Divyaanshu Makkar 0:50

Hi everyone, welcome to The Neon Show. I am Divyaanshu, your host for today. I am the co-founder of WizCommerce, where we are building AI-powered sales and commerce tools for wholesale distributors in the US.

I also have two amazing founders with me today, Siddhartha and Omkar, who are also building in very complex and traditional industries which have a fairly big impact on our day-to-day lives. I’ll let them introduce themselves. We’re very excited to have these conversations with the both of you.

Kumar Siddhartha 1:20

Yeah, sure. So, I’m Siddhartha. I’m the co-founder of Merlin AI.

And we’re building an AI-powered ERP for US construction companies. So, replacing the likes of NetSuites, JD Edwards, Oracle, all of your old, traditional 30-year-old ERPs that have not changed, and trying to make it more beautiful, more efficient, and more automated.

Omkar Patil 1:40

Hi, I’m Omkar. I’m from Pienomial. So, what happens is that in different biopharma industries, the journey of medicine it takes to multiple clinical trials.

And those clinical trials, once they are passed, the drug is approved. But it is a common misconception in the industry that once that drug is approved, it is in the market, and the work of that biopharma industry is done. Their journey starts there, where they have to do more research, and they have to find the product market for multiple different patient demographics and different needs.

And that’s where we come in, where we try to cut the time that is required to make that decision, the research, by 80-90%. So, that is what we have been doing.

Divyaanshu Makkar 2:18

Very cool. So, I think all of us are building in kind of industries which get zero love from Silicon Valley. So, I mean, I think we can just start with what gave you guys conviction that this can be a fairly large business and an opportunity you want to go after because these are complex industries, hard to break in, long sales cycle.

I mean, we all know the questions VCs ask us whenever we go for fundraising. So, what gave you conviction to sort of go against all that?

Kumar Siddhartha 2:44

Yeah, I mean, I can talk about myself. So, my background, I’ve been an engineer for the longest amount of time. And I have, you know, as engineers, you are used to using a different world of software.

You know, you’re used to Slack, the best in class, basically. And in 2020, my co-founder Sneha who is also my sister. So, she’s worked with several construction companies, helping them on the other side, you know, just pick software, do implementations and all that.

And I heard the word implementation. I was like, what is this? Why do you need to implement a software?

Like, you know, what’s the whole thing? And then I got curious. And so, I was watching that space for a long time.

I think in 2021, we built out a kind of a passion project for one of the construction companies. And I was in Phoenix and I was at their site for two weeks. Just deploying and seeing, you know, what is this whole thing about?

And that’s when I realized that the bar was very low. And their problems were like real problems that can be solved very well with good software. But it’s just that no one was paying attention.

At least not the kind of software that was being, that we were used to. There was no parallel at all. If you had the parallels, they were like super expensive.

Plus, the expectations were different. So, it took me about a good few years to understand them, understand the persona, their problems. Why is the space like this?

So, it took a while for me to build conviction, I would say. And that’s when, but then, you know, for me, as soon as I realized that the problem was real and surely getting in is like difficult, but once you build a solution that’s robust, that solves for them, you’ll get a lot of love from them and it’ll stick with them for a long, long time. And so, I was like, I don’t want to build yet another $10 per user per month kind of product.

I was like, let’s go for this. So, it requires a different muscle. But yeah, you have to build it with time.

Divyaanshu Makkar 4:36

I think I had the same experience. I was like, the kind of software that we are used to versus what they were using, I’m like, this cannot, this industry cannot be running on this in the next 10 years. Like, it’s just not possible.

And, I think that’s, there has been innovation in these categories, but it feels like it’s not come from outside and it’s like people from within the industry who are not very good at technology, but sort of understand broadly how it works and kind of build something. And then they just build one thing and then they just go away. Like, it’s 40% EBITDA, bootstrapped SaaS companies, just two people in the company.

And there’s a lot of stickiness. Like, the software we move people from, they’ve been using it for 10 years. I mean, this company hasn’t pushed a single line of code in 10 years.

So I think there are a lot of reasons why not to go into these industries. But I think once you go in and if you’re really good with regards to the support and services you give, this can be an incredible business. And these people show a lot of love and they really stick with you.

They’re not flimsy buyers. Like, once they buy in, they’re with you for the next decade. And that’s something that I really like about this space.

What about you?

Omkar Patil 5:41

Yeah, so for me, I love insights. And I think Biopharma provides very fascinating insights. So I’ll give you one example.

So think about it this way. So we see on Instagram there are celebrities who do dedicated hard work and diet and you see them losing a huge amount of weight in six months. And then you say, wow, such an amazing diet routine that they have followed.

Let’s go and follow that. So that’s more of a word that people love. People are excited to see that.

But if you think about, you know, think about what is happening in medicine industry. So, and this part gets very boring. So there are drugs called GLP-1 agonist.

And these drugs are used to treat diabetes. Now, this is another part of it. But you will see that GLP agonist and its prescription is 5% of entire US’s total number of prescriptions.

So does that mean 5% of US’s population is having diabetes? That’s not possible, right? So is there some connectivity between them?

And these kind of, you know, insights that you derive from business. I think bio-pharma is filled with these kind of insights. And, you know, I love those insights from the beginning.

So how can you build systems which can provide these insights at scale? That is something that we wanted to do for a really long time. And, you know, this became an opportunity and we went into it.

Divyaanshu Makkar 6:55

Got it. I think, let’s go back to the thing you said. Like, you know, it requires a different kind of muscle.

So let’s talk about what that difference is. Like, building a software for these kinds of categories versus, you know, right now, I think the expectation is that you launch something and 10 million ARR in one year and 40 million next year and it just will be viral. But these industries are very different from that perspective.

So what are those differences according to you and like the things that you’ve seen when you guys started?

Kumar Siddhartha 7:25

So I think the, a very important piece to understand is that you’re building business-critical software that is supposed to sit on the direct line of work. So if your procurement piece doesn’t work, the guy is not able to send out a purchase order and the commodities price go up tomorrow, they make an actual loss. So it has an impact on the job of the individual person who’s sitting inside.

And so it’s, you can’t afford to make the kind of errors that you can with a different kind of software. That’s very important to, it’s very important to understand that first. And how it impacts everyone else within the company, right?

So it impacts profitability, other people, you’ve got one large software that is kind of interconnected. All of those pieces are to be addressed. That’s one part.

And then the whole experience of how is that software implemented? A lot of these people are not hardcore software users. For some reason, I always used to feel is, you know, when you install an app on an iPhone, no one has to come and train you.

And that should be the same experience. But, you know, for whatever reasons, the way this industry has evolved is they expect to be trained. They expect it to be clunky.

They expect it to be bad. So by the time you show them a software that does something in three clicks, they are sometimes like, is this it? And they’re actually relieved about it.

But these are the kind of surprises that you would get once you start learning about them, at least in my space. And their main line of work is something else. Software is supposed to augment it or sit in the direct line and actually reduce their work.

If you’re not doing that, then you’re not helping them. And at the end of the day, it’s very, it’s quite brutal. So they look at it from does it really help me or not?

If you’re not a must-have, you’ll mostly be out of that business. So that’s something that we learned.

Divyaanshu Makkar 9:08

I’ve had people who come and you know, how do I do this? I’m like, there’s a button right here. I saw that button, but I wasn’t sure.

If I click it, would something break? I’m like, this is a very different kind of persona from that perspective. And I think that’s the point.

I think they have been burnt very badly with softwares. And that’s like a headwind that you constantly have to kind of fight against. Like everybody went through maybe a three-year implementation cycle which never really came to fruition, cost millions of dollars.

Kumar Siddhartha 9:39

Exactly.

Divyaanshu Makkar 9:40

I think we should talk about next, you know, like how do you build trust in a category like this? Especially your category, I think is…

Omkar Patil 9:47

Yes. It is… And the people that we are working with, they are extremely specialized.

So I’m talking about MD-PhD people. Okay, people who build drugs and those drugs cure humanity or treat humanity’s ailments for next 15-20 years. Now, it’s very hard to say, you know, that these people don’t know the technology.

They know the technology. They know those dots. They know that they want to reach over there.

But, you know, the piece that they don’t want to focus on is how to connect the dots. So, and here the little magic comes in is if you’re very good with technology and you understand what your client is talking about, you can empathize with what their requirements are, then these are lower hanging fruits. That is what I have seen.

So what you have to do is you have to just make, you know, make sure that you are connecting those dots. If you are able to connect those dots, they are early wins. And they are very easy wins.

So that’s what we have seen in this particular one.

Divyaanshu Makkar 10:40

Got it.

Kumar Siddhartha 10:41

And for us, additionally, I would say that you have to have exceptional support that really helps to build the trust. And it doesn’t take a long time.

Once they know that this is the pattern to expect, like, traditional ERPs would take like days and weeks to resolve an issue. Sometimes even months. And they would not prioritize things.

So once they see that you are prioritizing their problems, and you are actually solving for them, and giving them the kind of experience that, you know, we are used to seeing in this side of the world, they start, you know, that starts making up for everything else.

Divyaanshu Makkar 11:16

Yeah, I think it’s sort of every part of your process, right from sales to like customer support, you have to pick the things that they hated about the previous vendor. One of which is every query takes like three days to resolve, right? So if they tell you something on the first demo, if in the next one you can just show like, hey, we kind of like already wanted to show you how it works, we already have this live.

Those kinds of small, small things and taking them through very structured sales process, even in customer support, actually responding with like, and fixing those issues. I think that kind of helps build the trust. And it’s not a one-day exercise.

And I think this trust and the fact that they don’t trust you also has a lot of implications on your GTM.

Kumar Siddhartha 11:59

Yes, 100%.

Divyaanshu Makkar 12:00

Traditionally, like, you know, the playbook from India at least five years, six years back was, you know, a PLG with like some inside sales motion.

Have you seen that work in vertical categories in your categories? What’s been the learning on the GTM side in a category where trust matters so much?

Kumar Siddhartha 12:18

PLG? Not working. Doesn’t work.

Like, hasn’t worked in our industry for sure. If you had a horizontal piece, yes. And now with AI they’re looking more.

So if you start exploring in some categories for sure, but not the vertical space. It’s too risky and it’s too much for them to lose if they go and implement. There is a reason why they’re paying millions of dollars to get implemented.

And that they want to avoid all that risk. They want a person who can, you know, even the sales process is long, much longer because they want to feel that you’re understood, your problems are understood, that you are a reliable partner. They have to know you.

It has recommendations, pay a big, big, your reputation in the market is a big thing. All of those things add up. And that’s that the whole cost is actually not just the software.

It’s everything combined. And that’s what ends up building the brand as well. So how the market perceives you, your support, your product, what have you done in the last few years, everything adds up.

So in a sense, it’s also like not just building a vertical product, but like building a vertical org. You have like so many things to tackle. And that’s why I said it requires a different muscle.

And you only learn it with time. Like, unfortunately, I don’t know. I mean, I could not find a lot of stuff.

Maybe this podcast isn’t probably the first time. Too many people are talking about it. But yeah.

Divyaanshu Makkar 13:36

So what was your early GTM? Because you’re trying to replace an ERP, which is, it’s a big piece of software. And who, and I think you replace mostly, so they have already spent a million dollars probably in their last implementation.

So how do you go about like even getting your first few customers maybe?

Kumar Siddhartha 13:51

So the first few customers for us was we had to trial, do a lot of trial and experimentation with the market segments. And construction is like very diverse in the US. You’ve got residential, commercial, you’ve got like different people in the stack.

So there’s home builders, there’s developers. And you have to have an understanding of that market as well. There’s several layers to it.

So we’ve worked with architects, like I’m talking of the early days. We’ve worked with architects, we’ve worked with homeowners, representatives, we’ve worked with small builders. And of course, like my co-founder Sneha, she had, she’d always worked with these large home builders.

And after doing all of that experimentation, we realized that the, both the problem intensity was like super high within that home builder segment because they needed an end-to-end software that did everything in a good, reliable way. And that space was completely vacant. So you wouldn’t, you would find like generic stuff that they’re trying to reimplement into this industry.

But there’s probably few to none that exist from the ground up as an ERP replacement for construction companies. There’s very few. And so that was a good white space for us that it took us a while to discover, but then that’s when we realized.

And our customers were telling us and the market was telling us constantly that the, the maximum value for my users was when they had everything in one place. So the vertical was making sense to them. It was not making sense to us.

So it took us a little while for us to get the signals from our experimentation and understand that.

Divyaanshu Makkar 15:20

Yeah, the traditional thought process in software at least, that comes from the US or the Valley is like, make one thing, make it the best and just go for it. And that just doesn’t work. Like people just don’t want it.

People don’t want another point solution in their lives.

Kumar Siddhartha 15:34

So this is something that we can talk about in some time as well. But there’s several things that are like, as general advice for normal startups that has not worked for us in vertical. But we can do that at some point.

Divyaanshu Makkar 15:44

What about you? I’m assuming just to get mind shift from these people like PhDs, you need to be an expert yourself. So that they even listen to you.

Omkar Patil 15:54

So we have experts. So my CEO, Dr. Sanat Mohanty, he’s a former professor at IIT Delhi. And we started our journey at IIT Delhi.

My first company was incubated. My master’s project was my, you know, so I was working in GPUs in 2011-12. And so, you know, we know this place very well.

And that’s why it has been more of a straightforward journey. But we have built that muscle which requires to do the research part of it. We can go deep into that context.

We can, you know, we can understand what that means. But, like, leaving the innovation part aside, and again, you know, I am feeling that cost of innovation is going to go down significantly in the next few days. And I generally say this, but I don’t know whether, you know, this is true or not.

But take it with a pinch of salt is that the person that I am hiring from a local engineering college versus a person that, you know, probably Silicon Valley startup is going to hire from Stanford and all of that. The intelligence of both of them is going to be same. You know, because of the advent of GenAI.

Because the amount of research that is required to do, make a decision, that’s already the same. This person is also going to use ChatGPT, this person is going to use ChatGPT. That guy may be using $200 of solution, but, you know, that’s okay.

Now, the innovation part aside, which is going to get cheaper, the part where, you know, here in at least this particular industry I saw becoming much more important is the survival. So, the people that we talk to, what they have seen and we have seen, you know, heard them saying it, you know, that if you don’t see you survive for next two years, we are not going to engage with you. So new companies by itself, they don’t engage.

They want you to survive for two years and then they start talking. So they know that you are self-sufficient. So, this is a chicken and egg problem, right?

If you don’t have money, how are you going to survive for two years? But, again, you have to build muscle. You have to make sure that you are surviving in that time and then innovating and that’s a completely different Kichidi, but that happens.

Divyaanshu Makkar 17:56

How long did it take you to get your first customer? And, who was this customer? Like, was it somebody you knew or was it through a network?

Omkar Patil 18:03

So, again, it was for, you know, so what we have, so there are, you know, multiple layers of, you know, systems that we have built. The first layer of system was basically taking how can I automate Python waste workloads? That was it.

And we were working with, again, extremely regulated medical device manufacturing industry. So the problem that they had is it’s an implantable device that goes into your body and measures the body glucose content online. Like you can just download an app and see what is your body glucose content for, you know, six months or so.

So that was our first customer. And again, we got it through, you know, contacts and warm contacts is how you’re going to get your first set of customers. But once we got that, then moving to other sectors it was a slightly easier problem because that particular asset was already developed.

Okay, we understood that, okay, what kind of problems they are seeing. And when we moved from, and again, you would say that, okay, biomedical and biopharma, oh, both of them have bio, they will have their share similar. No, they’re completely different industries.

And the, you know, the discussions that happen, they are completely different. The kind of problems that they have, they are completely different. Medical device is extremely regulated industry, where biopharma, it’s also extremely regulated but slightly less regulated.

So with that, for biopharma, what was required, what kind of requirements were, you know, like the initial discussions that we started, the client in that particular pharma, it was a consulting company, they gave us certain, you know, interesting problems to solve and we were able to solve it because again, the technical part comes heavily down, the innovation part we had already got it. But as soon as we went into their problem, we were able to solve it. And then they become our first client in biopharma and then we expanded in biopharma.

We got more of those requests and we were able to show what we have done. And now, we are going to the clients and clients are saying that, you know, what do you have? Show us what you can do.

It’s not that, okay, I have this particular requirement. They are amazed by the technologies, things that they are bringing, you know, tell us what you do and we are able to amaze them. That’s the part that is important.

Kumar Siddhartha 20:02

So, getting in is somewhat difficult as Omkar, you just mentioned, you need to know some people, at least have enough understanding of the problem space in the market but staying put is easy if you do a good job compared to like, you know, other software where switching cost is non-existent. But here, the switching costs are really high. So, those are the upsides that you need to balance for sure.

Divyaanshu Makkar 20:25

But even for you, your first customers came from your co-founders network or was there some early GTM?

Kumar Siddhartha 20:32

They did. So, for a month, I have done cold calling sitting out of India. Doesn’t work.

Doesn’t work in this space because we had to make sure that all these hoops are covered. So, we’ve tried almost everything under the sun. But for us, the early customers came from the market and because the CEO, they have the network.

Omkar Patil 20:51

Warm welcomes do an amazing job. So, if the person C is being introduced by person A, you get to know that person and there is a cold shoulder that is given to you. And if it is introduced by person B, you will be regarded completely differently because person B has told you to meet that person.

So, the introduction happens twice but second time, it is completely different. You will get a much warmer welcome. So, yes, introductions work and warmer introductions work.

Kumar Siddhartha 21:23

And word of mouth is like, once you get some decent amount of traction word of mouth is incredible. It’s actually like I was, I am pleasantly surprised by how much powerful word of mouth is here compared to horizontal SAAS. People really take their competitors seriously.

When they know the other companies are saving money and doing better, I need to do this because it makes me less competitive. So, that’s their thinking is if you have something that makes everyone else better, I have to at least make sure that I meet that bar. That’s one significant piece that keeps driving people to you.

Divyaanshu Makkar 22:02

I have been surprised by how willing, even though all of these people know each other, they compete with each other. It’s very weird that they’re still willing to actually call the competitor for you and tell them like, hey this is actually a great software. This helped me in these many ways, even though they all compete with each other.

And they are very very open at least in the US I found like they are very very open to help you if you have really done a good job. We started our journey from India and that was not our experience in India. Like in India we were constantly would…They would not want to… They would like try to keep it like a hidden secret. Okay, we have this, we will not tell anyone. I don’t want you to work with these three people because we compete with them. But the US is a lot more open and there… People have actually just come out and said like you guys have done a great job for me. Let me actually, you know, introduce you to this person without us asking also. So these industries are like very very tight knit and it’s both good and bad because if you screw up and I know somebody who screwed up an implementation in some other industry and they got the biggest customer in that industry they screwed up the implementation whatever happened and then they actually had to shut the product because nobody else would give them a chance. So that’s also like if you do get in the door like, make sure like you just go above and beyond, it doesn’t matter what you need to do to keep them happy, but you need to keep them happy.

Omkar Patil 23:18

And another part, so what we have seen is US companies, they are not monoliths. Like, you you can’t treat a company as just one single entity. Even if you take research division for every different cancer, you’re going to have different research teams. And, you know, that person who is engaging with you, that person will happily introduce you if it works, they will happily introduce you to other companies and their, you know, counterparts because it works like, you know, entire industry is growing. What’s wrong with that? But yeah, word of mouth, either in good way or bad way, it does an amazing job of doing what it does.

Divyaanshu Makkar 23:55

Yeah.

So coming back to I think, just the complexity of the problem you have to go after. Sometimes the problem, I mean, from a software perspective, it’s all solvable.

You see, like, a lot of opportunity to, like, optimize things. But in these industries, again, you’re either… Like, people’s expectations now are, like, you know, solve everything for me.

I don’t want another point solution which doesn’t talk to other software that I have to learn. So you just, when you start, you just see this massive build, like, you know, how… I mean, just go-to-market just seems insanely long.

And then every, at least in my category, both people could be selling the same products. Could be both wholesale distributors. But they’ll come with, like, different ways in which they’ve organized their firm or very different workflows that they need powered.

So how did you guys go about, like, kind of… That, like, there’s no MVP in this industry. And there’s no…

The concept of early adopters, I don’t think exists. Like, I don’t think you can go, like, hey, be my design partner. They’ll be, like, fuck off.

Like, please, I don’t want to spend my time designing a software. When it’s ready, please come to me when you have some traction. Compared to horizontal SaaS, if you’re selling to a software…

I am happy to, like, be a design partner for somebody who’s solving a problem that I have. But these industries don’t work like that. They want a functioning product.

And they wanted to cover a lot more ground on day one. So how do you, sort of, go about that build-out on day one?

Kumar Siddhartha 25:21

So the design partners thing doesn’t work. I don’t think design partners work in general as well. Because they end up giving you very tailored information to their specific problem.

Which, after a point of time, is not applicable to everyone else. So you have to be… You have to have a good enough sample space to know that…

You know, where do you, kind of, draw the line over there. And that’s what we have seen. So it makes sense to basically find people who are willing to pay for an implementation.

Because then you know that there is a real problem that you’re going after. They’re telling you exactly what their problem is. And that’s where, you know, having understanding of the industry becomes very critical.

You need to know where to draw the line. Because, I’ll give you an example. Someone said, can you give us…

Can you manage our websites? And can you manage our inboxes? And we’re like, no, we’re not that.

But your users will not know where to draw the line. So you have to be very, very clinical about… This is what we do.

This is where we say these are the guardrails. Beyond this, you have to look for something else. And to them, that’s totally fine.

So you need to have some understanding of the space that you’re operating in. And the problems. Pick those problems.

Make sure that you have an expertise in those problems. It’s like picking… Instead of picking one problem to solve very well, you’re saying, I’ll solve five problems very well.

And that’s my scope. But I will limit it to those. And it doesn’t go beyond that.

So that’s…

Divyaanshu Makkar 26:41

But it’s still a fairly large build out, I’m assuming. Like, it’s not a…Because if you’re replacing NetSuite somewhere…

Kumar Siddhartha 26:48

It is. It’s like building 20 products in one go.

And you need… It’s not just building 20 products in one go.

It has… You have to go vertical enough so that you make sure that there is feature parity. And then you have to delight the customers. You have to go above and beyond as well. So you have to cover…

Divyaanshu Makkar 27:02

So building whatever they built in 10-15 years of existence. And then you’re building better stuff. So that they feel that, you know, switching makes sense.

Kumar Siddhartha 27:09

Makes sense, yes.

Divyaanshu Makkar 27:10

So tell us the… Actually, like for the audience and maybe founders listening who are trying to build in vertical.

What was the process you followed then? Did you pick one customer and just spent a year with them building the full product? Like, what was the early days like?

Kumar Siddhartha 27:23

Yeah. I wouldn’t recommend that.

Divyaanshu Makkar 27:28

Maybe what you did, what you did wrong and then what you should have done.

Kumar Siddhartha 27:32

Ours is a slightly different story because we’d seen so many implementations. Sneha had seen like several implementations. She’d done it herself first-hand.

And large implementations with big companies. So she had enough understanding over, you know, the past seven years of what’s wrong and what needs to be built. And so she was the one who was evangelizing us and kind of telling us what to do.

Where do we decide where to kind of put the guardrails? And that’s what we literally did. And we said that, you know, for each of these companies, we basically built out with them.

So I was actually the first customer that we went live with. I was in the middle of nowhere in Texas for about two months. And I’m a vegetarian.

So the only thing I could eat was margarita pizza. I mean…

Divyaanshu Makkar 28:16

I love Texas because of their barbecue.

That’s because I’m a non-vegetarian.

Kumar Siddhartha 28:22

But the thing is that, so because I was there, I could see it firsthand. I could build, I could fix, I could troubleshoot, I could do everything live. And that was incredible because it built my product in one go.

And I was doing nothing but coding for like eight weeks.

Divyaanshu Makkar 28:38

And to close this customer, did you have like a functioning product or was it very bare bones?

Kumar Siddhartha28:43

We had a…

Divyaanshu Makkar 28:44

prototype. What was it?

Kumar Siddhartha 28:45

We had a functioning product, but until the rubber hits the road, you never hit the use cases. You could very well sit and, you know, in a silo and keep building and keep building. And it could become an endless exercise.

But until some customer uses it, you will never know where it breaks. You will never know whether it works or not. What needs to change?

Where does it need to change? That feedback is very critical for the product to get out. So never build without a customer.

So we had the prototypes. We had some verticals that specific individuals were using. But by the time that we decided that we’ll go build for someone, we were clear that we wanted to have a partner.

We will sit with them. We’ll do whatever it requires to build with them. But we are not building and waiting for someone to use it.

Divyaanshu Makkar 29:28

What was this partner’s motivation to go through that process with a company that is fairly new?

Kumar Siddhartha29:33

So of course, I mean, they knew us.

Divyaanshu Makkar 29:36

So there was a trust.

Kumar Siddhartha 29:37

There was a trust factor that was there.

And the other thing was that they had done some implementations that had not worked for them. And so they were looking for someone or at least a partner that could customize some of their workflows. There was something in it for them as well.

And so we sat down with them and just made sure that we got all that done. So there were definitely pieces over there where we went above and beyond. I did their website, for instance, which was like super old.

I still have like their domains are on my system. So in some sense, I was their tech team for a good amount of time. But then that first customer was very critical.

Otherwise, we would have not been able to get to the second and the third and the fourth. So as I said, you get the first one right, the second one comes in and the third and the fourth. So from there, it becomes simpler.

But the first is difficult.

Divyaanshu Makkar 30:27

What about you? What was your journey from a product standpoint? And how much did you have to build to convince somebody? Or was it a more partnership?

Omkar Patil 30:37

So one thing that you have is your, you know, what is your earlier ground from which you start? So thankfully, we were at a ground that we can say that, OK, we need a paid project. And I have seen that what happens in a lot of industry, even for bigger industries, they’re going to say that, okay, you don’t do this for us. And then we will start to use your. Then we’ll start something and it, you know, it, it never goes anywhere.

And you know, a lot of bigger companies have also followed in that track. So we’re at that particular point where, you know, we started getting the paper. Number one is be ready to have failure because there are going to be, you know, some failures in the build that you do, but while you’re doing that, understand that, and this is said in completely different context, but I think it applies here as well, that last generations media is this generation’s content.

And that is what we have seen again and again and again happening. So in a way, what we had built, for an example, there was a, for our first client, we built a dashboard to understand like how their entire pipeline is moving. But we found out that whatever, you know, that particular pipeline, it took us six months to build. We found out that we can automate it.

How much we can automate it? Probably up to three days. Then we build that particular framework, which automates that. Now, this kind of automation mindset has helped us that even though I am failing, I am incorporating the failures into, you know, the, the framework that I have built and because of the genAI. GenAI has been a huge boon for us because it empowers the earlier frameworks that we had built.

So if now, if that same dashboard was taking probably six months, now it takes probably five minutes to build because we are incorporating those pipelines. And when you go with this particular thought process, we mindset itself, that automation is the way the world has worked so far and it is going to work so far. It, it, it simplifies a lot of your work and it simplifies a lot of your processes.

Divyaanshu Makkar 32:21

Got it. So were you co-building this with the customer? Is this the problem statement came from the customer?

You already had built something.

Omkar Patil 32:31

Yeah. So we, we had, so first piece that we built, that was, we, you know, that was more of, you know, can you build this for us? And we built that.

Divyaanshu Makkar 32:37

So, it was more of…

Omkar Patil 32:38

It was, yeah, and…

Divyaanshu Makkar 32:39

customer backward

Omkar Patil 32:40

It took us six months to understand like initial. So we, we sat, we wrote that Golang code and we build all of those things. Oh, really?

You know this, I could automate. So thankfully the notion part is what I’m going to emphasize. I don’t see that in a lot of Indian startups.

We have some space to innovate. So you won’t believe in last six months, we have filed around seven patents. Okay.

And you know, one of them has already gotten approved. So the one is in process. So this kind of innovation that doesn’t happen if you don’t give it enough space to work and, you know, patents, they work as, you know, for smaller companies or Indian, you know, ground realities, it may not work, but for US companies and specifically in biopharma, it works very well. It does. It creates a mode for you to make your make a play. So I think as time is passing, innovation is going to be the key.

So for the particular client, we co-build it with them. We, you know, they, they helped us. They gave us requirements.

And you know, we can automate it. Now we have automated that as well. We have automated whatever came out of that as well.

So automation has been the reality. And I think that is going to drive the future as well.

Divyaanshu Makkar 33:45

Got it.

Got it.

But basically like don’t build a lot before you have a partner.

Omkar Patil 33:51

And get paid projects. Get paid projects. Don’t do anything for free.

Divyaanshu Makkar 33:56

Yeah. Yeah. If they’re not paying for it, like, you don’t know, they’ll lose interest or, I mean, it’s just not a great.

Omkar Patil 34:02

And I have this feeling that, you know, no one appreciates things that they get for free.

Divyaanshu Makkar 34:05

Yeah.

Omkar Patil 34:06

Okay. Then what that means is you need like get paid 10 rupees, but you know…

Divyaanshu Makkar 34:08

And these people are used to paying. They’ve never seen a free implementation. Actually free means that you will do a shitty job to them. If you’re charging money.

And I mean, you could charge less because they’re their first customer, but if you’re charging money, they at least know that you, you’ve thought about what my costs are here and you’re not trying to like and they’re used to it. Like they’re used to paying. So why not?

So you build your first product, like, and you went like this approach of full ERP. I will replace this monolith, but there could have been other approaches to this problem as well. You could have maybe built on top of an ERP or things like that.

So what was the mind, why make that decision of like replacing the super monolith system?

Kumar Siddhartha 34:58

So this comes from my engineering, experience is every time you build on a platform, you incurred the debt that the platform has. And then you have to work within the guardrails that the platform dictates. So for instance, let’s say I was, you know, as we were talking a while back, my previous startup, we were building something on top of Gmail.

And, if Gmail decides just like Twitter did, right. Twitter shut down their API. I remember like a few years back.

And so all businesses that were working on top of it are like gone overnight. So that’s one, you never, it’s a very kind of risky proposition to build something on top of someone else. Plus in our space, like with ERPs, the same software works for oil and gas.

It’s supposed to work for construction. It’s supposed to work for like 10, 20 different industries. So almost every time a single user is not utilizing, the platform completely, and there’s a lot of deadwood that it has from like all of the other industries.

So we had the luxury of saying that, you know, we will not do all of that. We’ll just build out from the ground up. We’ll build it cleanly that phase.

We understand everything else. And what it has done for us is that now when we are trying to train our models on top of our database systems, our application layers, it’s much more simpler compared to an opaque platform where you don’t know what the code really does, but you have a bunch of interfaces to code your APIs with. So this is like much more simpler compared to what it was.

And I would say that, you know, our initial, hunch was that we’ll do it because I specifically didn’t want to incur a lot of that tech debt, but it has worked out for us in the long run. But that said, there’s like several companies that are trying to do like smaller pieces on top. So if you’re trying to automate a specific workflow, build a web browser plugin that does something, yeah, sure.

Go ahead and do it. but you know, just kind of building out your own thing as significant advantages and the other thing is, I mean, even if you take something like QuickBooks, right, which has an open API, go take a look at it. oh my God, like the amount of stuff that is in there.

And, you know, most developers like that you would go out and hire right now would not even know how an XML format works.

Divyaanshu Makkar 36:59

I, we connect to QuickBooks desktop because that’s the most used it’s like for SMB customer, that’s the ERP they use. I mean, it took us a while to figure out, like, how, how do you like even start, like, who do I go to who wants to like code in this language and understand how to connect with this system. And that that’s definitely like NetSuite, which you said is like, it’s a modern ERP, it’s modern and it has REST API.

So that’s like, if a NetSuite customer comes in with really, really happy, like, okay, this is going to be much easier for us to, go with. But I think, I think that trade-offs to that approach is that you will, it’ll be, you’re replacing your sales cycle are going to be higher, you will, like the takeoff will, will be longer. Like you’ll have to spend a lot more time.

Kumar Siddhartha 37:47

You are like really building on products in one. But then it helps because you’re, you know, your quotes, your support is much faster, debugging is faster.

You can build out features, ship them out much faster. compared to if you were doing it.

Divyaanshu Makkar 38:03

Was it a big debate internally or was it like, okay, this is how we are. That’s this is…

Kumar Siddhartha 38:07

No. So I think so. we are two technical co-founders on the founding team. So we’re like, we’ll just do it ourselves.

Divyaanshu Makkar 38:14

Yeah. Got it. We took a different approach.

We said we will not like go replace ERPs, but we will improve the customer experience with all the other products that sort of, but leverage the ERP data. So our wedge was a sales application, which was the most legacy, which we thought is the easiest to replace. Even within that, we realized that like there were four or five bootstrap competitors, but we picked one.

We said, okay, at least with this competitor, I’ll build feature parity. And then have these like five things that I will delight with. And then for the first four or five months, the only people we got like who switched to people who are using that competitor, but then I think we innovated on like, okay, we’ll connect with all of these shitty back office systems like Sage 100 and that like took so much bandwidth from our team to do.

Kumar Siddhartha 39:00

There’s another thing in this is that there would be some ERPs that are doing something and some are completely not doing something. So what do you do in that case? You have to go ahead and build it, but then you’re building, you’re slowly starting to already see that evolution of you building out your own stack.

so it’s, it’s inevitable. You would have to, at some point, start building out. That’s at least in my industry.

That’s true. plus the other thing that was clear for us was, because for our space, it was such a, there was a void. There was no ERP that did it grounds up.

so it made sense for us to kind of just build it and, you know, not have to worry about it too much.

Divyaanshu Makkar 39:34

Got it. What about, what about you? How do you think people should approach these vertical markets?

Like what works to get in early because you can build so many things. You can, I mean, most of these industries are old, so they will be on several legacy systems.

Omkar Patil 39:49

So on that trip, you know, I completely agree that you, you have to start something on your own. You can’t build on old, old systems. Because what happens is that you are going to get stuck, with the specific legacy problems that these legacy problem, you know, legacy systems derive.

So what we did is in, like earlier, we, rather than relying on web frameworks itself, we, we have built our own from grounds, technology stack from, you know, in Golang. So if you think about, technology that goes into ML, everyone has been thinking that, okay, PyTorch or Python is the way to go, we only use that for training in production, the code is in Rust. So whatever the efficient related efficiency related problems are, we focus on core languages and core problem core solutions to that.

So when you think, and again, we have some extra muscle to do that, make that happen. So when you do it that way, you, you do not incur that technical debt, but at the same time, you retain the agility. If you need to move it from A to B, or if you need to incorporate another feature.

And again, we have huge obsession towards automation. So if there is a feature that needs to be implemented or it needs to go in the pipeline, it stays in the pipeline. We never say that I’m not going to build this feature.

We will build that feature. Only thing is that does it go into the automated layer or does it go into automation layer? If it is an automation layer, it will take six months to, for us to build.

Otherwise it will get built probably automatically. So these things we have seen happening and because we own our technical stack, it has been very easy. But a question to you, like when you said that regarding QuickBooks and the XML part that you said, are you seeing Gen AI being used because all the documentation or whatever the implementation that there is, it must have been gotten, what do you say, percolated into those larger language models.

So are you seeing that kind of speed coming in for the future, you know, futuristic systems that will come in?

Divyaanshu Makkar 41:43

On the back office systems, I don’t think so. I don’t think any of these companies are really trying to even innovate. And with the thing with the ERP market is that a lot of these are still on-prem systems.

They’re not even, a lot of them are not on cloud. People haven’t even moved to cloud unless Microsoft comes and says like, Hey, we are just stopping the system. They will not move to cloud.

And QuickBooks desktop, for example, QuickBooks wanted to like basically shut it down and they got so much flack from the market that they did not. And now they’re actually actively even selling it. The resellers are still selling QuickBooks desktop.

So I don’t think you’re going to see a lot of innovation from them. I think for us, we thought like, you know, it’d be much faster if we don’t, I mean, of course, like an ERP touches very, a lot of different parts, but in our industry, at least we saw that there was a lot of need for innovation on the customer experience side. Like people wanted to buy differently, that e-commerce we’ve hadn’t touched wholesale.

It kind of completely skipped it. But COVID like was an accelerator. People were getting younger.

So we thought there are more than enough for us to do outside the ERP. And then over a four or five period of time, like we can actually replace that legacy system. So we started with the wedge and we then said, okay, to make this really work and sticky, we should integrate with these systems so that nobody has to come and upload data into our system.

Information is real time. And what that did for us was like, okay, once we did that integration, that dirty work, the first software worked, we kind of became a natural extension for everything else. Now, if they want to automate customer support, for example, QuickBooks desktop is not going to do it for them.

I don’t think there will be a startup who comes and says, okay, I’ll only solve this for QuickBooks desktop, but I’m already integrated. I have that data. I have that interface.

So I’ve become kind of the natural partner to innovate. So that’s what we are doing. We are saying like, okay, we’ll innovate around this legacy stack for now.

And I think if you stay long enough, everybody becomes that sort of one goal is to become that one platform, but sort of starting with the wedge, owning the workflow, and then sort of embedding us in other parts of the company. And that’s the approach we’ve taken. I think every market has its own nuances.

In our case, like ERPs really don’t do a great job on the sales side. Like they just don’t touch that. People have now started launching websites also, but again, they’re not really, they’re never been focused on customer like UX or UI.

And when you’re making customer facing technology, you want to be right. Like a wholesaler wants to have a website, which looks good. Right.

So I think that’s where ERPs were just not there and were not really playing a big role in that. So we, we, we found that, you know, a good wedge to sort of at least enter.

Cool.

I think one of the things that is a very popular statement these days to give is SaaS is dead. We should all pack up and go home and think of some AI native startup completely. What, what are your opinions on it?

Because is it, is SaaS workflow completely dead? Are agents going to run a wholesale business? They’ll deal with the factory.

They’ll order stuff. They’ll distribute stuff or what?

Omkar Patil 44:54

Someone has to design those agents. Right. Someone has to maintain those agents.

Someone has to build that ecosystem. I don’t think like, you know, even if you go into AI age as well, I feel that those AI native. So we are, you know, basically AI product completely.

And each of the product has the AI facing, you know, AI principle component that goes into it. But in the end, we are maintaining it for our clients. We are selling it as a software as a subscription.

So, you know, that works so that, you know, that, that particular mindset that came in because they thought like, you know, and again, I’m talking about the software giants, which are AI giants, Sam Altman and all of that, the AGI is coming in. So there is going to be this intelligent system, which is going to sit alongside you, sip coffee and, you know, solve all of your problems. Okay.

You know, let’s take it with a pinch of salt and pinch of salt is very much needed in, you know, everywhere when you talk about AI, but that hasn’t happened. Okay. So chart interface.

It is good when you’re conversing with an intelligent human, but again, there was a paper published by Microsoft around eight days ago or 10 days ago, where they said that longer conversations are going to go off rails. They are bound to go off rails. What that means is chat systems are not that kind of reliable system.

So there, there can’t be one interface that wins at all, but that doesn’t work. Then again, you come back to the original premise where you will have to have a good UX that the user loves, and that UX is going to change from person to person. There is not a singular UX that will go everywhere.

And that’s where the SAS is going to stay. Okay. Now the AI component is going to be much and much more profitable.

Either it will drive your effectiveness or it will drive your efficiency. Any one of these, but, you know, so be AI ready. You have to be AI ready, but while you’re being AI ready, I don’t think software as a service is dead because I don’t think my MD PhD gods are going to, you know, go and get this, because this is a small thing for them.

More important work is basically curing humanity. That is more important to them. Let us do this and you can take, keep doing that.

It works.

Kumar Siddhartha 46:49

Yeah. The interfaces for sure. I think I can tell you is that the customers surely want a chat bot that does everything that’s clear to us.

And just as you said, right, they don’t want to be scared that what happens when I click that button, they feel way more comfortable if there is a detailed set of instructions, even if it’s an agent that’s telling them all of that stuff. So from the product side, sure. The propensity, the market really wants that.

From a reliability standpoint as a business, are you really okay? If an agent makes a mistake and sends out a purchase order worth $10,000 that it was not supposed to in the first place, there’s a very real risk. so the financial side of it, the regulation side of it, all of that is like way too critical for you to kind of leave it completely to an agent at this point.

And so use case by use case, right? If it is something internal where you are okay with, let’s say a task getting updated, which was for the internal team. Sure.

Go ahead and do it. No one really cares. Do you want to send out a customer communication without you reviewing it?

That has a legal, legal implication. Maybe not. Do you want to review your purchase orders?

Your RFQs when you’re sending them out? A hundred percent, you want to do that. So you have to be very careful about all of these use cases. I think what the customers really want is they don’t want to spend all of that grunt work doing the same thing again and again and again.

And, you know, I was talking about the pipeline management management feature, just as you said, they want to do real work. They don’t want to sit on a dashboard and move things. So if you can move things for them, they love it.

They’re already, you know, you’ve already made their life much better. So even doing that is, which is like a fairly low hanging fruit at the moment is a great job for them. So I would say that, and you know, as I said, I’m an optimist.

So I’m waiting for these models. I’m waiting for AGI.

Omkar Patil 48:31

I’ll go and sit coffee at some cottage when the AGI comes. I’m waiting for that moment.

Kumar Siddhartha 48:36

AGI or an API. I think that would be best. That’s interesting.

Divyaanshu Makkar 48:40

Yeah. But yeah, no, I, I kind of agree. I think AI sort of embedded in their workflows works much better, gives them control.

Also. I don’t, I think like the startups that I hate the most are these AISDR startups, like, I mean, there’s certain rules that people really don’t want to automate. Like when we talk to our customers, like when I talk to a wholesaler and the lowest hanging fruit to automate always seems to be customer service.

Right. Because that’s like very defined queries that come to you. And you kind of feel like that’s, and there’s enough Klarna and Zomato automated it.

And you, but when I talk to my customers, they’re like, what’s your biggest differentiator? And in a lot of these industries, because you’re doing the same work, your differentiator is customer support. You said, I service my customers better than everybody else.

Now, would you automate the part that makes you unique, like completely, or do you just want to remove that layer? So I think there’s real world elements of it also, where people within certain roles just don’t want to eliminate or want an agent to do those jobs. They want to do those jobs more efficiently, which means you need to give them a workflow, but then automate all the redundant parts.

Like don’t make them go to like five dashboards to get some information or don’t, they don’t have to type the whole information, but like you can kind of generate it halfway. So meet them halfway kind of with AI.

Kumar Siddhartha 50:00

So several of these workflows is just obvious. Like if you know, a person like even, you know, anyone who’s been doing it for a long time, if they look at it, they will be able to tell you themselves. So the obvious ones, surely.

I mean, it is going away for sure. The more complex ones is more like a question of, do you want to cannibalize your business model around it? So does NetSuite really want to do away by saying that we will not charge you $400 per user per developer hour?

Do they want to do that? Or is that a feature for a new entrant? Let’s say like us, where, so for instance, what we, one of the features that we are building is, chat to an agent to get the UI ready for read scenarios, for instances, don’t have to call us for an accounting new dashboard that you wanted to create, which NetSuite will charge you, you know, $10,000 for, so do they want to disrupt their business model or is it an opportunity for us?

That’s something that, you know, the builder needs to debate.

Divyaanshu Makkar 50:58

Well, AI has real implications in how they experience these softwares, the entire software layer doesn’t just completely get disrupted, right? Like it’s, and then when people say SaaS is dead, I’m like, you, I mean, I don’t think you’ve run a business before, because if you’ve run a business and you have run an organization, you will know that, you know, it’s, it’s impossible to like, I, we still use HubSpot. Like I will not just go to an artisan and say like, okay, run my entire sales.

Right. You still need, and I still want control. Sometimes I want to make a dashboard.

I don’t want to just query and see. I just want to like see my data on a dashboard. So I think there’s certain parts of it, which like you still want there.

And then certain parts that of course, nobody wants to do. Like in our case, people get like orders on PDFs, right? Like if my wholesaler is selling to one of your construction home builders, they’ll send a PDF from maybe one of your ERPs will come to them and they’re team setting, just doing data entry.

Nobody wants to do that.

Kumar Siddhartha 51:55

Exactly. That is no hanging fruit.

Divyaanshu Makkar 51:57

You completely solve that problem. People really would love you for that.

Kumar Siddhartha 52:00

And it’s reasonably accurate now.

Divyaanshu Makkar 52:02

Yeah. Those things are reasonably accurate.

Kumar Siddhartha 52:02

That’s not a problem. But, as a product builder, building vertical SaaS, it would be such a relief to have, just go speak with that agent does everything.

Divyaanshu Makkar 52:13

Yeah.

Kumar Siddhartha 52:13

And you’d never have to go build up stuff, but that’s a, it’s a pipe dream. It’s a pipe dream right now.

Divyaanshu Makkar 52:17

Yeah. I think let’s cover a little bit more on the GTM side, because I think that’s where people get stuck the most, and have a lot of misconceptions about how it plays out. So say you’ve gotten your first two, three customers with their network, you’ve built with them, you have a functioning product, maybe two, three customers who can give a reference.

What really works in GTM? What has worked for you, for you guys? And also talk about what you tried from conventional wisdom that really just didn’t work.

Kumar Siddhartha 52:43

So, I mean, of course you always start with a wedge, right? You start with, let’s say one vertical that you really want to do well. You’re probably not even thinking at the start that you’ll probably build a vertical SaaS and somehow the market has, or the users are pushing in this direction.

So I think you have to be very careful about what you want to build. and how you build it, because, it might seem like a trivial task, but then, you know, you are literally building 20 different companies if you’re doing a full stack end to end. So on that side, you take the feedback, let the market guide you.

That’s what worked for us is we’ve let the market guide us in terms of what they want, but then we do the filtering. So we decide if it’s going too far or if it’s, and that is again, something that you only get to know over time and you will make several mistakes over there, a hundred percent. So we’ve had situations where we’ve thought we will build something, but then we’ve realized this is too off.

And then we kind of decided to stop that or hold that. And at some point of time, you know, some part of it is true, right? So to build a good product with enough vertical depth, you have to say no to certain things.

And that no is like very critical. We were not good at it. It took us time.

You can look at existing ERPs and existing products that are there to just see why did they do what they did, but then they’ve also kind of built it over 30 years. So, you know, use discretion and figure it out because you only have so many hours and so many hands to build it. And the one thing that the customers will not tolerate is a shaddy product.

So that’s, you know, you can’t push a product that is not ready for production. So in other industries, maybe you can have, you know, push something that’s maybe half baked or three fourths baked, but this one has to like work and you have to show them that it works. You wouldn’t even get into the door if it doesn’t.

So my advice would be is let the users guide you, let’s work for us and then decide where do you want to kind of draw the line and do take like a lot of feedback from users typically. a lot of people think that, you know, they can sit in a silo and figure out, I’ve been in that zone for a long time, but you have to allow the market to tell you where to go. And that’s the safest way to do it.

Divyaanshu Makkar 54:47

And from a sales perspective, I know you look at the product side, but what has worked and not worked on the sales and go-to-market side of things?

Kumar Siddhartha 54:56

So on the sales side, having like seasoned devs, the way of selling is very different and it’s not a one-shot selling. So you have to sometimes even groom prospects and get them to a specific stage. So it’s very detailed.

And if you’re starting from scratch, you would probably not have all of that understanding. So it’s always best to either have partners that you know, who’ve already done it in the past. So maybe someone from the competition who was not promoted and is hungry and they want to do it.

There are several such people that you can find out. But that really worked for us is making sure that people who’ve been in this industry, like our, our team has, Eric and Audrey, both of them have been in this industry for like decades. So people really trust them.

And when they’re seeing about us, then, you know, they’re not really questioning about anything. Then it’s like..

Divyaanshu Makkar 56:44

What point did you hire these?

Kumar Siddhartha 55:45

So we did this, after our demo day advices. So this was after, right after our seed stage.

Divyaanshu Makkar 55:51

Yeah. So you didn’t wait, like the advice that I was given early on was like, don’t hire anybody in the US till you hit, let’s say some people said a million dollar ARR mark, or some people said at least wait till series B, like all of those things, but what, what compelled you guys to like hire right after demo day?

Kumar Siddhartha 56:11

So we, we had already been selling before that. So we knew that, you know, people trust specific people from the industry. Trust people who they’ve seen, they’ve met, they’re meeting them at conferences.

They already have some form of a relationship. That trust is like super critical in this space because it’s expensive and sits on the critical line, critical path. So they’re not simply going to trust or nobody coming to them and saying, Hey, why don’t you use my software?

So you need that bit. So early on, as I, you know, it, it was the co-founder of the CEO Sneha. She was, you know, she was evangelizing.

She knew people, but then you hit a bottleneck beyond at some point of time because you only know so many people and that’s when to test out the market. Really. You need some people who can go do these kind of sales to prospects that you’d know, do not know in the market.

And that, that works. That’s a, I think that’s worked for not just us, but like several other companies as well. So that was one advice.

I think the other thing that we were told is do not hire a lot early on, but for vertical, you do need to hire like that. and I’d love for you guys to expand on that as well.

Divyaanshu Makkar 57:09

But what about you, both from, of course, I think you spoke about the product and what was the early days. But from, after you run out of your network, like what has worked for the company in terms of getting new customers.

Omkar Patil 57:24

So by the time you run out of the network, you should have gotten some traction.

That is, that is, important, but what worked and I think this is a, I feel that it’s a more of a generic advice is build something that already exists in the industry with minimum amount of efforts. If you can sit down, do it in one week, two week, build something already that exists in the industry, and then go to the places, just conferences and all of that, with all of these people converge, show them what you have, but at the same time, see what things are happening. Now, while you have gone there, your technological mindset from where you’re coming, it will give you an advantage that, okay, what is the state of the industry?

Because sometimes what happens in these verticals, it is not available in the general market. Like you can download Microsoft PowerPoint and you can start trying it out right away. That is more of a horizontal kind of thing, but these products that are not available in the market.

So see where, where that industry is, use those learnings and then actually get start down building, then you should be able to build those differentiators. And I am kind of mixing up the GTM and building part, because I think a lot of that, get a lot of customization is involved in this particular industry. But while that is also happening, the amount of time that is required until that particular product starts coming into effect, people start, you know, looking at the effects, it has drastically reduced.

Okay. Like in five, 10 minutes, people would want to see what they are, they were looking for. And that is possible with the current state of technology.

So with that, build whatever is currently exist in the market, and then either iterate on it or see what customer expectations are and build on it. That, that works very well.

Divyaanshu Makkar 58:56

Got it. Got it. I think a couple of things that we did differently was we hired salespeople in the US very early on because I, I, from my days of selling, I understood like, you know, if you’re trying to do this from India with an inside sales motion, this is not going to work.

And they, they want to, they want to, so it’s, it’s interesting. Like most of our customers, like prospects, when they come on a call, they would say things like, you know, I’m buying this for 10 years. You don’t hear that.

I don’t buy software for 10 years. I buy it for now. And I’m like, if I find something better, I’d switch, but they don’t think like that.

So if you’re buying something for 10 years, and it’s so critical for you, would you trust a team who’s like thousands of miles away, especially an early stage startup that might not even exist, that nobody knows about. So I think it comes back to trust. So you need to like go, I think the problem is that we trust playbooks, but their playbook should be customer backward.

Like it should not be that, okay, this is the playbook that works. I’ll adopt it. And that’s what I see people doing a lot of times, which I think just really be customer backward in your, your custom, people are still customer backward in product.

They’re never customer backward in GTM. They just try like, okay, these are the channels that work for three other people. Let me just try that.

But it doesn’t really kind of scale. And the second thing is that, I think show up in person for these categories, like just show up in person.

Kumar Siddhartha 1:00:13

You have to show up in person, meet them and, you know, be there for them when they need it, especially. So the trust bit is like super critical. That’s something that we also.

Divyaanshu Makkar 1:00:23

And the other, I think was no one in the market is that for SMP mid-market, you can’t have in-person, you can’t have sales, but like a sales led motion, you need like a more, I don’t know. Mid-market still people are okay, but they don’t believe that you should have somebody in the U S for mid-market sales. They think us based sales is only for enterprise.

But like toast was selling a literally free product, just a point of sale system, all feet on street, right? Because it’s customer backward. But it works like it’s, I mean, they’ve proven it. I mean they’ve proven it. So I think that’s another bad advice I think from people like if you’re selling to SMBs.

Kumar Siddhartha 1:00:54

Yeah.

Divyaanshu Makkar 1:00:55

You know don’t sell, you just can’t have a salesperson involved but you could like depending on what your expansion looks like and RR looks like.

Kumar Siddhartha 1:01:03

And also several times in these industries they’re probably they don’t even understand what PLG is. For instance they will never you know they use their time as stuff to do something else like you know family or do whatever hobby they are. So then they’re not like engineers or like developers who are sitting down and trying software.

So that’s never going to happen.

Divyaanshu Makkar 1:01:23

They’re not sitting on a weekend and thinking let me just try this software. I don’t think that ever happens.

Kumar Siddhartha 1:01:28

It never happens.

So and they’re used to people coming to them and selling. That’s the most critical bit is we feel that maybe it’s like a sale but no for them they are very structured processes around how sales works in the US and they’re used to it. So they don’t mind it.

They want to meet you and be exposed to that.

Divyaanshu Makkar 1:01:44

And then your shittier competitors will be in person and actually take away business from you no matter how better. I think the biggest misnomer is that a great product wins just with a great product. It’s just a complete farce.

It doesn’t win. Like you can have the fanciest interface but the shittier competitor will win if they really go in person build that relationship.

Omkar Patil 1:02:04

Relationship is required absolutely.

Divyaanshu Makkar 1:02:05

Because that’s what people care about. So it’s like vertical. You’re building a vertical company.

It’s like a combination of everything like your sales, like who you hire for your sales, how people want to in vertical SaaS. I feel when they talk to you they want to feel like you know their business and that is the wow factor. The wow factor is not a product.

It’s like oh man you get me. And I mean there are times when you know you’re able to do that in a you know meeting and you’ve just kind of sold it to them on the first go like before the product. If they really feel like you understand them.

Because fewer people do.

Kumar Siddhartha 1:02:42

For us as well like most of the time the product demo is basically the last. The final step where I am in and then they’re telling us some things. But typically it just closes after that.

But most of the work has been done before the demo.

Divyaanshu Makkar 1:02:55

They want to feel heard and they want to think like okay these guys get me. I think if you can create that feeling in your sales process, you kind of like almost there already. And I think then I think the things you said like around referenceability like don’t bomb an implementation.

If you bomb your first few like it’s very hard. But anyway moving on I think the one question that people keep asking me is like what is the why now moment for your business. And I think in certain businesses why now is very clear now like with AI.

Like you know there are certain kinds of why now which are quite apparent. But these are traditional businesses with problems that have existed for so long. And say no no billion dollar outcome is still out there.

So a lot of VCs just like what’s the why now. Like what’s changed. And what are your guys answers to this.

Kumar Siddhartha 1:03:53

So for us as I said right like the customers themselves are like kind of coming to us specifically to just look at products. They’re wanting to join webinars. They’re signing up for.

So there’s a lot of pull from the market itself where they want to look at alternatives while they are already on older platforms. And they somehow understand that these older platforms are not going to shift a lot. That is a critical piece.

So I think that’s happening across on and we’re seeing it across the stack. So we’re seeing it at the $500 per month product to the $5,000, $10,000, $18,000 per month product. Every customer is almost behaving the same.

And there’s a lot of excitement around what they can automate especially like how many dollars can they save. That is another critical piece. And then the more power users are like can you give me like a singular brain that understands everything and tells me.

So people are imagining all sorts of things and that’s a great thing to happen because then you know it’s an opportunity for you to showcase your product and get in the door. So that definitely is working at the moment.

Divyaanshu Makkar 1:04:56

Got it. So this general curiosity about what AI can do.

Kumar Siddhartha 1:04:59

Yes. And then you have to show them. So do a webinar.

Show them how you know and estimating let’s say you know for us we keep doing some workflows where we show them agentic workflows. How are they working? And then in those webinars when they actually see that you have a product and it’s working it sparks the curiosity.

So you’d see a comment where someone says can you try putting in that prompt. I want to see what it answers. And then it’s almost like them using it.

It’s like they are literally using it live though they are on a webinar and giving comments. So it tells them one that the product is real. So it’s not just some hogwash.

So that’s very critical for them because they typically are not very accepting people in terms of you know what software can do. So that’s critical. So doing things like these works.

Show them the product. Show them that it works. Be where they are.

All of that advice is you know all of that holds true.

Divyaanshu Makkar 1:05:48

What’s the why now for you guys?

Omkar Patil 1:05:50

So GPT-4 was a huge moment for us in the sense that where the industry was and what kind of things that. So it kind of brought the awareness. Accuracy not there.

It will never be there. But it basically opened up doors. That’s number one.

Number two access to knowledge is easy now. If you want to know about something you can just go to ChatGPT. Use free version and you will start getting it.

Now because of that explosion of knowledge has started happening. So over last three four years what we are seeing that number of publications every year regarding biopharma they are increasing by a factor of two. Like you know every year you will see you know double the amount of publications which have come in.

So there is an explosion in research that is happening and that you know value needs to be captured out of that explosion that is happening. And I think this is the right moment for us to come in and capture. Because in the end what what do we need?

We need you know good housing. We need you know good Medicare. Medicare in the sense that I don’t have to go to the doctor itself.

Like you know big pill and I am I am free of all the disease. That is where humanity is going. And I think we are at the right moment right spot where we can make the impact and you know get the best out of it.

Divyaanshu Makkar 1:06:59

Got it. I think some of the why now’s in these industries are also like just long-term why now’s right. Like the newer generation is taking over businesses and they don’t want to run it on spreadsheets.

Kumar Siddhartha 1:07:08

That is true.

Divyaanshu Makkar 1:07:09

They’re not like a singular moment why now that okay this is my why now and this is why it needs to exist. But they’re just long-term trends. I think just the macro overall has been so challenging.

People are looking for cost-cutting because of the challenging macro in the US right now at least. And so probably true for construction as well where yeah.

Kumar Siddhartha 1:07:25

So that’s the cost-cutting bit is there. Then there is lack of trained people who can join the workforce. That’s very relevant.

And for existing incumbents to catch up it’s kind of the same perplexity versus Google story but then it’s like a NetSuite versus us.

Divyaanshu Makkar 1:07:42

Google seems to be catching up though.

Kumar Siddhartha 1:07:44

Yeah I mean.

Omkar Patil 1:07:45

I predicted that that would happen one year ago.

Kumar Siddhartha 1:07:47

They have to and I would be surprised with the amount of ML and AI work that they’ve done over the decades.

Like we’ve used their APIs.

Divyaanshu Makkar 1:07:55

I don’t know how many startups died when in the recent Google I.O. Like with the things they launched like a lot of things that people are working on.

Omkar Patil 1:08:02

So if you’re a wrapper around ChatGPT or Gemini that is going to happen. But what that means is like you know people don’t understand this that there is so much knowledge that is available for you that you can try out and get you know a very even if you want to start from zero coming going from zero to hundred it is not going to take that much time that it take five years ago. So you know people are completely misunderstanding this particular fact.

So okay there is Google. They are doing something but you can start from scratch and do something amazing in a matter of you know short amount of time. So I think that part is there as well.

Divyaanshu Makkar 1:08:35

Yeah so it’s a combination of course all the macros the just expectations from technology changing with younger people in the workforce labor shortages which is a big thing that’s that’s kind of driving this. And in general I think people now at least believe that these things can be solved. I think maybe five years back they just believed that this can’t be solved.

Now there’s curiosity that maybe I can solve it. So that’s why they maybe come into the demos and they’re at least trying to look out.

Kumar Siddhartha 1:09:02

So they’re way more involved now.

Divyaanshu Makkar 1:09:03

Yeah and like every CEO is at least even in a traditional setup thinking about like am I missing out on something. There’s a fair bit of FOMO.

Kumar Siddhartha 1:09:11

There is.

Divyaanshu Makkar 1:09:11

I think.

Kumar Siddhartha 1:09:12

Huge FOMO.

Omkar Patil 1:09:12

Huge FOMO.

Divyaanshu Makkar 1:09:13

To at least look for solutions for everything. Got it. So with this accuracy concerns and I think it started with the co-pilot and now I think market is seemingly moving towards more agentic and less co-pilot.

So what are your thoughts on sort of the agentic?

Omkar Patil 1:09:32

So I think you know I will have a different thought process with rest of the people. Agentic systems again they are prone to a lot of errors and I said earlier that you know the longer the chat happens, longer the communication happens, it is possible that the system is going to derail. Now and why does that happen?

So think about it the decoder only models that we have in the market. They predict the next token and they do it with there is a softmax layers which basically predicts that there is you know 99% possibility that is there. But that 99% possibility it is probability that is you know extremely hyped due to that softmax layer.

Original probabilities are not like that. So what that means is you have to consider what is happening. What is the second next token to it?

What is the second next token to it? Now with that you predicted 1000 tokens. So with 99% accuracy multiplied that to each other you will get much lower accuracy and multiply that with for you know whatever the actions which are going to happen in the future.

I don’t think agentic systems will work the way they are at least in my industry. But they could work in a way where your code. So I’m moving two different things.

Number one there is action and there is state. If you are managing the state using these agentic systems but action is always taken by some deterministic system, it will work. That is how I feel that agentic systems will work.

Another way is core you know fundamentally changing the way language models are designed. So you know one thing that happened in this google IO is google introduced a diffusion model and this is you know to you know call throwback to their last you know birth like architecture. They are using that technology to predict in between world.

They basically come start with noise and they come up with in between world rather than next token prediction. So when you go on that kind of architecture it solves a few more problems. What that means is if your benchmarks if they have stagnated at 83% they will go to 87%.

But again your knowledge itself it is lossy compression. You are not storing all the knowledge that is in. So traceability is not there, repeatability is not there.

Accuracy is always going to go down. So I am kind of bearish on agentic system but I am looking forward for the day when the AGI comes and I am you know retired and sitting into cottage you know having drinking a sip of coffee and all of that. I would love to you know be in that future.

Divyaanshu Makkar 1:11:47

What are you I mean in your experience what’s are you are you guys seeing traction with agentic moving away from copilot?

Kumar Siddhartha 1:11:54

So very similar to what you said right is leave the decision making to the user and don’t act on behalf of them without them having been an approver. For the next at least good months our focus is help them become approvers but stop there. Don’t take actions on because the cost is too high.

You don’t want to be making errors and then be held liable for those errors. So that’s clear and from a product standpoint I think the way it’s evolved is earlier there was on-prem and then from on-prem it became SaaS. Now from the expectation is SaaS with an agent that is doing stuff for you where you are involved but sparingly.

So you’re mostly kind of dictating and then from there it goes on to how can you reduce the size of the organization. That’s where I see it tending to is some things that are working very well over a sufficient amount of time there you want to maybe you know whittle it down to those use cases and just completely hand off. But it has to be an iterative process that’s the only way that I think we can do it.

This kind of throwing caution to the wind and you know just saying that we’ll just allow the agent to run berserk and do certain things. I don’t think that’s I don’t think any customer will allow that.

Divyaanshu Makkar 1:13:01

Yeah I’m kind of very bearish on this whole AI employee concept. I feel like it’s taking it to another too extreme. I don’t think I mean there may be certain roles which could be like completely automated but for most roles that people do like I mean they’re making a lot of decisions on not every decision is like defined you know checklist of things that you know you look at and like assign weights right like you’re constantly making decisions on less information, gut, instinct, you know whatever like reacting to things. So I feel like small problems yes you can automate within a certain role but I don’t know this whole AI employees this seems like too.

Omkar Patil 1:13:44

So I feel that it’s an you know exponential function. So you know exponential of one is always going to be exponential of two. So whoever is using that particular system they will get exponential benefits but again it depends on what kind of problems are you seeking to solve and what kind of questions are you asking.

All of that is there but how are you going to take it, how are you going to implement it and at what speed. So in the end human is there a human is not going to go away and you know agentic systems you know they will work but when you go to state rather than the action.

Divyaanshu Makkar 1:14:14

Yeah no I mean there’s definite productivity increase you should pick up apart everybody’s job roles and automate but I think the final at least on critical things decision making has to come back to at least a human in the loop.

Kumar Siddhartha 1:14:26

I think that will become standard yeah in some time people will get used to the fact that they don’t have to do a lot of that grunt work at all yeah and so if you’re not doing that that’s the bare minimum that you have to cover as a new software. The very least.

Omkar Patil 1:14:38

And those are also products right like you know for an example Microsoft PowerPoint like you know or excel sheet people are using it on the paper PowerPoint came in or you know excel sheet came in and that was all automated so these kind of cycles they are going to happen this is one of the cycles yeah we’ll wait for AGI got to come in we’ll join the religion later or not that happens it happens.

Divyaanshu Makkar 1:14:57

Cool I think guys we’ve had a very good discussion today covered a lot of key topics how to build how to sell all the bad pieces of advice you’ve kind of received over the years but just helping other people who are trying to think about vertical markets or software what is the one piece of advice that you would have for that person and something that you probably knew when you were building out.

Kumar Siddhartha 1:15:24

So I think I lagged as I said right is build with patience it will take way more patience than it would take to build a horizontal SaaS so be prepared for that like at least mentally and this is a real marathon you know vertical SaaS is actually a marathon this is you would expect like a long long grind by the end of it I mean I would expect it to be you know a company that stays for another 30 years just like NetSuite it is but yeah it requires a lot more patience and you have to you know you will make several errors so be ready for that and you know just keep grinding there’s no day off you have to keep doing it.

Divyaanshu Makkar 1:16:03

Omkar?

Omkar Patil 1:16:04

I would say be empathetic and humble because people who are you know you’re working with this industry they are the gods basically you have to be extremely you know receptive of the knowledge that they are giving you have to be always learning and you know then you should be able to incorporate those learnings so you know take his advice and you know the ideas that stay survive for longer and at the same time incorporate those learnings, you will do just fine.

Divyaanshu Makkar 1:16:29

Yeah I think my advice would also be just be customer backward to the T in every decision you make not just product be it GTM be it how to set up your customer support like you have to just be customer backward and you’ll build a long durable business but you will if you’re looking for virality do not build in vertical SaaS.

Omkar Patil 1:16:44

No, no doesn’t happen.

Kumar Siddhartha 1:16:45

Doesn’t happen overnight

Divyaanshu Makkar 1:16:46

It will not happen overnight especially in these kinds of industries it’s just not how the industries work.

Cool I think we’ll move to the last section which is a rapid fire round. So let’s keep it rapid. I’ll I’ll say the question and then maybe you start and then you give the answer Omkar, right?

Google or perplexity?

Kumar Siddhartha 1:17:09

Google

Omkar Patil 1:17:11

Oh we built our own search engine NolQuest so you know everything from scratch from hardware level so you know this question does not apply.

Kumar Siddhartha 1:17:17

I still use Google and ChatGPT more because I find like Perplexity answers are too summarized I mean I just want to go a little bit deeper like you know it gives me the Inshorts feeling and I think that was the worst way to consume news was in shorts and that’s the same feeling I started getting with perplexity it’s not conversational enough, at least for me.

LLM open source or closed source closed source?

Kumar Siddhartha 1:17:39

Closed source for now. But we’ll see that might change very soon.

Omkar Patil 1:17:44

Both because they kind of complement each other so what happens is that when closed source are becoming much more arrogant there is some open source person comes in and punch them in the mouth but idea is that there are you’re going to need cloud solutions you’re going to need on-premises solution as well in here in this vertical industry and you should be ready for both so both of them work very well.

Divyaanshu Makkar 1:18:07

I think from a current perspective I agree like Closed. I’m looking for more accuracy and like applications but but I think from an ecosystem perspective it actually need both.

One software product you love to use at work.

Kumar Siddhartha 1:18:29

I really liked like all of these JetBrains IDEs so if you’ve ever used something like Eclipse etc you know what I’m talking about and then even you know Cursor and the rest are like a much better upgrade on top of this so you can actually it’s unsurprising because in JetBrains hasn’t caught up to how well Cursor has done and I would have expected these guys to kind of completely take that market away so yeah I mean all all these ideas are like great. Yeah.

Omkar Patil 1:18:55

I use MyScript and Nebo so it’s a it’s a application it’s an app on iPad and it helps drawing and note-taking much more easier so you know when you have to explain difficult concepts it becomes much more easier using that so I use that very thoroughly

Divyaanshu Makkar 1:19:15

My latest like software that I’m really enjoying using is Relevance AI actually no because for a non-technical person like it’s a very easy way to build agents actually like whatever…

Omkar Patil 1:19:26

Yeah new definition. Microagents was introduced here.

Divyaanshu Makkar 1:19:29

Yeah new definition. I’m not trying to automate my company away actually we are hiring so if you guys are looking for a job please reach out to us across all roles.

One business buzzword you think should disappear?

Omkar Patil 1:19:46

Role, like again you know the job role that people talk about I think it needs to disappear like you know it is and I think there was an earlier podcast here as well on Neon job role doesn’t exist in the age of AI you have to do what needs to be done to get you know take the ownership of that particular part that you are doing and execute it so whatever is required to learn that because learning it’s you know very difficult thing you have to go to years of training that knowledge is available to you own what you have so the role is is not what one should be looking for.

Kumar Siddhartha 1:20:19

Yah so I would stick to my industry I think the word ERP itself is a misnomer it’s like 40 50 years old it needs to go away and like people you know you have to name things as you envision them to be and that needs to really change so that people start realizing that that shift has happened stop using that word completely hopefully someday.

Divyaanshu Makkar 1:20:37

I think AI native as a word is being used and thrown around too much. I think everything is AI native till the customer comes and asks for a workflow to actually solve that problem I think. Yeah I think that’s a that’s a good buzzword I’d want to get go away with.

What’s one vertical SaaS company outside your sector that you admire and why?

Kumar Siddhartha 1:21:00

So there’s C3 AI they work with very limited set of customers big clients but then I totally feel that if you kind of with vertical SaaS if you reduce your scope so work with not too many customers specific customers have like bigger ACVs but then you can do like a much better job especially with there’s two schools of thoughts right now right is that you’ll have one kind of AI that does everything and then the other is you know you have to sit down bang your head with every single workflow get it done right and then release it but because every single company has like 50 percent they’re the same the rest of the nuances are different we’re all seeing it so you have to if you want your vertical SaaS to succeed inevitably focus on each company to a decent amount of depth which means you don’t have the time to focus on mass market which means that stick with specific customer…

Divyaanshu Makkar 1:21:56

Hyper verticalization

Kumar Siddhartha 1:21:58

Yes

Divyaanshu Makkar 1:21:58

Like even within construction, just several subsegments.

Kumar Siddhartha 1:22:01

Exactly and and it’s good because those companies that will utilize AI in that way will have way more to get out of it than the other companies so it’s a win-win for both them and for us so we would I would like for you know for us to kind of go in that direction and see them

Omkar Patil 1:22:14

I would say Palantir and you know. Lord of the rings reference that is number one like all seeing eye from Lord of the Rings but the kind of things that they have done in defense at least that’s amazing and you know they have done exactly what any vertical SaaS company would want to do build relationships build a good product you know make it sticky all of those things so you know Palantir is something that I’m very much uh you know I’m liking it I love it.

Divyaanshu Makkar 1:22:42

Interesting.

I mean I have two actually I really I mean and both of these are last generation SaaS but I really like Toast I feel like they’ve flipped a lot of the playbooks like you know SMB SaaS but in-person sales I at least have taken a lot of learnings from how they’ve executed it’s very customer backward but there’s another company called Slice which is more to your point they only sell to pizzerias within the US only independent very small like 40,000 pizzerias I think overall but they’ve done fairly well by going hyper verticalized which is like even selling pizza boxes to them customizing them like they run I don’t know now they must be doing it with AI but tele-operations for these pizzerias so you can go like fairly deep in that way.

So I think these two companies I mean I really like and I’m constantly debating like which one do I want to be like do I want to be for all wholesale distributors or do I want to be this for like one category it’s just like the question that keeps me up at night is that only. Like how deep do I want to go and what’s the framework around that.

Yeah perfect thank you so much guys this was a very invigorating. I learned a lot about AGI and about you know what’s working for you guys and I think you guys have had very interesting journeys so thanks a lot for sharing it with the audience and I hope new vertical software founders can learn something from these this conversation and avoid some of the mistakes we made early on thank you so much.

Omkar Patil 1:24:12

Yeah great experience.

Kumar Siddhartha 1:24:13

Thank you