He left a $2B ARR company to build AI agents—then hit $1M ARR in < 6 months | Amit Shah, Founder of Instalily

Amit Shah (00:00:00) :
The vision was very clear. How do we amplify the human work, instead of how do we replace human work? Because our belief always has been that we want to empower human beings and human teams to do the most demanding cognitive work, and amplify that impact using AI technologies. Instead of a product market fit, I always think about operator market fit. Do you understand the operational rigor and the operational nuance to bring any solution to that vertical market? Because our belief is one size fits none. So you have to really understand the nuances. We are accelerating rapidly. We were able to hit that first million ARR within a matter of months and then we are in the second year of our sales cycle. And confidently past that triple, triple and actually accelerating our growth. More than ARR, what I measured myself is what impact have we created for our customers.

Previous Guests (00:01:06) :
That's product market fit. Product market fit. Product market fit. I called it the product market fit question. Product market fit. Product market fit. Product market fit. Product market fit. I mean, the name of the show is product market fit.

Pablo Srugo  (00:01:18) :
Do you think the product market fit show, has product market fit? Because if you do, then there's something you just have to do. You have to take out your phone. You have to leave the show five stars. It lets us reach more founders and it lets us get better guests, thank you. Amit, welcome to the show, man.

Amit Shah (00:01:35) :
So excited to be a part of this show, Pablo. Thanks for having me on.

Pablo Srugo  (00:01:39) :
I'm excited to have you here, man. You just, so you've launched a company called Instalily, and it's in the kind of AI agent space. You just raised a $25 million Series A, we'll get into all of that. Maybe as a starting point, tell me just a bit about your background. You know, just the stuff leading up to starting this business.

Amit Shah (00:01:56) :
Yeah, so I've had a, awesome experience in what I call sort of two decades in operationally complex businesses. So after starting my career at McKinsey, I ended up doing a bunch of startups and ended up in operational roles. That involved a lot of complexity from grounds up. So think about businesses that rely on deep distribution, supply chain, logistics. Companies like Row Flowers, 1-800-Flowers, Blue Apron, that cut across very deep verticals that underpin not just the U.S., but the global economy and what stuck to me was that the pattern was always the same. It's really motivated smart people trapped in dumb systems. So what stuck me in sort of those twenty years of operational work, was how much human work was trapped within the systems of record and really not getting unleashed to the full potential. So that was the sort of deeper motivation I had when, I was thinking about how do we really resolve this now that we have a system of intelligence around us. Can we really bring this, this sort of new age of, you know, it's almost like when electricity first started. It was not just about bringing that electrical medium to work that was being done in analog state, but really rethinking about the role of human and humanity within that. So it's a very deep motivation, but born out of years of, I would say, scar tissue and really being in the trenches. Observing, managing, and ultimately trying to derive an outcome in that medium. And just to give you a scale of it, you know, almost half of the global GDP is ninety percent still manual. So if you think about deep verticals, everything physical that surrounds us goes through that same complexity and that is the part that is still deeply manual. So it's, you know, the world talks a lot about services and those businesses or startups. Those are easier, I would say, things to resolve. The real depth is in resolving this sort of physical goods economy, as I call it, and how manual it is. Because it actually affects all of us every moment that we live around.

Pablo Srugo  (00:04:37) :
Maybe for context, before you started this company. What was your last role, the last thing you were doing?

Amit Shah (00:04:42) :
Oh, sure. So right before I started this company, I almost spent a decade at a public company called 1-800-Flowers. Helping them scale from just under $500 million to about $2 billion plus.

Pablo Srugo  (00:04:55) :
What was your role there?

Amit Shah (00:04:56) :
I started off managing all the product and engineering around mobile. Which was the growth channel when I joined it. So think of it as building a growth team and then over time, I took over all of the growth and marketing channels. So I was the chief marketing officer and then after that, I was the president of the company.

Pablo Srugo  (00:05:16) :
How do you decide to leave that to start Instalily?

Amit Shah (00:05:19) :
Yeah, I think so. As I was saying, you know, I was deeply motivated by sort of this in-situ experiences. If you will, of managing such complex businesses and not just at 1-800-Flowers. Even at a board level, when I joined Blue Apron as a public company director, I observed the same sort of complexity. So there were two sort of intersecting origin forces that gave me and my co-founder confidence to start this company. First was this realization that all the systems of record that surround any human operator in these businesses, actually take them away from the customer. So there was this increasing distance between sort of the atomic unit of work and who that work impacted or who they wanted to impact through that work, right?

Pablo Srugo  (00:06:13) :
You're talking about logging like on Salesforce? Spending all the time getting all the stuff in, so that everybody above you can see what's going on. But it doesn't actually drive value for end customers.

Amit Shah (00:06:22) :
Exactly, and think about the quality of human cognition being used. You know, it is clicks and taps to just get from one point to the next in that system of record. Lift and shift of data, answering emails, extracting out ledger entries. This whole sort of ball of work was literally the least exciting human work that was possible and yet it was compounding at a massive rate, right? Because more and more software is getting written, more and more software is getting deployed, but the quality and quantum of human work was getting degraded.

Pablo Srugo  (00:07:01) :
What's driving that? Is it really a management thing, where management is just so far removed from end customers. They need all these lower level people to do stuff, so they can see what's going on but ultimately is just so much wastage in order to deliver that visibility?

Amit Shah (00:07:15) :
When we think about sort of automation and sort of this complexity, and drudgery. I don't think it is only about sort of back office work or lower end work. Think about the whole stack, so there's almost a three-layer stack that surrounds a lot of work that we see. At the bottom layer, it's sort of the data stack, if you will. Wherever that system of record resides, whether ERPs, CRMs, emails, people's texts, what have you, right? On top of that, you have this sort of systems of record. As we call it, software and the promise of software was always that it would make the distance between sort of the cognitive layer. Which is the third layer, the human being on top of the systems, easier. But where did we end up? It started to make our lives more and more complex. Because now it's not just the management saying, hey, we can automate out certain things, we can make certain things more better, if we capture all customer details in a CRM. But then we became keepers of that system of record versus the system of record doing anything for us over time. So I joke with people, you know, we have a great working relationship with the systems of record. Whether sales forces of the world or SAPs of the world. But I'm yet to find a very happy salesperson who puts up his hand and says, I love waking up every day and entering all my sales records. It just doesn't happen, right? So I think it is not just the lack of intention. I think management always had the intention of how do we empower the teams more. But actually the nature of how work itself gets done was getting more and more mediated by the systems of record versus the clarifying question that we should be asking. Is that, if work really is about taking actions. How do we build systems of actions? To take over this cognitive work that is built on this top of systems of record?

Pablo Srugo  (00:09:18) :
So this is kind of, one of the issues that you're seeing. That you've been seeing for a long time, I think you mentioned there were kind of two forces that came together.

Amit Shah (00:09:24) :
Yeah, so this was the first force and I think the second force. Which is what we are experiencing now, is the incredible. I would say, growth in both the aptitude and I would say ability of this large language models, AI systems in general. To create atomic units that take over the cognitive work. You know, we refer to them as AI agents and when we started this company, it was very early days.

Pablo Srugo  (00:09:55) :
When was it?

Amit Shah (00:09:56) :
This was in spring of '23. So it was super early and it was just around the same time the first. I would say, look of what can happen beyond sort of the ChatGPT starting point, started to emerge.

Pablo Srugo  (00:10:12) :
We were very much in that just, post ChatGPT moment.

Amit Shah (00:10:15) :
Just that moment and it was still very early. Because I think the point that time still was that there is a lot of hallucination. There is no way you can have an atomic unit of work getting done at all, but there were some early research papers that the team started reading and we got really motivated. Because I really appreciated the fact that I came from an operator background, so did my co-founder who worked with me for twelve years at 1-800-Flowers. That actually if we can take this AI agents into production within this environment. You can really start changing how work gets done and it was very early. And I think, you know, the first six to eight months when we built out the core technology. We used to get a lot of pushback, that this dream or this vision is never going to be realized. Because this technologies are still at an infancy.

Pablo Srugo  (00:11:10) :
What did you want the AI agents? Because especially back then, and even today, I think a lot of the AI agents are hyper-focused. What was your vision of what an AI agent would do? You know, we're talking over two years ago at this point.

Amit Shah (00:11:23) :
The vision was very clear. How do we amplify the human work, instead of how do we replace human work? Because our belief then as now always has been that we want to empower human beings and human teams to do the most demanding cognitive work. And amplify that impact using this AI technologies.

Pablo Srugo  (00:11:51) :
And what's an example of that? Especially back then, how might you augment a salesperson? For example, at an enterprise company?

Amit Shah (00:11:57) :
Sure, and I think it's actually a very interesting conversation. A lot of companies that we see around us, think about amplifying a salesperson's role by automating certain things that they hate doing.

Pablo Srugo  (00:12:10) :
That's where my mind goes. Oh, fill the CRM for example.

Amit Shah (00:12:12) :
But if you actually think about what makes the salesperson achieve, let's say the highest commission. It is actually ability to discern signal from noise. How do you really figure out which signal is the most prescient one to work on right at that moment, and then measuring the impact of whether or not that action led to a better outcome than doing a random selection through the day. It is going to that level of understanding and then making sure that not just that each of these salespeople has that individualized plan of action. But also developing a system of learning that is improving over time.

Pablo Srugo  (00:12:53) :
It's interesting, it is almost like a third bucket, right? Because you've got your replace bucket. Which is, here's a job, AI is going to do it. You've got your augment bucket, which usually does look like, what's the worst part of the job? Okay, let's automate that and then this is almost a new class. Where it's like, nobody's doing this stuff right now. Nobody's going and asking a thousand different salespeople what they're seeing, disseminating it, looking at actual hard data, putting it together, and then delivering personalized plans to every single salesperson. It's just not something that happens in an organization. So it's almost this new layer that then, if it can work as you describe it. Then ultimately augments everybody but it doesn't do it by replacing part of their job or taking part of their job away. It actually just does it by providing new insights and information at the right time to the right person.

Amit Shah (00:13:39) :
And deeply personalized. And working for the human versus in spite of the human. I'll give you another deep sort of insight that early on we had working with our early customers and that is the idea that, the work that we think can be mediated by AI and AI agents is not just the work that we see exists right now. But it's actually work that the organization was never able to get to, because either the scale required or the cost required to get that done was extremely high. So I'll give you an example, if you can read every purchase order and every contract written with a B2B supplier. Let's say, that you might be using. It will reveal deeper patterns about not just the obvious ones that, hey, where did we make the most margin. But say which set of suppliers actually misbehave the most in very insidious ways, right? We just continue doing those works. So I think what we have uncovered also is by deploying code as work by having this AI agents that we call instaworkers and we can go deeper into sort of our stack. By deploying this instaworkers, businesses are unlocking things that they were never able to do before. So the value creation on the third lever that you are talking about is not just about the obvious places where the human beings are currently even working. It is actually reframing it into places where the human beings were not even able to go to. We have another customer, for example, a $2 billion distributor of commercial restaurant equipment, and so think about the back end of Taco Bell. When a fryer goes down, they get the call to send out a service technician with a part and so they were able to service a certain number of this QSRs. But they were not able to go to a deeper level of servicing a larger number of them. Because A, there's a labor shortage with service technicians. It's very difficult to find, but the second thing is that they had to go anytime they got a call on some fryer that was very deep in terms of trying to find a part. They would have to literally go to the supplier website, download the manual, read it and answer. Now that what we have done is that this instaworkers surround each one of this service technicians and they are able to amplify them 20x to 50x. So suddenly the business can now go out and bid for more contracts. Because they have not just escalated and rapidly increased their workforce, but they can do it with the same level of confidence as hiring a human workforce. So this is now starting to go into uncharted territory and literally compounding the outcomes, and the ability to do things which were not possible before. And I think a lot of entrepreneurs miss this sort of deeper detail that it is not just the existing market that can be unlocked. But you can actually position competitively the company in a very different shape and form to extract out even more enterprise value.

Pablo Srugo  (00:16:47) :
So now that we get kind of, the sort of things that you're enabling. Maybe walk me through the storyline. So once you decide you want to start this with your co-founder, what's the first step? Do you go out and raise a round? Do you start building with a design partner? Where do you go?

Amit Shah (00:16:59) :
Actually, because I joke with people, instead of a product market fit. I always think about operator market fit and I think very deeply about, do you understand the operational rigor and the operational nuance? To bring any solution to that vertical market and by choice from word go. We were very focused on vertical outcomes, because our belief is one size fits none. So you have to really understand the nuances. Because ultimately, a lot of impactful work requires domain specificity, right? So what we started out thinking about that if we take this three layer stack, data, software, human cognition on top of it as sort of this trifecta of how work gets done. What we started realizing is that we needed two sort of starting points. To really make work come down to the code level, you needed first what we call an InstaBrain. A rich contextual layer that allowed you to codify the tribal memory and the knowledge within a company. So think about not just your metadata in your Salesforce or SAP, what have you, but how is it really getting used by the operators. You know, and sometimes it is in fact yellow post-it notes around people's monitors. So first of all, codifying that into a contextual layer that we call the InstaBrain and the InstaBrain. The key thing that we discovered early on, which I think is a differentiated IP that we have. Is that very much like the human brain and actually, one of our founding engineers is a neuroscientist and a computer scientist. And really helped us think through this, is the idea that it needs to be editable. Like the human memory is both pruned and is edited on a daily basis. So, for example, if a sales team last quarter wanted to just focus on top line growth. A lot of the sales processes would be written to say that, hey, to chase top line growth, give discounts to customers, book the same. Whereas this quarter, let's say they shifted to margin focus. If you run a simple agent, it will be overwhelmed by the context that existed three months ago versus three days ago, right? So you should be able to edit it. You should be able to almost treat it as human memory that, hey, the context has shifted. I need to now edit out the memory, right? Or I need to bring in that memory cube internally into that. So think of this as a continuously learning rich layer, but with all this depth of utility built in and then on top of that, we have our AI agents that we call instaworkers. Which are specific domain trained Insta workers. So they clearly understand what is the difference between a certain size shingle versus a material that you put down in the floor. So this ability to understand that richness that exists within the vertical makes them not just able to get the work done. But then they are able to bring the feedback into that InstaBrain at a much deeper, richer level. So ultimately, what we are delivering to companies is not just code as work. But really systems of learning embedded within these systems of action that are self-improving, self-organizing, and getting better at every round. And this is what I mean by winning games in succession versus games in isolation.

Pablo Srugo  (00:20:28) :
We have tens of thousands of people, who have followed the show. Are you one of those people? You want to be a part of the group. You want to be a part of those tens of thousands of followers. So hit the follow button. So did you build all of that before going to a first customer or did you build that in tandem? Did you get one design partner and then start collaborating with them? How did you set it up?

Amit Shah (00:20:48) :
We actually evolved. That's a great question. I would say, where technology existed in '23. A lot of our focus was first making multi-agent, multi-step architecture work. Because if you think about most human work, it is not just about automating an agent. There's a lot of frameworks that will do multi-agent single step. But the real sort of complexity comes when you do multi-agent multi-step, right? So we focused a lot on making the multi-agent multi-step in instaworkers really perform within these systems and the other choice that we made. Which I think is very different from the starting point of other companies, is that we deeply thought about the starting point, not as rip and replace. But as a human worker would do on the first day of the work, literally have the agents log into the whatever systems existed. So it gave us a faster starting point, and that is true of most environments, right? You are not going to go in and just say, hey, rip out all your systems and suddenly AI will start working, AI agents will start working. So that was our initial area of focus.

Pablo Srugo  (00:22:01) :
And who was that with? Who was your first customer? How did you land them?

Amit Shah (00:22:04) :
So we had two first customers that we were very lucky with, one was Parts Town USA. Which is this $2 billion distributor of commercial restaurant equipment and the second was SRS, which is now part of Home Depot. Which was at that point about a $12 billion distributor of building products.

Pablo Srugo  (00:22:23) :
How did you get them?

Amit Shah (00:22:24) :
You know, on two aspects, we were very different from other companies that were building around the same time. One was sort of deep operator networks. So we knew within the distribution space, sort of who are the right players to go and start having the conversation with. But the second thing was also that a lot of our focus is trying to figure out, which verticals are actually deeply ensconced by this manual work. So we went out and targeted those verticals as well. And we continue to do that, whether it's construction vertical, industrial goods, we think those are trillion dollar verticals. They have a lot of space to grow and if you could dominate any one of those verticals, that's literally billions in ARR. So it was more of a deliberate choice and I think it was based on sort of the knowledge that we had of who the right operators were.

Pablo Srugo  (00:23:17) :
How did you structure it? Was it these paid pilots? How did you get access to their data? Because it's a big ask if you think about it as a small startup. As much as you have, you know, impressive background. Still a very small startup with a kind of crazy idea. Certainly at the time and a very big company that has to give access to, you know, with proprietary data and systems, and so forth.

Amit Shah (00:23:37) :
It's a very good question. Look, as much as every entrepreneur will claim that they had some magic sauce, I think it was a mix of luck and progressively leaned in operators. You know, I think when we first started having the conversations, we have a great partner. Still a great partner to us, Patrick. At SRS, Patrick Garcia, who is the chief digital officer and the head of AI there. And he almost was open to trying out this vision and navigating with us the internal powers that be to open up those systems. And I see the same opening now. I think there is a desire on part of enterprises to really see if they could make a difference both in their sort of human workforce and the quality of outcomes that they deliver to their customers. So I think if you speak the right language, you are at least able to get the starting point and then the second thing I would say, which was critical for us. Was because we were chiseled within the enterprise governance and risk mindset. We were able to go to the CISOs or the CTOs of those companies and say, hey, we know how to build systems that will go through the scrutiny of InfoSec. We know how to really write agents that can be audited and can be evaluated on those constructs. So from day zero, we focused on SOC compliance. We focused on sort of making sure that our systems were robust instead of flashy outcomes. Really thinking about, hey, how do we make sure the risk profile fits the appetite of this billion dollar enterprises? Because otherwise there is no starting point to your point, right? Of having this conversation, because you'll get blocked over time in terms of accessing the data.

Pablo Srugo  (00:25:29) :
For enterprise, you kind of have to get the boring stuff right. I mean, otherwise, it's not going to go anywhere and talking about something that is a little bit. Let's say, more status quo is how did you set it up in terms of success? How do you define in these pilots, we're going to go, we're going to build this InstaBrain, we're going to build these workers, and ultimately, we're going to augment all your workers. How do you define that? Is there kind of an ROI thing that they're looking at?

Amit Shah (00:25:52) :
A hundred precent. Having come from an enterprise outcome. I always sort of joke that, if someone promises me that they can slice the bread better with that technology. I generally end up eating the loaf as is, because the total cost of operation after you slice the bread is generally way more than what was promised to you. So we were very metrics driven from day zero and the way we define sort of outcome was three ways. One was time to value. How do you deliver measurable outcomes within weeks, not even months? Very clear win for that salesperson, right? Did they achieve a higher velocity of sales? Was the quality of sales better?

Pablo Srugo  (00:26:37) :
That's a huge one, by the way. Time to value I think is an underrated, underappreciated concept just in general. Less value but in less time, a lot of times beats massive value in a very long time. Because you don't get there.

Amit Shah (00:26:49) :
And I know this because I've always been, I would say most of my career been on the buy side. Instead of the sell side. So I asked the question to begin with that, hey, does this pass my smooth test or not? And it has to be in weeks, right? The second is sort of expansion velocity. How do you how do you really help once you have the InstaBrain down? Can you really help the operator achieve a higher velocity? So I was giving you the example of our OEM equipment distributor. That was suddenly able to go out and bid for way more contracts. Because now they have an expanded workforce that they can rely on, right? The third way we measured it is sort of engagement depth. So not just that you have created value, but you are able to create continuous value, which is compounding. So it's almost like now our customers have wish list of our instaworkers that they come to us with. You know, that, hey, can we deploy this agents instaworkers. Especially now that you have the rich contextual InstaBrain already pulling in data and the systems of record from this fragmented internal systems. Can we take it to this function internally or that function externally even with us? So what we are seeing is that the overall value that we are delivering is not just defined by what outcomes we achieve. But we are measuring the outcomes that the customers achieve. Because I think then only we are a believable generational company. That we are not just creating value for us, but that we are able to sustain and create value that compounds over time for our customers. And so that's how we measure ourselves. That's what excites us, that it is massively accelerated and it's probably the first time in my life where I've seen that actually value accretion. When done in a very focused manner, in a very vertical specific manner. Can actually create sustained outcomes and we are seeing that in C2 very much so.

Pablo Srugo  (00:29:00) :
One of the things that I find founders want to know most about is go to market. Maybe and especially, enterprise go to market is kind of notoriously hard just for obvious reasons. What are some of the unique things or maybe the most important things that you've done in go to market? After these two customers to go and get, you know, the next five, the next ten?

Amit Shah (00:29:17) :
Yeah, I think both of our LinkedIn feeds is replete with all these claims about what really works and our mindset with GTM is a very engineering mindset. We think about what can we test at scale and learn, and continuously adopt. So even our GTM team is a GTM engineering team. We have engineers in that team and we treat it in a very agentic manner. So everything in our outbound is driven programmatically. We have self-constructed our GTM stack, which powers it. We think very deeply about face time. So we love events, we love going to a lot of these industry conferences this year.

Pablo Srugo  (00:29:58) :
I've noticed lately that events are just becoming more and more important. Is that something that you've seen as well?

Amit Shah (00:30:03) :
Especially when you think about we were going back and talking about specificity, and vertical expertise. You really need to be where the tribe is in a way. You need to go and gather alongside them, right? So this year we are on pace to have attended more than a hundred sort of trade shows. 

Pablo Srugo  (00:30:22) :
Wow. 

Amit Shah (00:30:23) :
As a team. So we are very, very focused on making sure that we are not just getting in front of the operators. But it also makes us more aware of what really is cognitively and in general operationally really a drag coefficient for them, right? So we are not going in and talking about some random thing. That our solution does versus being very specific that, hey, we know that with tariffs. This is the problem that you guys are facing and here's the sales team's issue right now. That we heard literally in a session around table that you guys spoke about it yesterday. So that has been working really well for us and the third thing is like, look, I do think that if you deliver compounding values. Your customers start being your salespeople. So we have been privileged to be called in now, not just because of our outbound motion or events. But actually customers referring us to other customers and saying, hey, we have seen great value working with Instalily.

Pablo Srugo  (00:31:22) :
I mean, that's one of the clearest signs of product market fit is when customers start referring.

Amit Shah (00:31:27) :
Yeah and we are continuously testing all of these strategies, and continuously evaluating it. But our mindset is to test and learn. We obsess over what sort of each trade show that we went to, what was the quality and the quantum of the interactions. Not just, you know, did we get a deal out of it or not. But the same ideas that we have that a system of learning will overcome all these systems of record over time. Internally also, we think that we are built and primed to learn collectively as a team. So I'm more thoughtful about, hey, at that trade show, did we get to the decision makers? And did we really capture what was the pain point of the decision maker? Did we sit through even the boring presentations? Because a lot of people will just jump to, hey, can I get in front of that VP? Well, the VP's pain points are actually reflective of the line management's pain points and you're better off going and talking to that frontline worker. And really asking, hey, how do you start your day if you are a leader of a sales team? And you will learn a lot more, and you'll be a lot more thoughtful when you really go in front of the decision maker. But that test and learn mindset is a foundational element of who we are.

Pablo Srugo  (00:32:43) :
The other thing I want to touch on a little bit is just like AI talent. I think there's so many ways to deliver value, because of what AI has opened up. But now the question is, who's actually going to get that pie? And a lot of that has to do with people, and the quality of AI that you create. How do you think about getting AI talent? How do you think about hiring engineers these days?

Amit Shah (00:33:02) :
It's very interesting, you know, I think the starting point is that the bottleneck isn't intelligence. It's actually execution. So when we think about hiring talent, we are trying to find the equilibrium between these two. But we are almost trying to focus on what's your aptitude and attitude in this new world. So our starting point was to say that, hey, if we are going to be surrounded in an AI first world, we ought to first go and start working alongside the AI natives. So the first thing that I and my co-founder did is really go to the top engineering colleges. And we are a US-based company, proudly so. We went in and started talking to a lot of these graduating seniors. Whether at undergraduate, master's, or PhD level and we built out an initial nucleus of people who had just graduated in '23. Which was very different from how a lot of other companies were starting out.

Pablo Srugo  (00:34:04) :
Yeah, usually you go for more senior talent, more expertise, these sort of things.

Amit Shah (00:34:07) :
Especially when attacking the enterprise surface area, right? But our feeling was that if you foundationally understood AI technologies as a nativist. You are able to build with a sheer confidence and you are able to execute with a sheer confidence as well. So we actually ended up, I remember the first year we probably did. We had about a thousand applications and we were trying to just hire four people. And I and my co-founder did close to two hundred interviews to really get to who really has this ability to ship fearlessly.

Pablo Srugo  (00:34:42) :
The other thing that I've seen lately a lot, is just insane growth when it comes to some of these AI companies. How have things been for you on the revenue side? How long did it take to hit a million ARR, for example?

Amit Shah (00:34:51) :
So I know the triple, triple, double, double is kind of discarded, but we are accelerating rapidly. We were able to hit that first million ARR within a matter of months and then we are in the second year of our sales cycle. And confidently past that triple, triple, and actually accelerating our growth. We have had a lot of inbound interest for our Series B and sort of the next funding cycle. Which we are about to sort of start thinking through as well. But it's been great and even more than that, Pablo. As I said, more than ARR, what I measured myself is, what impact have we created for our customers? So just for one customer, we are on track to deliver more than $150 million in annualized growth.

Pablo Srugo  (00:35:37) :
Wow.

Amit Shah (00:35:38) :
So these are very serious compounding engines. So my thinking is, not just sort of the book of business now. But as I said, you know, I'm in this game to build a generational company and I think the early signs that we see, and the momentum that we see makes us super confident that we are on that path.

Pablo Srugo  (00:35:56) :
And when did you feel you had found true product market fit?

Amit Shah (00:35:59) :
I would say all of these three values came together, you know, when we started seeing our customers refer us to other customers. When we started seeing the quality of the code that we were deploying, not just past master in a POC period. But way beyond that, to the point that now we are probably, I would say, one of the only companies globally that is not just transforming work using agents. But is building these systems of learning that are rewiring work, right? No other company that I've come across has the depth of deployment in this really complex industries. Where the code is actually working and augmenting the work. And then the third thing I felt was when our culture was attracting the best talent in the game. That's when I knew, I would say probably six to eight months in our journey. You know, I and my co-founder looked at each other, and by the way, we were profitable before raising Series A. And, you know, we were very excited by Chrissy and the Insight team when they came to the table. We had an overwhelming interest from both East Coast and West Coast VCs. And we were super excited that, hey, it seems like we are on the right track. This is the right time to start adding some jet fuel to our journey and finding the right partner. So that's why we actually decided to even take on a Series A.

Pablo Srugo  (00:37:26) :
Was there ever a time where actually, on the flip side, you thought things might not work out?

Amit Shah (00:37:30) :
You know, very early on, because I ran a company called 1-800-Flowers. When I would pitch sort of our vision of agents. People would be like, are you really talking about phone agents? You know, are we going to automate voice? But leaving that aside, I think the early days. I'll give you an example of it just to tell you where we had some questions for us. When we first started deploying this multi-agent, multi-step architecture. We would get caught up in something called a gratitude loop and the gratitude loop was that the agents would just pat each other on the back, and say, great job. But they would not go to the next step, because the reward function was that they would get plus one if they just complimented each other, right? And so those early days, it was the Wild West. Because you were literally experimenting with code and hoping that it would work. And those days, I would say both from a customer appreciation perspective, and frankly, where the systems were. Were kind of challenging days, but because I think we focused on sort of a very vertical mindset. Very deep understanding of, hey, what problem we want to attack. We didn't really feel that there were deep drag coefficients that were external to us. We always felt it was about how do we keep doubling down on the focus and also believing. I've been around this game a long time to know that evolution occurs in sort of a slow pace, but when the stack change occurs it surprises all of us. So the better focus is not to worry about the rate of change. The greater focus should be on how am I preparing for the step change and that's why we decided to double down on our three C's. And our value system more than what underlying technology will evolve out to be.

Pablo Srugo  (00:39:23) :
And last question, what will be your number one piece of advice for an early stage founder?

Amit Shah (00:39:28) :
I would say my number one advice would be pick your lane and double down on your values. Because I think we are still on our zero of day zero of this game and I think there is a lot of evolution left to happen. And a lot of shifting forces. So unless and until you are deep down. Deeply aligned on what lane you want to play in and what values you want to manifest as you play the game. You're likely going to come up short as with everything else in life.

Pablo Srugo  (00:40:02) :
Amit, thanks so much for jumping on the show, man. It's been great having you.

Amit Shah (00:40:05) :
This was such a fun and a real pleasure to talk about Instalily, and our vision. Thank you for the opportunity, Pablo.

Pablo Srugo  (00:40:13) :
Wow, what an episode. You're probably in awe. You're in absolute shock. You're like, that helped me so much. So guess what? Now it's your turn to help someone else. Share the episode in the WhatsApp group you have with founders. Share it on that Slack channel. Send it to your founder friends and help them out. Trust me, they will love you for it.