About our guest

Thothathri Srinivasan (who goes by “Tho”) is the Head of AI and Engineering at Flip, one of Indonesia’s leading fintech companies. With over a decade of experience in Silicon Valley, Tho has built products and systems at some of the world’s most recognizable tech companies including Groupon, Netflix, and Pinterest.

Originally from Bangalore, India, Tho began his career in electronics and software engineering before moving to the United States in 2010 to pursue his master’s degree in computer science. His Silicon Valley journey started at Groupon in 2011, where he worked on merchant-facing products and learned the fundamentals of dynamic pricing and A/B testing. He then spent six years at Netflix Studio, building the infrastructure that powers content rights management, deals, and contracts – essentially everything from “pitch to play” in Netflix’s content ecosystem.

After a stint at Pinterest working on shopping monetization, Tho briefly explored the AI startup space before relocating to Southeast Asia with his family. His transition to the region led him to Flip, where he now leads AI and engineering for Indonesia’s growing fintech market. 

Tho is also an alumnus of Insignia Ventures Academy’s Certificate in Venture Capital program and became involved in angel investing through the XA Network, bringing his technical expertise to early-stage startup evaluation and mentorship.

Timestamps

Watch Part 1:

(00:00) Kickstarting AI Transformation in Flip;

(03:28) Measuring ROI on Flip’s AI Initiatives;

(07:04) Learnings from Tho’s Make or Break Moments in Tech;

(10:21) Leading Engineering teams in Southeast Asia vs Silicon Valley;

Follow us on LinkedIn for more updates

Subscribe to our monthly newsletter for all the news and resources

Directed by Paulo Joquiño

Produced by Paulo Joquiño

The content of this podcast is for informational purposes only, should not be taken as legal, tax, or business advice or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any ⁠⁠⁠⁠⁠⁠Insignia Ventures⁠⁠⁠⁠⁠⁠ fund. Any and all opinions shared in this episode are solely personal thoughts and reflections of the guest and the host.

Transcript

AI Implementation at Flip

Paulo: Now we’re gonna get into what you do at Flip. I think especially with AI, we’ve had a conversation previously with Carl, COO. We talked about AI transformation and I’m interested to hear about it. One from an engineer and two from a more Indonesian company. The whole role of AI, I think your scope is pretty interesting. They could have just gotten you to lead engineering, but then they also added an AI aspect into it. What was that conversation like with Rish and Rab? I guess that was already part of the conversation from the beginning.

Tho: When I came in, I was just head of engineering. Primarily I was heading to the B2B side. To give folks context, Flip has three divisions. One is the consumer side, which is basically your wallet, like you have money that you can use to send it to friends. What the taxi driver showed you. Make payments, you can top up your electricity bill, wallet, all that.

The second one is the B2B side, which I had, and primarily two things. One is disbursement that we help out. You can think of large groups, 10,000, 20,000 employees making their salary payouts, for example, using our disbursement tool.

Then the second one is a way to accept payment. We support all of the different e-wallet bank integrations and QR integrations as well. That’s the other part.

Then the third part of our business is the lending arm, which is Lippy.

I had the B2B side, and I’d been doing this for a while. Obviously, given my background in AI, I remember going to Luman and saying, “Hey, Luman, I want to also do things on the AI side along with the B2B side.” That’s how that conversation started. Prish and Luman said, “Sounds good. Go for it. Tell us what you need and whatever you’re trying to do.” That led to where we are now.

Three Pillars of AI Strategy

Tho: The first thing that we did in terms of AI was we said let’s focus just on the tooling side. We started talking to Cursor and trying to have our engineering teams use Cursor pretty significantly. Obviously we saw some improvements, but the bigger parts that we saw were on the test engineers. The test engineers are using it. Difficult test coverage was, let’s say 80%. You’re trying to expand that to 100%. The delta seems small, but takes a very long time to get there. They said, can we use AI there to close that gap down?

That’s one of the big things that we did. Of course now most of our, actually all of our engineering team uses Cursor. That’s purely on the tooling side.

The second side, which is a very obvious use case for AI, was on the customer support side. That’s another thing that we started exploring saying okay, is there something that we can do there to make all of our customer requests a little bit faster and more efficient?

The third part obviously comes to what we do as a FinTech company. Whether it’s in terms of risk management, all of that’s what we want to continue. Especially on the lending side. I think those are the three overall things that we are doing.

Paulo: Customer support of course are very obvious first use cases. But as you expand into the third use case a little bit more, that would be interesting.

Tho: I also want to add, as part of the third use case is what we’re trying to do currently with the data. Flip obviously has a lot of data. You have folks from ops, sales, business, engineering, product, basically trying to say “Hey tell me how much money we made this month. What are the top merchants? How much revenue did we make, what are the margins?” All of these different things.

Typically they would come to the data team and say “Hey, data team, give us all of these numbers. Crunch it for us.” There’s always variations of that because different account reps handle different accounts. They just want to know some specific attributes. They’re not obviously well versed in writing the queries themselves. Now we’re building an AI bot that will enable this for them, so they can ask in plain English. It’s a kind of internal business intelligence.

Paulo: To obtain any information that’s needed.

Tho: Exactly. I think that has been a pretty big win. We’re building on that as we go forward.

ROI and Measurement Framework

Paulo: I’m curious to know how you have framed ROI for these types of projects? Obviously recently featured Flip a few weeks back, I think Ari did an interview on Tech in Asia or something like that, like 5x growth for the B2B business. I’m not sure how much of that was AI driven or just your team’s hard work, but how did you pitch some of these ideas from an ROI standpoint and how do you then measure whether it worked or not?

Tho: I think specifically for the tooling side, what we did was we said, let’s start as a small pilot. Before we expand this to all of engineering, let’s look at the value added specifically for the B2B engineering side. We saw a pretty big lift in terms of productivity. Even if you’re saving 20% of an engineer’s time, that’s a lot of value compared to what you’re paying for the tools. We said okay, this is working pretty well and I really like what it was doing from a testing standpoint. That’s how we gradually expanded that out.

Everything was purely data driven. Everything has to be data-driven. Once we saw that value add, I went and pitched it to him and he said, “It’s a no brainer. Let’s go ahead and do it.” That’s from a tooling standpoint.

Now for the other things, for example, what we’re trying to do from a data standpoint, that is an interesting one because what we want to measure is how many people have stopped going to the data team to ask these direct requests. That was causing one or two folks from the data team to be dedicated to this. These are all ad hoc requests, you don’t know when it’s gonna come. They need to keep their sprint cycles for it and so on. That’s the second part. I think that also is something that we’re actually tracking right now.

Paulo: It was time cost savings, productivity gains.

Risk Management and Future Applications

Paulo: Have you already thought about the third one, the third kind of use case for Flip as a FinTech. How would ROI look for that?

Tho: From a risk management perspective, obviously we have rules and engines and everything in place, but moving away from the traditional sort of rule-based logic or existing rule engine that we have into using some interesting data points that we already have inside Flip and using AI to make that process faster, quicker. That’s something else that we are investing in as we speak.

Paulo: You guys have 13, is it 30 million already or somewhere about that, 10 to 30 million users, so there’s obviously a lot of data per user based on how they use the app and all those things. Obviously when it comes to risk management, there’s maybe some variables that would cost much less to obtain, but then you guys could actually use just by virtue of automating a lot of how the data is obtained. Quite interesting.

Leadership Advice on AI Implementation

Paulo: What advice would you have for leaders listening in who are also on this journey of trying to figure out how to better automate or leverage AI tools?

Tho: I would say start with a smaller set because obviously you hear things about AI being 100x productive or any of that. I still feel like we are in an in-between stage where we are using AI to enhance our day to day and make it a little bit more efficient. Since that’s the stage we’re in, look at it from a smaller perspective instead of trying to implement it from a broader perspective. Do it on a small team, see how it works out, see what’s working well, what’s not. Try to constantly get feedback.

I noticed some basic things. Even Cursor, for example. Some of the models are really good. If you use Sonnet in Cursor, it’s great. If you use some of the other models, it’s not so good internally. You need to tweak it around and become very good at prompt engineering, basically. You want to get to that good stage.

With a smaller team, it’s easy to control. You can see a lot of analytics and stats about it and then expand out from that. I think the tooling is definitely a no brainer because it definitely adds value in terms of 10 plus percentage. At that point, it makes a lot of sense.

Career Lessons: Make or Break Moments

Paulo: Something I always ask our guests is like a make or break moment in their career or journey that they think would be really helpful for our audience to learn from. What kind of moments do you share with us?

Tho: I think one of the things that I’ve learned both being in big tech as well as startups, obviously the biggest difference is adaptability here in startups. Adaptability in startups basically means, obviously you have to do what is right in the next one week quickly, as opposed to you having an overall North Star, which is still applicable for startups, but every week you’re building things to please the customer and finding out what the customer’s problems are during that time.

You’re quickly trying to solve that. Of course you still have a vision for six months, one year versus obviously in big tech companies where you have broader timelines, and you’re still trying to solve problems, you’re trying to do it to the scale that is actually needed.

I think make or break, I would say is always okay stepping out of your comfort zone. Whether it’s moving from big companies to startups or vice versa. Trying to understand what exactly is the value add that you’re doing.

One other thing that I want to call out here, which I think recently I learned, is that as you step into any company, everyone knows, especially if it’s not your own startup, you’re working for someone. Always try to understand what is the value add that you are bringing to the table. How much are you getting paid and how can this company make even more money out of it?

This goes back to Netflix where we would think of ourselves as we are a sports team. You gotta figure out who’s the best and who do you have to cut. That’s the harsh reality of life. But I think that is something that I’ve started accepting a little bit more. I’ve said a great example here is Lionel Messi moving teams. Why does Miami have to pay so much for them? You gotta be a brand where the brand pays for itself. So if you come into a company and say “Hey, this is how much you’re paying me, but I’m going to save you guys the same amount of money. So basically I’m working for free.”

I think that is a very interesting mindset that I’m thinking of more now. How do I save costs and build amazing products?

Leadership Philosophy and Team Management

Paulo: I’m interested to know how that mindset actually influences your leadership style, especially being a, I guess for lack of a better term, like a middle manager of sorts. Between the engineers you work with and also trying to align with the management direction. How is that mindset that you just described influence that?

Tho: I think for everything that we invest in, we start looking at, okay, what is the ROI here? Purely from either traction that we’re getting or amount of time that we are saving or anything else. That’s one part of it. Second thing is, this again goes back to Netflix, but thinking about the keeper test. Who do you want to keep on your team, who’s an absolute stellar performer? Who comes in and says “Hey, I’m gonna leave tomorrow” and you’re gonna feel bad about it. You tell them “What can I do to keep you?” because you’re an amazing asset to the company. I think both of those things tie into how you handle teams and how you build teams over time.

Obviously, every team has a range of people on that team. But you’re working together on a very common goal. Exciting the team and making them aware of what is the value add that they’re having. I think that also is a huge part of it. You’re thinking of it from an individual perspective as well as from a team perspective. What is the value add that we’re bringing together as a team? What is the overall revenue that we are generating as a team? Are we focused on the right ideas?

Cultural Differences: Silicon Valley vs Southeast Asia

Paulo: That actually brings up something that we talked about earlier, which I wanted to also mention on the podcast, which is how has it been working here in Southeast Asia versus back in the Valley? Especially from a culture and engineering standpoint.

Tho: I think it’s very different. Silicon Valley, I think again, given that the engineering teams and so on have existed for that long. I think you give folks a task, and again, this is going back to Netflix, but you give someone a task, the expectation is it gets done.

Here maybe there’s a little bit more handholding that you need to do in Southeast Asia, but I think talent wise, there are very good talent pools in Indonesia specifically, where we work. You just have to find the right talent and bring them together and drive that mission forward. I don’t think there’s any lack of talent, it’s just you need to look at the right places.

Paulo: What advice would you have for founders who are saying that, oh it’s too hard to find talent here. There’s too much competition for the best talent?

Tho: I think you need to go to, obviously Jakarta is a big city, but you need to go to the smaller ones like Bandung, ITB, and some of these other universities producing solid engineers. It really is blowing my mind. Obviously even in the context of AI. Especially in the context of AI. Because I think the engineers coming out now are able to use some of these tools better than engineers probably like me, who’ve been around for 15, 20 years. I think they’re able to use that and leverage that faster so they’re able to produce code faster. The ultimate goal is to produce bug-free code and build great products. Then how does it matter if you use AI to get you that? They’re able to use those tools much better.

Angel Investing Focus

Paulo: On that note, I would like to thank Tho for coming on the show. We’ve had quite a discussion about a whole range of topics including his own kind of mini biography. If you ever come to that point, I would love to write your story. But in any case at this point in time, I am really happy to have seen your progress through the IVA program. Great to know that you’re working with our portfolio company at Flip and as you’re immersing yourself into Southeast Asia.

If you guys are, I guess since you’re angel investing, like any particular companies, AI startups that you’re talking to? Or what specific sectors or use cases are you looking at?

Tho: A few of them, but I think they’re all much bigger. But definitely looking primarily at the AI vertical. If you’re building something in AI, if there’s anything that you guys are building that you think I should check out, please feel free to ping me and I’ll definitely take a look.

Paulo: Let’s connect people with Tho. Thanks again for coming on the show.

Tho: Thank you so much for having me.