benefit of, for example, smartphones? Can it reduce cost? Of course, it reduces cost.

You can now communicate with whomever you want. So it has like wide reaching impacts. In fact, so much so that it is actually difficult to quantify exactly how much is the impact of smartphones.

And it’s just such a revolutionary technology. AI is one of those things. And then usually these transformative technologies play out over a long period of time.

Just to try to take your mind back to when the smartphones were started off. Do you remember that Nokia, it was like this red brick sort of a thing. And then I think my mama used to pay about 1 rupee 60 paisa for incoming calls at that point in time, right?

Now you take your iPhone and then you compare with that phone. It’s taken us 20 years. And now one feels like smartphones are saturated.

Everybody has it. There’s no new innovation possible, isn’t it? So I think this is a 20 year mega cycle that I’m sort of speaking about.

In the case of AI, right now people are getting obviously quite excited about what they are seeing. But the reality is that this is just the tip of the iceberg. Nobody could have guessed how you are going to watch movies, how you are going to watch cricket, how you are going to listen to music, looking at that Nokia brick phone.

So similarly, I would measure that you and I have no idea what AI is going to do to us, both positive and negative, sitting here. We are just seeing the first 2-3 use cases. Of course, search is everybody is using that already.

Coding is another large use case that’s come up. Voicebots, these are all some use cases that we currently know about. But this is just the tip of the iceberg.

And then we are obviously seeing the trailers of what is about to come. You must have seen Tesla Optimus. Now you obviously take a ChatGPT and hook it up into a Tesla Optimus.

It has hands and legs. So it can now start doing things for you. Then full self-driving is another thing that we are seeing.

So I think things will sort of continue to evolve. They will keep cranking away towards AGI. So all of this will play out over time and there will be newer use cases that you and I are not able to imagine today.

So that’s the 20 year AI mega cycle.

Siddhartha Ahluwalia 44:39

And you also during our conversations mentioned that today almost 80% of the Exotel revenue comes through enterprise and 20% through SMB. But when you had started Exotel for a long period of time, you were a SMB first company. So what had to change and how did you change the DNA of a company to become enterprise first?

Shivakumar Ganesan 44:59

Absolutely. So it was actually a long transformation. It didn’t happen overnight.

So we had to change the way we think about sales. There was a point in time when I said nobody will go and meet a customer.

Siddhartha Ahluwalia 45:12

Why is that?

Shivakumar Ganesan 45:13

Because we are an SMB company. So we cannot offer to now send salespeople to go and meet a customer. So we said everything is going to be remote.

Everything is going to be self-serve. Everything will be inbound and content. And then of course I had to contradict myself once we started becoming enterprise.

So then we had to now first of all hire salespeople. We have to change the culture. Then I myself had to go and tell them, hey, don’t come to office.

Go and meet customers. So a complete flip-flop. Enterprises buy very differently.

They take much longer. Very important, you get infosec correctly. The quality of the product is very different.

They are not really interested in self-serve, how quickly you go live, but they are very interested in how stable the product is, how secure the product is. Somebody’s job is at stake. So they’re not going to take that purchase very lightly.

These are large purchases. So our product development has to change. Our engineering mindset has to change.

Actually everything has to change. We had to now bring in a delivery team, a professional services team. There has to be paid support.

There has to be on-site support. There has to be resident engineers. So we did all of that.

And then it’s not an overnight journey. In fact, I would even argue that we are still in that process. I think we are there maybe 70-80%.

We are not there yet fully. But it’s taken us good four years, five years to actually go from SMB company. Actually, we did this transformation in three steps really.

First was the SMB company, long tail, what we now call as our scale-ups business. Then we actually ended up becoming startup focused or digital natives focused, which also is sort of like quasi-enterprise because we’re still talking about Flipkart and Swiggy and these are not small companies. And then there is the HDFC bank and the SBI, which is a completely different ballgame altogether.

I think this is the journey that we sort of went and it’s still happening if you ask me.

Siddhartha Ahluwalia 46:57

Do you advise founders to go from SMB to enterprise? Because SMBs are the easiest to crack. And maybe you can share it.

Enterprise takes like one year to go from conversation to a contract.

Shivakumar Ganesan 47:10

That’s right. So it is easier to start from SMB and then migrate gradually over to enterprise. So this is a very interesting conversation.

So I’ll say a few things. First of all, in almost all markets, more so in India, 80% of the value of that market lies in the top 500 companies. So you may try all you want, actually volumes will be high, churn will be high, but you’ll never be able to scale beyond a point.

And that’s exactly what happened to our scale-ups business. So frankly, I think as the company becomes larger, you don’t have a choice but to go towards enterprise. Even more so in India focus actually.

And I think the distribution of wealth is slightly better in the US, even more better in China, but in India, it’s even more skewed. So you have to be top heavy. So this is point number one.

Point number two, it’s not easy to just stand up one day and say, I’m going to start a company and go to SBI. They’re not going to entertain you. So it may make sense for you to actually build out the product for SMBs and then gradually go to mid-market and then gradually sort of go to enterprise.

This may actually be the way to do it. This is exactly what Exotel did. And so what I’m now implying, I don’t know if you already picked it up.

So because of our enterprise focus, we actually now have left a lot of open space in SMB. And so if you are now thinking about getting into our space, actually the way to attack Exotel is to start with SMB and then grow up. And then the cycle just keeps on continuing.

So as companies start and then they go up market and then new companies come and then again they go up market. And I think this is the cycle of life.

Siddhartha Ahluwalia 48:47

But do you think there’s enough space for founders or learning for founders that they can start day zero enterprise first?

Shivakumar Ganesan 48:54

There are a few examples. Cogno AI is itself a great example. And the way I think it happens is many of these enterprises also know that unlike 10-15 years ago that startups actually sometimes build cutting-edge tech.

So they do have a department, some processes that they encourage to go and work with startups and try. And that might be your angle which is what Cogno AI did so beautifully to sort of reach out. And so I think what you want to do is to be able to give great service, build cutting-edge technology and offer it really cheap.

And then this makes it very easy for enterprises especially through the startup programs to really take notice of what you’re building.

Siddhartha Ahluwalia 49:38

So Shivaku, what I keep on observing from your journey and learning from your journey, it’s not a typical startup journey that like started from zero and scaled right. Do you think like there are certain things in this journey which you can give to other founders that hey don’t do this, do it like that especially maybe in B2B software?

Shivakumar Ganesan 50:06

Sure. So see actually it’s a little difficult to generically give advice because I firmly believe that advice has to be strictly personal because the context and where the person is all very different. So I’m mindful of that.

So I think the audience also should be mindful of that. But I can say that market and timing is probably very important, much more important than how good of an entrepreneur you are. I think you might be the best entrepreneur in the country but if you are working in a small TAM or if you’re starting five years too early this is nothing you can do about it.

So I think the key skill then would be to predict the future preferably ahead of everybody else but not too ahead which is what I think you should call as timing and the ability to sense what is going to be very large and what is not. For example, web 3 sort of fell flat. We’ll see what happens to AI.

My guess is it won’t be, it won’t fall flat. It adds real value. So all the people who spent years on web 3, I think it’s sort of like time wasted right.

So of course there are some aspects of web 3 that will stay but generally you know what I mean right. So I think timing and ability to perceive the market ahead of time. I think these are the two underrated but very important skills.

A lot of people talk about perseverance, grit, intelligence, resourcefulness, hustle. All of these are important but I think far more important will be timing and sensing the market. I would also now say that there are two ways to look at my journey.

One is to say hey you know I’ve stood my ground for 14 years so perseverance, grit, you know resilience. These are the things usually mostly people associate with me and my story rather. But then one could also argue that are there ways in which I could have saved time.

Could I have now started much later and nothing would have made any difference. Like I could have started Exotel five years later or could I have sold the company when we were growing 40-50 percent year on year and could I have now started something else. So these are all the other ways to think about life as well and I do think about some of these things once in a while.

But the truth is I’m actually in a good place. I think 99 percent of the companies would have died by the time they are 14 years old. So but Exotel is still here.

We’re still profitable. We’re still making money. So my heart is filled with gratitude for all of that.

But along the way I have made several several mistakes, strategic errors that possibly could be avoided. But then again there’s no way for you to know what these errors are before they actually hit on you. There are some basic rules.

You should get your business design good. You should get your unit economics well. You should be clear about your GTM model.

You should think about risks ahead of time. You should think about your branding and positioning well. And you know you should think about building great products.

Your UX must be great. Use the latest stack for building tech. Don’t leave too much tech debt for too long.

So these are some basic principles and I can maybe give you like 15-20 of these things that I think everybody can be aware of. But it is the problem is when the rubber hits the road. If there are 15 things what do you focus on right?

So this becomes problematic. So then you have to hire the right people. You have to build the right culture.

What exactly is right culture? Sometimes it is better to focus on what’s happening today. Sometimes it is not.

Sometimes it is better to do what the customer is asking you to do. Sometimes it is not. So I think what overall I think in my journey as an entrepreneur and by the way some people get this very naturally.

I do feel that there are some communities where I think they breathe business. And it is I think a skill just like learning a cycle. And for some of us I’m not from a business community and so it doesn’t like my father has never ever spoken to me about money.

He’s spoken to me a lot about education. He’s spoken to me a lot about various other things in life. But you know our dinner time conversations have never been about money.

So unfortunately everything that I have to learn about money I have to learn it by myself. Maybe he didn’t know a lot about money. I don’t know why he didn’t speak.

So he’s no more. So I don’t know. So for some communities or some people it comes quite naturally.

Just like how you might be pretty good with coding or you might be pretty good with dancing etc. But for some of us we’ll have to learn it the hard way. Make mistakes and then iterate and things like that.

So I actually belong to that category. I just followed my dream. I wanted to start a company.

So I started a company. So that’s all I knew at that point in time. And everything else I kind of learned along the way.

So I’m not even like an MBA. So I’m just I’m just a coder technically. A product manager if you will.

So I think these are some broad thoughts around thinking about. I think broadly it is not. It is just a skill.

Actually making money is just a skill. People make too much out of it. It’s one of those things.

Once you know how to ride a cycle you don’t think about how to ride a cycle right. So in that sense makers of money or people who add value, they just have the skill of how to add value and how to spot trends. Everybody else has to learn it.

The faster you learn it the better.

Siddhartha Ahluwalia 55:40

You know what you have done I think for the Indian ecosystem is amazing. Like you have shown founders at a playbook that hey India for India software exists. And I think you might be among the top 10 or 15 companies which are just focused from India to India market.

Shivakumar Ganesan 55:56

Absolutely. I would say even lesser.

Siddhartha Ahluwalia 55:58

Lesser.

Shivakumar Ganesan 55:59

I’d say the top five.

Siddhartha Ahluwalia 56:00

Yeah yeah. And do you still think or do you want to break this narrative for anyone that hey there’s a market large market in India in other domains as well. Let’s say if you go to solve for banks in other domains.

Shivakumar Ganesan 56:15

So one has to be picky and choosy about the markets. So there are some markets that are uniquely large in India but is not very large elsewhere. If you land up in those kinds of market I think India is a great.

I mean obviously the market is only here so you don’t have a choice but to do it here. So I think that’ll be a good. So telco is one of those things right.

So like I said due to a number of reasons 10 years ago there were almost 8, 13 telcos in India. The per minute rate was the cheapest in the world and the number of people who had a phone phone was highest in the world. And then the number of phone calls that they were making per per day was the highest in the world right.

So what better place to start a telco business telco ish business than India.

Siddhartha Ahluwalia 57:08

Yeah.

Shivakumar Ganesan 57:08

So you’ll have to find such such markets. Now if you actually take this idea and go to a country where let’s say the call rates are very high and let’s say the calling propensity is very low and the signal is not that good and then you try to replicate this is going to go nowhere. So what are those kinds of market segments in India that is uniquely India specific.

So I’m thinking data annotation for AI is one of those things right. So where I feel like actually the labor is very cheap here. You know people have multiple languages.

They know a lot of languages. The BPO people already know foreign languages also. So it should be possible for us to recreate a good AI data annotation data labeling company out of India.

So this is another example that sort of comes in my mind and so I think one has to think through what markets exist that is uniquely Indian and solve for that.

Siddhartha Ahluwalia 58:01

And I think you described it earlier very beautifully that hey 80% of all value captured in India is among the top 500 companies in India.

Shivakumar Ganesan 58:10

Absolutely. Absolutely.

Siddhartha Ahluwalia 58:11

So the and today I assume Exotel would serve like 100 of them at least.

Shivakumar Ganesan 58:17

More. I think we have a total of about 200ish odd enterprise customers now.

Siddhartha Ahluwalia 58:23

Okay.

Shivakumar Ganesan 58:23

So about half of the top fortune 500 will be Exotel’s customers.

Siddhartha Ahluwalia 58:27

In India.

Shivakumar Ganesan 58:28

In India.

Siddhartha Ahluwalia 58:28

Right. And the more as a founder if you’re operating in software you thinking start thinking on on that framework that ultimately all the money is going to be made among those 500.

Shivakumar Ganesan 58:39

That is correct. That is correct. Absolutely.

Siddhartha Ahluwalia 58:42

And you have proven that these guys now paid earlier the narrative was enterprises make your services company and labor make your labor of the services company because India is not used to paying for software products.

Shivakumar Ganesan 58:54

I agree. Actually it is absolutely possible for us to build an enterprise software company focus in India to begin with at least. I think these enterprises are also now trying to compete.

Our own our enterprises are trying to compete in a global stage as well. So you know think of Tata Motors or think of you know Bajaj FinSev and these are like fabulous companies. So absolutely I think there is a you can totally build an enterprise business in India starting now.

Siddhartha Ahluwalia 59:24

And would you argue that the market has moved away from price sensitivity to quality sensitive?

Shivakumar Ganesan 59:31

No. No not at all. So I believe that that will happen when the India per capita GDP crosses $15,000 right now.

Siddhartha Ahluwalia 59:40

20 years away.

Shivakumar Ganesan 59:41

Yeah. I think we are one generation away. So India is a price sensitive market.

You have to do cost innovation. This is I’m connecting this back to the AI part. So I think cost innovation in AI is a game that we will play very well.

Another 20 years one generation later I think the India market will not be price sensitive.

And right now, they are price sensitive for a good reason because ultimately the end user is

price sensitive, right. So, the average Indian on the road wants a bargain, wants a good deal.

Siddhartha Ahluwalia 1:00:08

Yeah.

Shivakumar Ganesan 1:00:08

So, the companies that are serving them also has to, will want a good deal.

Siddhartha Ahluwalia 1:00:12

Yeah. Well, obviously then, the software companies that have to make big in India, they have to build it in such a way that they deliver that kind of a value with the price sensitivity in mind.

Shivakumar Ganesan 1:00:24

Correct. And then the good news is that people from outside can’t do this. So, actually the market is easy to grab if you really put your mind to it.

Siddhartha Ahluwalia 1:00:32

Yeah, yeah. Thank you so much, Shivaku. This has been such an amazing conversation, very different from the one we had last.

A lot of the previous conversation was a lot of centered around your journey as an entrepreneur. This conversation has been what happened in Exotel post COVID and some of the key lessons in scaling that you learnt over a period of time as an entrepreneur.

Shivakumar Ganesan 1:00:57

I thoroughly enjoyed this conversation as well. So, it just, I can’t believe that we have been speaking for over an hour now. Thanks for inviting me and it was a pleasure talking to you.

Siddhartha Ahluwalia 1:01:05

Thank you so much, Shivaku.