Inside Mercury's Founding Story | Co-Founder & CEO Immad Akhund

We really had product market fit from the first day we launched. I really had no expectations of this happening. Within actually 4 days of our launch, someone sent like a million dollars to a Mercury account, and I'd never spoken to that person.

When I started in 2017, "new banking" was hardly even a phrase. For the first year, I was really worried about, like, over-investing too early, and so we were really, like, always behind. Like, we'd never had enough customer support.

Like, everything was breaking all the time for the first year. And then eventually COVID hit, and at that point, you know, the whole world was ending. But then a month after that, our growth grew even faster because all the bank branches were shut down, and there was this huge explosion in, like, digital businesses and e-commerce.

I think the hard thing about Mercury is trust. That's taken years to build. When I think of a new user, I want to try to reduce my marginal cost of serving them to zero, which means I can give them better features for new things, like net new things.

Really, it's about getting feedback and improving the feature. Get a small team prove it out, and once it's, like, proved out enough, then we invest it. I feel like every 6 to 12 months, like, the paradigm shifts entirely.

We actually have an AI enablement team building a lot of kind of platform tools, so AI was still quite early with financial services. Now we're working on a ton of customer-facing features that, like, really make sense in a way that it didn't make sense last year.

So, you know, one big thing we're pushing is you kind of have to, like, be a little contrarian. Unless you have, like, a very different perspective on, like, how everyone else is doing it, it'll be very hard. But I do feel like the thing that matters is—

Immad, welcome to NEW ECONOMIES. I've been so excited to have you on the show for so long, and we finally managed to make it happen, so here we are.

Yeah, thanks for having me. I guess the audience will get two British accents, although mine's a bit diluted.

Exactly. Hopefully they stay around for the whole episode, because we have some pretty exciting news. Mercury is announcing some interesting new developments. What's been happening? It sounds like you guys have been pretty busy, just a little bit, in the last few months.

Yeah, I guess startup world is always busy, but we're announcing our Series D. We just raised $200M at a $5.2B valuation, with TCV leading, and a lot of existing investors like Andreessen Horowitz, Sequoia, and Cotu Sparke participating in the round.

So pretty exciting news. We did a round last year, so about a year ago we did a round with Sequoia leading, but we made a lot of progress last year. Grew, you know, more than 50% on most of our metrics, and we had a lot of interest, so we did another round.

Amazing, congrats. You know, it feels like—well, actually, I'd love to just, like, get inside for a moment with that fundraise. Why do you think was now a good time to fundraise again, and how long did it take you?

I'm a big fan of fundraising when you don't need the money. At least in my previous companies, I've always been in this, like, "Oh my God, we're going to run out," kind of thing. With Mercury, we've been profitable for 4 years, so it's very much, yeah, "Hey, we made a lot of progress from last year, there's a lot of inbound interest," and we wanted to capitalize on that.

You know, we've grown really fast, partly because of AI kind of creating this kind of new wave of entrepreneurs. Especially in Q1 this year, our growth was 2.5x higher than Q1 last year, which was already—in terms of number of applications.

And, you know, Q1 last year was already a pretty big quarter for us, so that's been great to see. And then we also applied for a bank charter in December, and we just got kind of conditional OCC approval for that.

So yeah, those are, like, two places where we're seeing a lot of momentum, and I thought it was, like, a good time to kind of, yeah, raise a little bit of money, get a higher valuation. Obviously has some, you know, marketing, hiring, and acquisition benefit.

It feels like fintech is having a bit of a moment. I was looking at all the fintech players who, similar to Mercury, who actually launched at roughly the same time as you. You know, you've got the likes of you guys, Revolut are making some interesting developments with launching in the US.

You've got Robinhood to launch a new, like, venture product to get access to, you know, interesting AI companies. Brex obviously just got acquired, Ramp are doing some interesting things. Like, it's all kind of just happening at once. Why is fintech having a moment now, do you think?

You know, it's funny. It's like the seeds you plant, like, 8 to 10 years ago are going to take a while to bear fruit,right? I think in 2021, about, yeah, 2021 was a crazy year, but fintech was, like, kind of part of the biggest part of the craziness.

I think 20 to 25% of all funding went to fintech. So I think, yeah, a lot of those companies ended up not doing well. But the ones that did, you know, we're seeing them now doing really, really well because, you know, these markets are huge,right?

Like, financial services is, like, you know, $2 trillion in the US or something like that. So, you know, every time I do a fundraising slide, I'm always like, "$2 trillion!" You know, it's like the best kind of, or one of the best, like, TAMs market sizes you can get.

So now that, like, you know, we've been doing this for a while, us, Revolut, new bank, whatever, you know, we've reached a certain level of maturity and, I guess, like, you know, now it seems a lot more obvious.

When I started in 2017,right, new banking was hardly, hardly even a phrase, and, like, a lot of it was, like, just convincing people that, like, anyone would use a startup for banking or that we even needed this,right? Like, the questions were much more, like, basic, whereas now it's much more, like, it feels inevitable that, like, people will get most of their financial experiences using apps.

And, like, I don't know, it seems obvious to me that that's, you know, that fintechs and kind of product-centric companies are the ones that are going to, like, be delivering these experiences in the future.

Every time you update those pitch decks, it feels like the whole market has changed as well. I mean, just in the last 12 months, this whole AI category has changed completely. How has AI changed in the last 12 months versus before you did your previous fundraise?

I mean, AI is changing every month. I think there was a huge inflection point kind of end of November when Opus 5.4.5 came out. You know, that's the model from Anthropic that, like, really made coding work, like, extremely well, and, you know, ended up, like, leading to cowork and, like, Claude Code and, like, a lot of the—I feel like there's been a real inflection point.

And, like, you can see it in the Anthropic numbers,right? So I think a year ago, yeah, we said some things about AI, but AI was still quite early with financial services. Now we're working on a ton of customer-facing features that, like, really make sense in a way that it didn't make sense last year.

So, you know, one big thing we're pushing is this kind of ability for you to use your AI, whether it's ChatGPT or Claude Code or Cowork or whatever, to connect into Mercury. So we have a Mercury MCP and a Mercury CLI, and we're seeing, like, the numbers there go through the roof.

And, you know, our thinking is, like, you're spending all your time—not everyone, but some people are spending all their time using Claude Cowork, and that's, like, their central kind of brain. We should just

expose our features within that. And there's some really cool use cases, like, our accountant controller internally is using AI to see if there's any task for her to do on Mercury to approve, like, payments and things like that.

And there's tons of people using it to, like, run financial planning and all that kind of stuff. So that's one big use case for us. The second big thing is kind of having AI features within Mercury. We launched something, I guess, a month and a half ago called Mercury Insight, where you can ask questions about, you know, what happened last month with your finances.

And if you remember a year ago, everyone was like, "Oh, you can't touch numbers with AI because, like, AI would just make up numbers." But now you can make much more kind of deterministic things. So the, you know, in our Insights product, like, AI will make a statement about numbers.

Like, it'll say, like, "Hey, you spent $50K on AWS," but it'll actually, like, every time it mentions a number, it'll, you know, it normally makes it so you can click on it to see, like, what are the actual transactions added up to it.

But it also, in the background, it does the calculation. Like, we make sure it's, like, doing the actual math rather than, like, trying to guess its way to the number, which is what AI used to do. And then lastly, we're launching another really big feature called Mercury Command.

I guess in the summer or so. And the idea there is actually to make it so you can run all of Mercury and all of its workflowsright within the AI if you want. And it's not just a chat interface.

Like, you can say, like, "Hey, set up a virtual card," and it'll, like, kind of prompt you with a UI component so you can, you know, still see what's going to happen, and you can click yes. Like, that looks good.

So, yeah, I mean, I think AI has just got so much better that, like, you can actually build these, like, really compelling features within Mercury.

You're soright. You know, I've had a few ambassadors and founders on the show, and one of the biggest points we were talking about is, "Oh, no, no, AI. I don't want to, you know, have anything to do with my financial services."

But actually, now in reality, it's actually, like, starting to happen,right?

Yeah. But there's aright way to do it,right? There's also, like, a wrong way to do it,right? Like, I think, you know, one thing we've done is we have an AI customer support bot, but, like, you know, we have a human in the loop.

So you can really easily, you know, if the AI has an answer, it's instantaneous, that's a better customer experience. If it doesn't, it either automatically escalates to the human or it, like, you can easily escalate to a human.

So I think you have to be thoughtful about, like, you know, AI has a place, but also it can be annoying when it, like, you know, thinks you're talking about the wrong thing and there's no way to get out of, like, its box.

And I think if you start with that kind of idea, as AI gets more powerful, you know, it can answer more questions, it can do more things for you,right, within the same interface.

It has a place, but it also has a moat for a lot of these companies. But the moat only becomes so strong until everyone just copies each other,right? How defensible do you think the AI features are until you have to just keep evolving and evolving?

I find this whole defensibility thing kind of funny because, you know, when I first became an entrepreneur, which was like 2006, it was all about, like, "What's your moat? What's your moat?" Right? And it was like, you know, you needed a network effect or you needed data or you needed something to be your moat.

And then we kind of went through this phase where, like, you know, a lot of it was B2B SaaS. And I don't know, there was no one could really say most of the time what the moat was in most of the startups.

I think we actually, like, kind of swung back a little bit where, you know, software has in theory become a lot easier to make.

There's a lot of AI startups doing a lot of things. And I think you get back to the world where you need a real moat. And to me, that's like, you know, are there features that are better if, you know, more customers are using it?

Like these kind of classic network effect features. Are there data network effects? Is there a reason your brand is better? You know, how we think about it is, you know, number one, we're starting with banking, which is, like, a freaking hard thing to build,right?

There's, like, so much more than the software piece of it. Like, Mercury, you know, I used to say, like, to build something like Mercury is at least three times harder than to build just a software thing. Given that software has actually got easier, it's probably, like, 10 times harder now to build something like Mercury.

You need, like, all of this kind of compliance regulatory infrastructure. You need all of these back-end integrations, either, you know, either with the sponsor bank or obviously we're going and getting our own bank charter. So in some ways, like, doing the hard thing first and obviously, like, doing that at scale is by itself a moat.

And then over time, you also, you know, we're a very, I think the hard thing about Mercury is trust,right? Like, you're trusting Mercury to hold all your money. Like, it's actually, like, a pretty big deal. And, you know, when we first launched, within actually four days of our launch, someone sent, like, a million dollars to a Mercury account.

And I'd never spoken to that person, and I thought, I really had no expectations of this happening. I thought it would be a real, you know, grind and in-person kind of relationship to get someone to, like, really trust us that much.

But even then, no one would have sent us, like, $10 million when we first launched. You know, now, obviously, we're 7 or 8 years in, and, you know, we have several customers with more than $400 million on the Mercury account.

So, but that, like, trust builds over time, and that's part of, like, building a brand, building trust with customers and doing theright thing by customers and building all the products and features that they need that we can, like, scale that up.

And that's taken years to build. And, you know, obviously, we're not done there in our journey. But that by itself is, like, its own kind of moat.

And you also know when you've almost found product-market fit,right? If customers start to trust you, then that's really, like, when the network effects start to happen. When in the Mercury journey do you think you actually found that product-market fit and you really, you know, figured, "We have something here"?

So we really had product-market fit from the first day we launched, which I think is, like, relatively unusual. But, yeah, I think for the first year it was mostly I didn't believe we had product-market fit because I was really worried about, like, over-investing too early.

And so we were really, like, always behind. Like, we'd never had enough customer support. Like, we didn't have enough engineers. Like, everything was breaking all the time for the first year. And then eventually COVID hit, like, about a year after we launched.

And at that point, you know, the whole world was ending, and I was like, "Thank God we didn't invest too much." Like, you know, we still have, like, a ton of cash in the bank. But then a month after that, our growth, like, even grew even faster because, you know, all the bank branches were shut down and Mercury was one of the few places you could sign up.

And there was this huge explosion in, like, business formation and digital businesses and e-commerce. So we started growing even faster than we were pre-COVID. And at that point, I was like, "Okay, you know, even if a global pandemic doesn't slow down our growth, like, we definitely have product-market fit."

Like, that was a year in was the first point where I was like, "Okay, I'm willing to believe we have product-market fit." But, yeah, the reality was, like, yeah, I often tell people, I mean, at least in consumer and kind of startup and SMB, product-market fit is, like, when you're not hunting for users, like, they're just coming to you and you have no idea where they're coming from.

Like, that's been my experience with product-market fit. It's just, like, you wake up on a Monday and, like, you know, some, like, you've grown way more and you don't know quite why and you're just trying to keep up with that growth.

You know, one of the things I've been super surprised about, and I will be honest, when I saw all of these fintechs launching in 2017, 2016, you guys kind of, like, all launched at roughly the same time. Back then, I was like, "Surely all of these large conglomerates or, you know, legacy, you know, banks are just going to launch their own products."

But actually, if we look almost 10 years on, very few of them have. Why do you think they were just so slow to adapt and why do you just don't think they just didn't do it themselves?

So this was one of the questions I wanted to answer when I was, like, researching Mercury. You know, I really, this is my fourth company. I had already done startups for 10 years. I really didn't want to do something unless I was going to really commit for 10-plus years.

So I really wanted to understand why someone wouldn't, you know, why wouldn't JPMC, which is the biggest bank in the U.S., just copy us very quickly or do their own version. So there's two, I guess, two or three things I realized.

You know, number one, deposit banking, which, like, you know, when we think of banking, we're like, "Yeah, you get a checking account and a debit card and all these things." Like, that's not the main product that a bank thinks it's selling.

Like, that's a cost center from a bank's perspective. The main product a bank thinks it's selling is loans,right? That's where they make revenue. Like, if they give you a mortgage even, you know, a million-dollar mortgage,right? They're making $30K a year from that.

That's like, that's ACV,right? And that's small,right? Like, if they give you $100, you know, if you're a business and you get, like, a $100 million line or something like that, that's, like, serious millions of dollars in revenue. So the deposit side is a cost center, which is why they only invest the minimum possible in it,right?

That's, like, kind of the mindset of banks. They're trying to do the minimum possible. They don't want to make loss on small users. They're really, like, you know, they charge lots of fees and things like that. So that's one really big thing.

The second thing is the whole cost structure around this deposit side is, like, you know, very expensive. They're, like, very committed to bank branches. They've got thousands of them. And, you know, when your cost center is, like, pretty high, you have a very different approach to a small user.

You have a different approach to marketing. Like, you're really, you know, when I think of a new user, I'm like, "I want to try to reduce my marginal cost of serving them to zero," which means I can give them better features.

I can give them more economics, you know, better cashback, better yield, whatever it is. And I can serve everyone. Like, I don't have to say no to some people just because, like, my math doesn't work. I can make the math work everywhere.

So that's number two. And then number three, a lot of the banks don't really have engineering teams. What they have is, like, IT teams. And, like, there's a fundamental difference between an engineering team and an IT team. It's like, an IT team is, like, taking something.

There's this whole industry called banking cores. So the IT team takes these third-party vendors that make mobile apps and make websites, and they kind of hook them together and make sure they all work together. Whereas, like, you know, engineers actually build the product,right?

Like, they build and they have, like, a lot of, it's just a very different mindset. Like, a lot of the time, if a bank wants to make a fix, they have to get one of these cores companies to actually, like, do the work for them.

So it's hard for them to really innovate in that kind of paradigm. So those were, like, the three things that I thought. And, I mean, the last thing I'd add is I think AI makes the whole situation, like, 10x worse,right?

Because the, like, the rate of change is much higher. So if you imagine, you know, the traditional kind of incumbent banks haven't caught up with Mercury or, you know, the other newer banks in the space 10 years in, now everything is changing every six months.

Like, the, you know, you really have to be nimble and be able to, like, launch an MCP, you know, do all these features within, like, a few months so that, like, you know, you're actually keeping up with customer demand.

So I think the faster the world changes, the harder it is for, like, a slow-moving entity to kind of keep up.

The faster the world changes is, the faster the world changes makes it then easier for you guys,right? And it's almost like for these large, you know, legacy banks, if they can't move fast enough, then, you know, it's an amazing opportunity for you guys.

When you get to the scale Mercury is now, 103 US dollars use you guys on a, you know, pretty regular basis. I'd love to actually get inside the product team. When you think about launching new features, is it just a case of, "Let's just try it with 1,000, 10,000 users and see what works"?

How does the actual, like, product mindset, you know, evolve at the scale you guys are now?

So it really depends on what the feature is. Yeah, we definitely, like, you know, we'll do alpha testing, we'll do slow rollouts, and we'll make sure it, like, all works for, like, basically any scale feature. We'll try to, like, understand what feedback we can get.

And really, it's about getting feedback and improving the feature.

We, you know, for new things, like net new things, you know, we try to come up with, like, really small teams. So sometimes it's just, like, two engineers or it might be, like, you know, three engineers and a designer and one PM.

So we're, you know, for something completely new, like we launched Mercury Personal kind of earlier this year, which is our consumer banking. You know, that started off with a five-person team that really proved it out and got a waitlist product live and, you know, showed that there was, like, some product-market fit and market pull there before we really invested a lot more.

So I'm a big fan of, like, having these small teams, even if, like, it sounds big,right? Like, it was a move from, like, business to consumer, which is, like, you know, a huge deal in some ways. Like, you know, a small team can often prove a lot, especially if it's, like, kind of, you know, banking is something we already did.

So we understood the kind of basic core stuff pretty well. So that's how we think about, like, new products. Like, get a small team, prove it out. And, you know, I'm like, you know, once it's, like, proved out enough, then we invest it.

But even then, you don't go blow up the team. You just kind of incrementally invest and, like, wait for them to hit more milestones over time.

You know, I feel now it's so easy to launch features and the opportunity is so immense. I'd love to get inside the actual Mercury product team a little bit more around if you have a new idea, is it, do you just come up with the ideas yourself?

Is it, if Ollie works at Mercury, can he come up with this idea and come to you and say, "Look, hey, we have an amazing opportunity. I think we should give this a shot." Are you guys quite open to that as a culture fit?

So I think of, like, two types of ideas,right? There's ideas that are within. So we try to have, like, product teams that own different areas. Like, there's the cards team that owns our corporate credit card. There's the, you know, the personal banking team that owns a personal product.

There's the invoicing team that owns our invoicing product. So within those teams, they have a lot of autonomy and they can come up with ideas. Like, hopefully anyone in the team can come up with an idea and, like, change the roadmap and go implement it.

And often without, like, having to, like, work it up the chain or something like that. Like, I think that's the power of these kind of small, and we break down, like, all of our problems into, like, smallish chunks that, like, each team can kind of own end to end.

Obviously, if they want more engineering resources or they, you know, they need it to integrate between lots of different products, like, you know, it'll have to kind of go up. But, yeah, it tends to often be driven by the team.

Sometimes, obviously, if I have an idea, I'll try to, like, get it on their roadmap. And, you know, I guess I'm the CEO, so they have to listen sometimes. And if it's, like, a net new thing, like, it doesn't fit any product team, then it's, I would say it's often top-down,right?

Or someone in the team has an idea, but we do have to get, like, alignment from me and, like, my co-founder runs product and design, so from him. So there's, like, some alignment work that has to happen, especially if you have to instantiate a new team.

And if we're trying to instantiate a new team because we want to do something, like, net new that none of the teams fit in, we try to generally take a few people from existing teams. Maybe we'll hire one or two extra, but, you know, you kind of want the nucleus of the team to at least be, like, filled with, like, at least half existing team members that, like, understand the code base, understand what it takes to, like, ship something new within the Mercury product.

What I found amazing just witnessing how fast AI is changing is the nuclear power of talent is so immense now. How do you think about keeping teams super small versus also just hiring the very best talent?

I mean, I think the best talent was always important. Like, that's not, I don't think, I don't know if AI has changed that. Maybe the types of talent you look for have changed a little bit. You know, I think one of the slightly negative parts of AI is, like, I'm finding for a lot of different roles, you kind of want more senior people.

So, you know, if you can imagine with AI

encoding, a lot of the kind of junior roles, like, AI can kind of do. And you really want, like, an engineer to be, like, you know, helping architect, like, doing code reviews. Like, it's a different job and it's, like, a slightly harder job for engineers to some extent.

But that's also true in sales,right? Like, the entry-level sales job was SDR and, like, often, like, you can automate a lot of that kind of work. But, yeah, talent is even more important and the shape of the talent has changed over time.

And, yeah, maybe especially for new things, like, small teams can get a lot of stuff done. Yeah, existing very large code bases that have a lot of complexity in it, like, you know, AI is still, like, increasing productivity, but, you know, it's not completely changed the game in the way, like, you know, for net new features.

It's become way, way faster.

How much of Mercury's code is written by AI today?

It's hard to, you know, it's hard to get an exact number because, you know, everyone's using AI all the time. Like, you're using Cursor or you're using Claude Code. And I often think when you try to get a number, then it, you know, people who put out numbers often are, like, exaggerating, in my opinion.

But almost all of the code that is written in Mercury is AI-assisted, if it's not, like, written entirely by AI.

This whole AI category is 1,400 days old, give or take, since ChatGPT launched within November the 30th, '22. It's just changing so quickly. You're an investor. You're obviously a founder. You've been in this space for so long. Have you been surprised just how quickly this whole shift has happened and it's just changing so quickly?

Yeah, I think

the slightly, maybe I shouldn't be surprised about it because, like, you know, these very AI-built people always say this, but I do feel like the thing that matters and the paradigm changes quickly as well,right?

Like, open Claude and, like, now Claude Cowork is, like, that's a, like, a year ago, no one was thinking or talking about, like, hey, something's going to, like, be on your machine, connect to all your services, and, like, do everything for you kind of thing.

That's very new. But, you know, I feel like every 6 to 12 months, like, the paradigm shifts entirely. And that's, yeah, I love it. Like, I love change. And I think, as I said earlier, like, I think it creates a lot of opportunities for, like, startups that can be nimble and adapt to the change.

But, yeah, it is surprising. I think there's definitely, like, an element of where you're like, okay, you know, this is what it is and I expect it to continue. Like, even if you think about, like, okay, how does the AI improve from where it is today, you kind of think about, like, incremental improvements, but, like, these kind of step-change improvements that happen every 6 to 12 months are, like, really hard to, like, think about or predict.

I think one of, yeah, and I think one of the hardest things, like, you kind of mentioned that everyone, all the new entry grad-level roles, whatever you want to call it, all the new people coming into the workforce today, whether it's the universities for three or four years building, like, how relevant is that today?

And, you know, the crazy thing is these universities are super anti-AI. I don't know if you've seen this. So they're not, like, you know, I was talking to the CS grad and I was like, oh, yeah, you know, did you learn, like, any of this stuff?

Like, did you do Claude Code or any? And they were like, nope, they didn't touch any of that at university at all. So they're now equipping people with, I mean, obviously, people can learn for themselves, but, yeah, it's definitely, like, hard out there, which, you know, ideally, they would do a better job of equipping people with this.

But universities, unfortunately, are, like, similar to incumbent banks. Like, they're not very fast to adapt to things.

Coins of our show will take, but I don't know how relevant universities are going to be in the future. We had this amazing founder just on the show called Jolene. She's the founder of Scoro AI. And essentially, it is a Chinese company, but it's super interesting.

And they're essentially replacing teachers with instructors. And we got talking around actually, like, just how relevant, you know, university is going to be in the next few years. That is an open question. Don't know where that's going to go.

But the correlation question here is, I think educating people inside companies around actually just what's happening with AI is super difficult because it's changing so quickly. Do you guys have, like, an AI channel where people can get, you know, updates on the latest AI news, features?

Like, how does that education actually work internally?

We have so much stuff. So, yeah, there's definitely.

That's awesome.

There's one channel that's, like, yeah, everyone, we call it Learn AI. Often teams have their own channels where they're talking about things. We do a lot of, like, per-team and company-level kind of demos where people are like, hey, this is what I built.

Like, we just did one the other day where someone's like, oh, I built an AI chief of staff. This is how I did it. This is all the services I plugged in. These are the types of questions. This is me working through it.

So, yeah, I mean, like, the thing that you do with AI changes over time. And those types of things where other people see how other people are doing it creates, like, a lot of inspiration. And then more recently, we actually have an AI enablement team and they're building a lot of kind of platform tools.

So, like, we have our own internal version of, like, a ChatGPT type thing and you can also, like, make apps with it. And it's already plugged in with all the data that you might need and, like, you know, you can deploy very easily with it.

And this hasn't quite happened yet, but we're hiring for it as we're going to have, like, these kind of internal AI forward-deployed engineers. So, like, we want to make it so, like, there's an engineer that's got all these AI tools and can go directly into the sales team and automate their processes and, like, you know, clean up their lead list and, like, do all of this stuff and, like, just bring that, like, you know, not every team can be at the forefront of everything happening in AI.

And obviously, like, you can get people that, like, really know what's up and all the tools to be used to, like, deployed at, like, you know, in compliance, sales, marketing, all these kind of different departments.

And we're still in the early stages. We're only four years into this AI change. I don't know where it's going to happen in the next 12 months, where it's going to be. But I think for the CEOs and the founders' role, it makes it super exciting, but even more challenging, frustrating.

When I speak to founders on the show, one of the biggest questions the audience always loves to understand is actually what does the CEO do on a day-to-day basis? So if I was going to be your chief of staff for a day, what would a day look like?

That's a good question. So I have a, you know, bunch of standing meetings. I stopped doing all one-on-ones. This is a whole, actually listened to Jensen Huang over at NVIDIA and he says he doesn't do one-on-ones and he has 60 direct reports.

So I'm not as extreme as him, but I was like, okay, you know, actually I don't find that much value in the one-on-ones. So I dropped all one-on-ones and instead I did, like, group meetings around, like, different subjects.

So I have a meeting with the GTM execs. I have a meeting with the engineering execs. I have a meeting with the EPD and compliance and legal and risk. So I do these group meetings and we can, like, actually get into, like, meaty topics that, like, solve things rather than, like, these one-on-ones, which tend to, at least I don't find, like, you actually, like, go do things with it.

It's just, like, you know, have all the people around the table to do things. So I have those types of meetings and then I have this, like, thing on Fridays called Exec Workshop where, like, you know, product teams and other people can, like, come to us and present, like, what they're doing and get feedback and get direction and things like that.

And then over time, I tend to, like, dive into whatever's, like, you know, the biggest opportunity or the biggest problem in Mercury. And I'll, you know, it surfaces for whatever and I'll make it, like, my problem for, like, two or three weeks and I'll go talk to, like, 10 people about it and come up with a plan on, like, how to fix it.

And, you know, and I end up doing this across the whole organization whether it's, yeah, it doesn't matter what the problem is. I'll try to get a handle on it and see what we can do. So, yeah, sometimes it's, like, product-related.

But, yeah, I think, yeah, there's these kind of standard things you have to do. I'll say one more thing is hiring. Hiring is always, there's always someone that we're hiring that's, like, important that I try to be involved with the hiring for.

But, yeah, I often think, like, you want to give yourself enough bandwidth as a CEO to be that chief kind of problem solver or chief opportunity chaser. And it's, like, a little dangerous if you just, like, fill all your time with, like, just, like, this routine and you're not, like, you can't, like, dive deeply into things that really matter.

I think the biggest challenge for founders is how do you prioritize time? And I heard a great quote at the beginning of this year, which actually I started learning, is you just have to be selfish in a good way.

If one-to-ones aren't making any direct impact into your life or the day-to-day of the company, then it's not as relevant,right? You have to be continually chasing those opportunities and, you know, problem-solving those challenges. Again, if I was to be your chief of staff and we're figuring out this huge opportunity, was this a recent opportunity you spent those the last two to three weeks building, maybe at some point in the last six months, but it's now become a reality inside Mercury?

So actually, like, a lot of these kind of new features we're launching around, you know, I talked about that inside product, the AI product, and the command product started off actually a year or actually not even a year, like, seven, no, eight months ago, I wrote, like, a doc.

And do you know this process? It's called working backwards. It's like an Amazon process. But you basically write the press release and then you work backwards to, like, yeah, because, like, a press release is kind of a good document because it talks about, like, well, at least this kind of press release talks about, like, this is the thing we're delivering to customers and these are the features and this is what people are excited about.

So I wrote this for, like, these features. Like, I was like, hey, you know, we had very little AI that was, like, directly in the product and I was like, okay, you know, this is what I'm imagining it could do and this is why people would be excited about it.

So I wrote this, like, basically, like, six, seven months ago. So a lot of our roadmap this year kind of followed from, like, that doc. And I'm really proud of the team for, like, you know, delivering on, like, a fairly vague doc that was just like, oh, yeah, I think all of this stuff is possible.

Like, go do it. But, yeah, that was, like, something that was fun. It was a fun intern process because, like, you know, I had to kind of really think about, like, what we should do and how we should think about it and the team kind of, you know, ran with it from there.

So something I was looking at last week, I was noodling on Crunchbase and PitchBook, who are the most active investors in the ecosystem? And your name popped up. 3,400, I think, personal angel investments you've done now or something?

Yeah, since 2016.

Why, how, first of all, how do you find time to, you know, figure out, like, you know, what companies and startups to back? And then, you know, what's keeping you most excited at the moment?

So once you kind of get going in the investing space and people know you invest, you get a lot of leads. And obviously, because of Mercury and, you know, the fact that one in three startups use us, a lot of people know about me.

I have, like, 250,000 followers on X as well. So, you know, getting leads is not hard. In fact, it's a harder job to, like, filter them and, like, find the ones you want to talk to. I actually work with someone on the investing side.

Yash is the name. So, you know, that helps kind of filter those leads and things like that. And then, you know, we don't, or I don't lead around. I, you know, I tend to participate in rounds that other people are kind of coalescing around.

So that makes my life a little easier. You know, when you're leading around, there's a lot more pressure to do, like, this kind of complete diligence and do a lot more work. So, you know, to some extent, someone else is doing a lot of that work and I'm kind of joining in around with, you know,right now I write normally a 250K check.

So I can make a decision pretty quickly. It's normally one or two conversations and a little bit of research. And, yeah, I mainly do it because I think it's just so much fun to talk to entrepreneurs. Like, I think, you know, to go back to your question or, like, where do you spend time, I find that if you're doing things that you have fun doing, it's not work,right?

Like, it's just, I'm like, oh, yeah, tell me about what you're working on. Like, you know, these pitches, I don't let people, like, go off and talk on a deck. Like, I'm just having a conversation with entrepreneurs and I try to be helpful.

So that gives me a lot of energy. And it also helps Mercury just because, like, you know, you get to be at the forefront as an investor and see what, like, everyone's working on. And also, like, yeah, sometimes I, like, look at a startup and I'm like, wow, like, these three people achieve so much.

You know, why can't Mercury with, you know, we have 1,200 followers. Like, why can't we achieve more? So kind of it keeps you rooted and, like, yeah, you know, what does a high-paced environment look like? Which, like, obviously, a tiny startup is a very high-paced environment.

So I think it's like, it's been a good practice for me.

I love that. There are so many opportunities now. Where do you think are some of the most interesting and underappreciated opportunities?

Yeah. You know, one of the problemsright now is, so, you know, a ton of investments happening in AI. I end up investing a lot in AI. Everything that becomes, like, a mainstream opportunity gets, like, instantly swarmed. So, you know, there was a bunch of startups, I guess, about a year and a half ago that were like, we're going to do accounting for AI.

And they were like, you know, building tools for accountants and, like, they were building accounting software and, like, so many,right? I even invested in, like, three. And I'm still, like, excited about the category, but, like, you know, it gets, like, all those get, like, funded and a lot of them got, like, you know, funded to, like, 10, 50 million, whatever it is.

But that space gets very hard to then do a new startup in. So, like, at this point, doing an accounting, unless you have, like, a very different perspective on, like, how everyone else is doing it, it'll be very hard to, like, enter the space.

You know, the same thing happened with legal. Like, a lot of these spaces get, like, done. I've been really interested in, like, trying to find things outside the AI space just because I do think it's, like, kind of overinvested and overvalued to some extent.

So, you know, I quite like deep tech, space tech. I've done a bunch in space. Done, like, you know, like Stoke Space, which is like a rocket company. I've done some nuclear stuff like Arlo Nuclear, which is kind of doing a small nuclear reactor and doing really well.

I did do Etched, which has done really well. That's like a semiconductor company. That is AI-focused, but, you know, I guess now AI semiconductor startups are cool, but, like, I did it two years ago when, like, people thought it was crazy that anyone would do a startup in that space.

So you kind of have to, like, be a little contrarian, like, just go, like, what are people not caring that much aboutright now? And I think it's very hard to invest in, like, the core of, like, AI apps or things like thatright now.

I just had Eric Hippo from Larry Hippo on the show and we got talking around these AI wrappers. Are they actually going to last the next 12, 24 months? What's your take on that? Because as soon as an AI wrapper comes out and the likes of OpenAI or Claude CL, this is quite an interesting tool.

We can just replicate it. Then they're going to have a hard time competing.

So I think it depends. I think if it's a very horizontal tool,

then, yeah, I think it'll get, you know, like, something like OpenClaw, whatever,right? Like, it's such a horizontal tool that, like, you know, Claude's got co-work and I think Google has something that does it as well. So, yeah, these horizontal tools are really hard.

But I think if you pick a vertical, I don't know, if you do an AI thing that works really well for dentists or works really well for

gyms or I don't know. There's, like, so many of these, like, it really has to be differentiated enough that, like, you know, that those people find, like, big needs for big benefits from it. But I don't think the risk is, like, you're going to be copied on every niche idea.

I'm not saying it'll be easy because you're also competing with everyone else that's coming up with these ideas. But I think if you have, like, a really, you know, what's not easy is a deep understanding. Like, if you deeply understand a space,right?

Like, you've been building, like, software for dentists for the last 10 years and you're, like, you know, you know how to distribute to them. You know what their pain points are. You know, like, you know, what actually their needs are.

Like, that's very hard to compete against,right? Like, that's, I don't think AI has changed that that much,right? Like, I think there's always, like, you know, if you have this depth of understanding, then you can, like, always build a big startup and, like, AI has just made a lot of opportunities there for people who do have that deep understanding.

I think also an interesting opportunity is around how do we allow people to create and have new experiences, get people back, like, actually meeting in-person community. I know it's super hard and the returns for investors haven't been that great.

Is that something you've looked at?

I think it's hard. I think all consumers are quite hard. You know, Particle is doing really well. Like, I don't know if you use this thing, but, you know, I'm finding that, like, yeah, events and in-person has, like, moved a lot to, like, Particle as an app.

At least I use it. All my friends seem to use it. So, and that's, like, traditionally been a pretty hard category. So, you know, I mean, they've been around for five years or something, but, yeah, consumer is, like, in general, a really hard space.

I think there are probably a set of new consumer startups that will arise because of AI. I don't know what it is, but I also think it'll be very hard to build a big business in this space. But, yeah, there will almost certainly be, like, something that comes from it.

Particle is like this event hosting management tool,right?

Yeah.

Yeah. I saw a tweet the other day and I thought it was just an interesting take that a lot of product builders and startups now are using Particle to actually launch their new products because you get automatic, like, WhatsApp updates, which I thought was an interesting, interesting way to use it.

Yeah. So if you guys ever do another, you know, product launch, then, you know, maybe use Particle for that.

Yeah, that'd be an interesting way to do it. And then I guess you can build up an audience there. That's cool.

Exactly. You get like these WhatsApp updates and so on. Pretty, pretty cool. So much is happening, but to end the show, we always love to do a quick-fire round. So a couple of questions, rapid-fire, and just get a bit of interesting more insights into the personal life and also just what's happening wider.

How's that sound?

Yeah, love it.

Awesome. So who is your favorite AI founderright now, which you most respect?

I mean, this is maybe too obvious an answer, but you really got to give it to Dario. You know, I think all eight co-founders at Anthropic are still there. And he just hired another OpenAI founder. Andrej Karpathy just joined.

So, like, they've, you know, they've built a company that has that culture where people want to stick around and build a company for the long term, which I think has been hard in the AI space. And just like the, you know, the rate of growth, the product velocity, it's just incredible.

And, you know, they keep delivering. And actually, if you think about the story, it's a crazy story. Like, they came, it's kind of like a spin-out from OpenAI. Like, they came from, like, you know, basically zero to, like, now making more money than OpenAI.

Twelve months ago, I thought OpenAI is totally going to dominate the space, but now my views have changed,right? Same for you?

Yeah, 100%. I mean, I don't, I think it keeps changing and I don't think OpenAI is, like, out. Obviously, it's a very strong team and brand. But, yeah, Anthropic has definitely pulled ahead.

Which is your favorite AI company outside of these AI models?

I mean, a couple I'm invested in, so I'm kind of biased. Decagon has done really well in the CS space. And then I'm also invested in Etched, which is really doing well in the semiconductor space.

I think founders and CEOs' mindset is always so interesting. There's always just one thing that perhaps some people don't quite know about you. What is the one CEO characteristic that maybe even your team don't know about you?

The two most important things for me are, like, at least, like, characteristics that I try to have that I think are, like, really positive characteristics are being optimistic and being curious. And I really feel like if you have those two, like, you're optimistic and curious, like, your life is kind of set,right?

I don't know why everyone isn't like that because, like, if you're optimistic, like, everything is just, like, you just assume it's going to work out well and you feel, like, much more positive about things. And if you're curious, like, you're always excited to learn.

Like, you know, there's always something to learn from and changes, like, exciting to you. So it's always seemed to me like those two things are, like, such free attributes that you can just, like, if you can adopt them and maybe you can't change your personality, but if you can adopt them, like, your life is just so much better.

So everyone should just do that.

I love that. To remain optimistic, do you mostly just read and spend time with founders? Obviously, like, you're building for founders, so you probably get so many cool insights anyway. Is that the best way to stay optimistic?

I mean, to some extent, it's a choice. Like, I think you kind of have to, you can easily, yeah, there's a bunch of habits,right? Like, you can easily go into, like, these kind of social media doomer holes and be quite pessimistic.

So, like, yeah, it's like, what's your, like, information diet and, like, how do you process things? But, yeah, I mean, I'm also just, like, very grateful for what we have. Like, I think it's very easy to get complacent with things you have,right?

Like, I'm just happy that I get to eat really well every day and, like, have a roof over my head and, like, have this, like, iPhoneright here. That's, like, this amazing kind of piece of technology. Like, I just feel like if you're, like, very grateful about the world, like, you tend to be pretty optimistic.

I think when you become, like, a little more, I don't know, entitled and, like, just expect everything to be given to you, then you're, like, a little more, like, pessimistic about the world maybe. But, yeah, I mean, what do we not have optimistic to?

Yeah, what do we not have to be optimistic about? Like, I think it's like, you know, we have this, like, crazy technology that's, like, changing the world. So I feel like it's, like, a choice to some extent.

If you weren't building in fintech, any other categories that you think are super hotright now that you'd go all in on?

I mean, the only other thing that I get really excited by is space tech. I think there's, like, both a rational and irrational reason. Rationally, you know, SpaceX has made it cheap enough to, like, send things to space.

The ecosystem has got mature enough that, you know, there's, like, all these, like, microsat companies and, like, you can, you know, you can get things in space way, way cheaper. Like, I've been investing in space for 10 years and, like, 10 years ago, it still took, like, $30 million raised to, like, get anything in space.

And now people are doing it with, like, a couple of million dollars. So I think the opportunity to do things in space is big. And I think there's a lot of ideas there. And then on the irrational side, I think it'll just be kind of cool.

Like, you know, I've grown up, like, reading a lot of sci-fi and watching things. And I think the idea of, like, you know, touching space is, like, such a cool idea. I'd love to do that.

Well, we have the SpaceX IPO happening at some point in the next few weeks, hopefully. And then.

Is that a few weeks away?

Yeah. Yeah, I think so. Yeah. And hopefully at some point we'll get the opportunity to go up there.

Yeah, hopefully. I'll only do it if I'm, like, almost guaranteed I'll make it back.

The same. Hey, Immad, this has been amazing. Thank you so much for your time and congrats again on the fundraise. This is super exciting and it creates a lot of great things.

Yeah, thanks for having me, Ollie.