Interview of Praful Krishna from Coseer (Video)
Praful Krishna shares his story of how he got the idea of Coseer and how he grew it without taking any VC funding. He also mentions that marketing channels that worked and the ones that didn't. If you are an aspiring entrepreneur, make sure to hear his advice at the end.
Neerav: Hey, this is Neerav from Growfers and today we have Praful Krishna from Coseer. Praful would you like to tell me something about yourself and company?
Praful: Absolutely, thanks for having me over Neerev sounds like a great idea that you're embarking on good luck with it. So, I am Praful I run this company called Coseer what we do is use data science and digital technologies particularly if it’s around national language to automate or accelerate workflows in fortune five hundred. If you have any workflow or any process, that is centered around any processing any kind of language. We can in most cases automate it or accelerate it tenfold so that your employees and customers get to the right information quicker than they are doing now.
Neerav: You are trying to embark on this NLP / AI journey which is very much in Vogue right now.
Praful: We’ve been at it for about 8 years now. We are kind of on the mainstreams so to say.
Neerav: How did you come up with the idea behind Coseer? What were your background before and how did you come up with this idea?
Praful: This idea really came from when I was with hedge fund my job was to invest into companies and look after the investments find the right time to accept. That is where I found that I was spending far more time on looking for more information in thinking about it. I would be. You know in the world of finances so many relevant information floating around and so much irrelevant information floating around. It would take more time to find the right negative insight that I would be after. You know I would make an investment decision. That’s where the idea started, I was like “In this day and age why can’t a computer do this to me even if it does it at an eighty percent level. And you know the rest of it the human can complete. But that itself would be a huge help that really is the genesis of Coseer and from there the journey began.
Neerav: What happened is that you started in finance and then you caught this idea of Coseer in finance. When you first created your product or when you thought about the product, where you thinking of applying this to financial domain only or thinking of a much more horizontal plain from the start?
Praful: I would say it has gone through many influences in its journey, When I started, I started with automating my own job like what I was doing. I started saying “Can I make a tool that investment analyst like me can look at.” Essentially you know whatever ten or twelve hours they work on every kind of market. They are kind of gone they already have it there. That’s where it started, I think once we had the product, we had the customer feedback, we developed it and once the technology starts to become stable, we realized that you know that this is just a powerful start that we could actually make it horizontal and sell it to enterprises as well. instead of selling it to small firms. Just the unit economics was looking for attractive there. We made that decision and that’s how we kind of move from finance to enterprise. Once we got into enterprise, finance continued as a kind of a focus sector but pharma quickly took over, and today infact pharma is our biggest focus sector.
Neerav: Your biggest focus sector switched from finance to pharma. (Praful Interrupts.)
Praful: It happened right away; it was not our conscious decision to do that.
Neerav: Did you have any people in pharma you knew? How did you break into pharma?
Praful: You know for a small company serendipity is as a big a factor as anything else and it’s essentially getting our first customer in pharma through my personal network. Because the quality of the solution is so good that the word of mouth took over and we also invested a bit behind it. We realized that pharma has this classic problem were the life of the knowledge (inaudible) the knowledge around it hands down for decades. Whereas the person working on that knowledge is working on that for two to four years depending on different companies. There is a huge mismatch we realized that our solution could be very fruitful there and then we started investing behind it.
Neerav: Got it. That’s very interesting you got your idea and started this company, getting your own edge and somehow found by your own connection and serendipity found an audience in pharma industry which had been way bigger than what it had earlier in the finance industry.
Praful: You’re right and they have the willingness to pay were they were also more open compared to finance for using external vendors versus trying to build it themselves. In finance you’ll see and technology also there is an urge to you know built it yourself and have it in house and that’s a good philosophy people think of it that way. But from a vendor’s perspective like us a pharma is much more amenable to talk to some vendors.
Neerav: You’ve been at it … in this company for 8 years. You’ve been bootstrap that’s a pretty big thing that it’s very difficult to bootstrap a product company stuck with it for eight years and grew it. Did you ever think about taking V.C. money, or were you just focused in making this profitably?
Praful: Both,We have been focused on making it profitably. Covid has been hard ‘till 2019 we have had a good growth our way. Venture capital is suited for a certain type of company. And Coseer in question is not that. Venture capitalism is very, very effective when you have a very defined product. What you need the money for is for scaling the market or going to market activities around that product. We could never get to a point where productize it and copy and paste it in any environment and it will start working obviously there’s reasons for that. We could never get to that point where we could have taken “X” million dollars from venture capital and then invest in sales and gotten a positive outcome out of that. So, we’ve never taken venture capital for that reason.
Neerav: Every installation or every customer deployment is likely different.
Praful: Slightly different in two of three ways, right? The basic problem or the basic opportunity whichever you look at it is that A.I. has to be trained again and again. Right now, for example in the news this whole emphasis on GPT, where GPT essentially stands for Generally Pre-Trained people are trying to have pretrained models that can apply anywhere. If you look at it if you kind of peel the cover, you’ll the accuracy of these models are very low when applied to or in doing very specific or company specific problems you have to do what’s called fine tuning. That’s fine Tuning itself is a massive exercise, they call it fine tuning but there is nothing fine about it, you have to go through like millions of dollars. Even our A.I we base our solution to on something called calibrated quantum mesh which is something we develop ourselves. Obviously, there is a level of pre-training that we do but it has to be pretrained in every situation. If you have been doing that the next part is the plumbing around it. First of all, it’s data byte so everything relies on data so data coming in and data going out. If you are doing these two things anyways. At that time, we realized that if we do the third part which is the solutioning part on top of it. Somebody needs tables, and somebody needs the question and answering phase, and somebody needs the prepopulated data basis, and the engine behind it is exactly the same which is A.I. engine and the data plumbing has to be different for every customer so we can just chain the solutioning aspect on top of it. And widen much more value to our customers than to trying to force fit a product. That’s how we started following a solution strategy but as I’ve said before in the solution strategy V.C. money is not the most positive thing.
Neerav: Got it. Awesome. After launch when you found customers in pharma to your network and everything. After that first set of initial set of customers, what worked to attract and retain more customers and what did not work?
Praful: We’ve tried everything. There was a point in time that where an ex V.P. of sales and S.E.P. was working for us and (inaudible) selling, we tried the insider call selling and everything. What eventually worked was taught leadership inbound email marketing. I or the people from our company would be going to each and every conference we can find and talk about the value of a solution like this, talk about what we have done with others. Eli Lilly for example is a big company for us. They were so happy with us our work that they actually publish or co-author an academy papers based on our work. That combined with word of mouth, you know somebody picks up the phone and say “Hey talk to Praful, these guys are on to something, great something great or to Praful in general.” That has worked more for us so far and that had been the source of our growth so far.
Neerav: What are the things you did try which did not work? Probably a long list but that can you think of?
Praful: We have tried account-based marketing, we have tried inside sales, account-based marketing is that you create massive automated systems algorithms will figure you out and keep your messaging. Obviously, there is still human’s involvement as well. But keeps emailing you until you say "Yes.". We've done some of those things. We've tried insider sales where somebody acutely picks ups the phone and starts calling people and gets hang up on you know almost all the time, and then we tried building really consultative sales somebody had to level up SVP of SAP, It was kind of working for us trying to open doors and converting to sales kind of very heavy with or making kind of a sales model. And we also tried the regular type or kind of sales channel in terms of have a normal sales people knocking on the doors.
Neerav: Got it.
Praful: We tried all the data gamut. The other thing that did work, by the way is partnerships with either consulting firms of system integrators caught us a lot of business as well.
Neerav: Nice. Good to know that partnerships.
Praful: Partnership can come is multiple forms as well, Right? You can have a partnership like each of these sales model. Again, the partnership also that kind of work for us was thought leader kind of partnership. We will go in behalf of McKinsey in go enters into the company while McKinsey would be saying "Look Mr. Client we are bringing some cool company to talk to you" And then we show up there is already a huge amount of credibility already put there in that conversation. Similarly, we are working with bane in that regard and then Deloitte and Accenture and Wipro, right. This kind of the entire gamut, where we would come in as a very kind of a very credible and warm introduction to the client.
Neerav: And a lot of our audience probably will not have an idea on how to actually go about creating these partnerships, right. So, for example, if you want to partner up the Bane or McKinsey, how did you even go about it? Because these are big organizations? How did you as a start-up, how did you partner up with them?
Praful: Yeah, I always say that, you know, if you're doing a b2b startup make sure before you launch, make sure you identify five of your best friends who can give you either business or don't Roadsters business. Right? I think as a founder, when you're starting unfortunate as it is a lot of serendipity, but a lot of kind of personal networking, begging, pleading, and you know, making people from whatever it takes. Stepping back, I mean, you have to be creative. You have to open the doors. These are huge organizations, they don't have any systems you know, when I was making so I was at McKinsey, right. that kind of helped me a bit. But when I was at McKinsey there was a kind of a joke, we had that you know there are rules for everything. And most things happen outside of rules, right. that is true for almost every big organization that is true for almost every big organization technology is a second thing, right? You have to be creative; you have to be resourceful. You have to network your way in. You'll find that if you have something cool then people are basically helpful people you just kind of go and knock on the right doors and people will open the door for you people will invite you in. And that's how you just grow.
Neerav: Okay, that's a great advice, thank you.
Praful: That's my experience.
Neerav: In this whole journey, what have you learned? And what advice, do you have other entrepreneurs who are just starting out or thinking about starting their own business?
Praful: Yes, I think a couple of learnings, right? As I said, first is figure out how would you get the initial momentum if you're doing a b2b? (inaudible) five best friends that will give your business on day one if you're a B to C, figure out how, what are the milestones. We'll have to hit to raise about 15 to $20 million because we're a b2c business. That's the kind of money you need to establish a brand. Right? And that at least one plan for that should be clear. And obviously, the planning will go to hell as soon as things start, and that's fine. but at least one plan should be clear. Don't just start with the technological edge. Right. And you know everyone says that to be honest and entrepreneur we'll have to go through some of these mistakes you understand what I'm saying? but that has been an important one for me. The other thing I would say is that right now there's so much hype around A.I. We don't even try to sell AI anymore, right? I mean, we don't we don't talk about you. I mean, we talk about AI in our written kind of material that you will find on website. But when I'm talking to a person and I'm trying to understand the problem and see if we can solve it, we don't even mention the word A.I. And the reason is that AI comes with certain expectations and certain kind of fears. So as soon as you as you drop AI your audience will assume one of two things. One is that Oh God, another snake oil, snake oil salesperson, right? Or they will assume that you have some wonderful tool that can solve every problem for them. And it's best to manage that expectation early on. The best relationships you make are those where you say upfront that look you know this is what you want but this is what you actually need and it's different. And by the way of what you need. This can be done by AI or technology whatever you call it and this is just vaporware. And we can try but you know, it's not going to work I think we have to be careful a bit about AI and managing expectations. So I will say these two things have been my biggest kind of learning in terms of go to market strategies.
Neerav: And the last question, where can the audience learn more about Coseer?
Praful: Oh, everywhere we try to be as visible as possible. So the website coseer.com is a great place to start you find blogs and webinars use cases, what have you I was recently was featured on Udacity course where we talked about AI Central so you can look at that video. There's much of other videos floating around. I mean, just type Coseer or my name and on Google this bunch of stuff will just pop at you.
Neerav: Awesome. Praful it was nice talking to you. And thanks a lot for sharing your journey and advice for future entrepreneurs. And I wish you good luck for the future.
Praful: You too. Thanks for having me. As I said its great initiative that you're on. Wish you all the best.
Neerav: Thank you.