Afleveringen
-
James Cadwallader and his co-founder Dylan Babbs were working six days a week in an office before they had a company, or even an idea. James would drag Dylan into their coworking space on Saturdays, ban laptops, and fill a whiteboard with post-it notes; for a stretch, the leading idea was lithium mining. He calls it hardcore from day minus one. Twenty-one months after launch, Profound is a $1 billion company with more than 700 enterprise customers and over 13% of the Fortune 500 on the platform.
The thesis: marketing is about to change more than it has in the last 25 years. The old internet playbook was built for humans browsing the web. The next one is being built for AI agents interpreting it. When ChatGPT recommends a product, it doesn't just give you a link. It opines, 3,000 characters at a time, and itâs a marketers job to drive whatâs in those characters. James watched the last platform shift from inside the room, building Kyra and running creative for Nike, Unilever, and Prada while social and TikTok ate the old marketing playbook. Now he's building the infrastructure for this one.
In this conversation, James maps the new marketing operating system. Why you're no longer marketing to John Smith but to John Smith's agent. The âMarketing Engineerâ, a role Profound coined and Google has already hired for. Why ideas become the bottleneck when content costs nothing to produce. And the culture James calls the best product Profound has built: six days a week in office, and you can still leave at 6pm without anyone batting an eyelid.
Listen on:
⢠Spotify: https://open.spotify.com/show/5KorHDryscDzLwPkWEgAH1
⢠Apple Podcasts: https://podcasts.apple.com/us/podcast/knuckle-up-with-nakul/id1893213920
James Cadwallader is the co-founder and CEO of Profound, the AI visibility platform that helps brands measure and shape how AI assistants like ChatGPT, Gemini, and Perplexity talk about them. Profound serves enterprise customers including Target, Walmart, and U.S. Bank. Cadwallader started Profound in New York City with co-founder Dylan Babbs after the two met at the startup incubator South Park Commons. Before Profound, he co-founded Kyra, a London-based creator marketing company.
Where to find James Cadwallader:
⢠Profound: https://www.tryprofound.com/
⢠X: https://x.com/thejamescad
⢠LinkedIn: https://www.linkedin.com/in/jsca/
Where to find Nakul:
⢠Audacious Ventures: https://www.audacious.co/
⢠X: https://x.com/nakul
⢠LinkedIn: https://www.linkedin.com/in/nakulmandan/
Where to find Audacious Ventures:
⢠Website: https://www.audacious.co/
⢠LinkedIn: https://www.linkedin.com/company/audaciousventures/
In this conversation with James Cadwallader:
00:00 Who is James Cadwallader?
03:56 What happens when ChatGPT recommends your product?
08:33 How do you market to machines instead of people?
11:10 How is this different to SEO?
15:00 What is a Marketing Engineer?
17:23 What is the antidote to AI slop?
20:56 When marketing copy is free, what's the new bottleneck?
24:25 Can Profound get bigger than Adobe and Salesforce?
28:04 Why does Profound's marketing team have only eight people?
33:57 How did Profound become a unicorn in 18 months?
37:22 Post-it notes and lithium mining: how did Profound begin?
40:49 Which AI tools actually run Profound day to day?
47:29 Why does James call culture the best product Profound has built?
53:54 What breaks when a startup crosses 250 people?
1:01:12 What is James most paranoid about?
1:04:44 Quickfire: workflows to kill, AI-era brands, business books, AI SDRs
1:06:00 What's the recommended 90-day plan for a Fortune 500 CMO?
James's sharpest lines from this conversation...
On the new question for every brand:
"In the old world of search, it was just this question of how do I rank? And in this new world, the question is not just do I show up, but how do I show up?"
On the antidote to AI slop:
"The antidote to slop, from my experience, is more context."
On what AI still can't know:
"There are things that you know about Audacious right now that Chat and Claude simply don't know... You're an API between reality and AI at this point."
On his ambition for Profound:
"I'm never going to have another opportunity like this. This is it. This is my shot. So yeah, I'm kind of planning on trying to take it all the way."
On enterprise speed:
"We work with giant Fortune 10 brands who are a lot faster than you'd expect."
On culture:
"I've said a few times that our culture is the best product we've built. It's the product that I'm most proud of."
On who thrives at Profound:
"We look for masters of craft and people that are default hard working. That's what they do if left to their own devices."
On what he tells the team at all-hands:
"I'm very curious to see who in this room sticks it out when things actually get hard. Because right now we're playing the game on easy mode."
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com -
Most founders get more stressed as the company gets bigger. Immad Akhund has spent years engineering the opposite. âIâve tried to make it so that the pressure kind of goes down as success goes up, which I think is relatively rare,â he says. âIf youâre more successful, why are you more stressed about it?â This may be surprising coming from someone running one of the most loved fintechs of this generation. But Immad has reached this point through a deliberate operating philosophy that he shares in this discussion.
Mercury serves around 300,000 customers, runs at a $650 million annualized run rate, and has been profitable for four years straight. One in three US startups banks with it, and in April 2026 the company won conditional approval to establish Mercury Bank, N.A. Immad has done all of it on his own terms: anti-996, still remote-first while Silicon Valley snapped back to fully in person, and hiring for curiosity and humility. The throughline is that he optimized early to be a founder for life, not for a sprint.
In this conversation, Immad walks me through how that philosophy actually runs a 1,200-person company. Why 996 may not make you more productive. What Mercuryâs famous curiosity interview is really testing. Why he killed one-on-ones after listening to Jensen Huang. How AI has turned the disruption of banking further in Mercuryâs favor. And why, three startups in, the thing he would tell his younger self is to build something he would actually love to exist.
Listen on YouTube, Spotify, and Apple Podcasts
Immad Akhund is the co-founder and CEO of Mercury, the fintech banking platform founded in 2017 that serves around 300,000 businesses and runs at a $650 million annualized run rate. In April 2026, Mercury received conditional approval from the Office of the Comptroller of the Currency to establish Mercury Bank, N.A., and the company subsequently raised $200 million at a $5.2 billion valuation. Before Mercury, Akhund co-founded Heyzap, a mobile developer platform that went through Y Combinator and sold for $45 million in 2016, and earlier co-founded Clickpass. He has served as a part-time partner at Y Combinator and is a prolific angel investor backing more than 350 startups, including Airtable, Substack, and Deel. He was born in Pakistan and raised in the UK before moving to the United States.
In this conversation with Immad Akhund:
00:00 Who is Immad Akhund?
01:53 Is culture just the founderâs personality externalized?
07:53 Is 996 actually less productive?
13:14 Why is Mercury still remote-first when Silicon Valley went back?
15:52 How did Immad hire Mercuryâs first ten people?
20:33 What is Mercuryâs famous âcuriosity interviewâ?
25:06 How do you hold the hiring bar at 1,200 people?
28:48 What does Immadâs day-to-day look like?
34:10 How do you know when an exec is no longer right for the job?
37:32 Why does Mercury run on small, autonomous product teams?
40:27 What breaks when a company runs purely on metrics?
47:03 How is AI changing Mercury?
54:09 What does becoming a chartered bank actually change?
56:19 How did the SVB collapse end up helping Mercury?
1:03:40 What does psyche management look like 20 years in?
1:11:19 What does it mean to be at the founder âbonus levelsâ?
1:15:21 Quickfire: overrated traits, AI blind spots, and being proven wrong
1:17:52 What would Immad tell his 25-year-old self?
Where to find Immad:
* Mercury
* X
* LinkedIn
Where to find Nakul:
* X
* LinkedIn
Where to find Audacious Ventures:
* Website
* LinkedIn
Immadâs sharpest lines from this conversation:
On pressure and success:
âIâve tried to make it so that the pressure kind of goes down as success goes up, which I think is actually relatively rare. If youâre more successful, why are you more stressed about it?â
On the math of a 996 day:
âThe extra two hours, was that really 20% more productive? I will be shocked if it was 20% more productive. Itâs maybe 5% more productive.â
On where culture actually comes from:
âIt does start with founder personalities, but then you try to structure that, try to encourage that internally, then you hire against it. Then it becomes a self-perpetuating thing, and it also shows up in the product and how you talk to customers.â
On why he killed one-on-ones:
âIâve never found one-on-ones help. One-on-ones is just an avenue for them to try to sell themselves. You need to create avenues where you have visibility into what that team is doing.â
On holding execs accountable:
âI donât think execs can have excuses. I think thatâs the job, to deliver on that.â
On why AI favors Mercury:
âI highly doubt Bank of America has an MCP. We are extremely well situated for that because we are a technology company, and thatâs just what we do.â
On reinvention:
âI love change. I live for change. Change is the main thing I want in life.â
On his advice to his 25-year-old self:
âBuild something that you would really love for it to exist, rather than being too spreadsheet about it.â
Full Transcript: Immad Akhund on Knuckle Up
Immad Akhund:
9-9 is 12 hours. The extra two hours, was that really 20% more productive? I will be shocked if it was 20% more productive. If you have five employees and you hire a sixth employee, thatâs 20% more productive.
Youâre probably a little less creative if youâre not doing anything else in your life. I have my best ideas when Iâm on a flight or driving somewhere. Sometimes these decisions or these ideas are actually worth 10x more than an extra few hours of work.
I started as entrepreneur. I loved it, but I just wanted to survive.
We got to a billion dollar valuation in 2021. At that point, which is like, I would say I really got into the bonus levels on this game where I was like, âI donât even know what my next goal is.â Iâve tried to make it so the pressure goes down as success goes up, which is I think actually relatively rare.
If youâre more successful, why are you more stressed about it?
Nakul Mandan:
Immad Akhund has built one of the most loved FinTechs of this generation. Mercury serves 300,000 customers and is at a $650 million annualized run rate. One in three US startups banks with them. The company has been profitable for four straight years, and as of last month, they have conditional approval to become Mercury Bank. But whatâs interesting is that Immad has accomplished all of this on his own terms. Heâs anti-996. Heâs still remote first at a time when Silicon Valley has bounced back to fully in-person. He hires for traits most growth-stage companies donât even look for. He said, âItâs dangerous when everything has to be driven by metrics.â And he said openly that heâs optimized early on to be a founder for life and not for a sprint. Today we talked to Immad about how he runs Mercury, the operating philosophy that got him here and whether that philosophy holds up as Mercury becomes a bank and AI-shaped company and a business serving companies well beyond the tech ecosystem. Knuckle up.
Immad, welcome to the show.
Immad Akhund:
Ah, thanks for having me.
Nakul Mandan:
Yeah. So I want to start with something youâve said that actually cuts through a lot of the culture theater in Silicon Valley. Youâve said that the culture is about the founderâs personality, not about some values doc, not about some rituals. The founderâs personality externalized. And I was wondering as I was preparing for this, that works for 10 people, 100 people company, maybe even stretch it to a 500 people company, but at the scale at which Mercury is today, 1000-plus people, does it still work?
Immad Akhund:
Oh, yeah. 100%. I mean, I think I often talk to people who talk about culture, but I think they donât know what it means, and I think itâs like an ill-defined word and Iâm an engineer, so I hate ill-defined words. So I really struggled with it for a long time. I was like, âIs it like having beers in the fridge and going, âThatâs what it was,â or is it like what programming language you choose, et cetera?â So I think there are probably some people mean culture and they mean something else. I think the easiest way to define culture that is very definitive and means something and you can measure it and all that is what are the personalities of the people you hire, what are the personalities that you encourage, what are the types of traits you encourage, kind of thing.
You kind of have to decide that very early on because what happened in my previous company is I didnât really understand that. So we had a bunch of salespeople that kind of ended up dominating the culture because they were the loudest voice in the room and we didnât really have a counterculture to fix that. So it was fine, but it wasnât like the type of company I wanted to build with Mercury. So we sat down and I think there was only four people and weâre like, âOkay, what are the personalities that we really cared about?â And for us it was kind of product-minded, humble, helpful people that were kind of curious and at that time we didnât like meetings so we were like no process, go build things kind of people.
And so we wrote that down and we tried to measure against it. We had interviews like every engineer, everyone we did product interviews with, and we would try to test people for, okay, are they high ego? So thereâs a bunch of things that followed from that and we have a bunch of things within the company that encouraged that behavior. We pick one thing, like we have a channel called Grateful and people at the company, like every day thereâs multiple posts where people are like, âOh, Iâm really grateful to X and Y because they helped launch this project or whatever it is.â A lot of that behavior is organic. I didnât make that channel, it just came across. But I think itâs important and I also think different personalities can work.
I always think of Uber as an example of a very competitive, high eco kind of culture, at least in the Travis days. Obviously they built a very successful company doing that. It doesnât work for my personality. I also think it doesnât work for the product weâre building. Yeah, maybe it could work, but I think Mercury is a very customer-centric kind of product where weâre very transparent and open with our customers and we try to be. That authenticity is part of the reason I think Mercury is successful. It does start with founder personalities, but then you try to structure that, try to encourage that internally, then you hire against it, then it becomes like a self-perpetuating thing and it also shows up in the product and how you talk to customers.
Nakul Mandan:
Does it potentially lead to also though having a lot of the same personality in the... Can a different personality come in into the company?
Immad Akhund:
Well, you need a little bit of just different energy sometimes, but it can sometimes ruin things. Sometimes you let someone who is a bad fit come as a leader. I mean, I think itâs especially an issue with exec level. I think at IC level, you can have a lot more variance and I donât think it spoils things too much. But at a leadership level, if you have someone and then they hire people against them and for whatever reason that clashes with the rest of the company and then you can have a team thatâs in a silo that... It can lead to a lot of bad things, and we are very careful at the exec level where we take a lot longer hiring and we try to test against our culture fit quite a bit. But itâs not like... You never have 100% conformity and you donât necessarily want that either. You need enough of it that you are really being true to that culture and not letting it deteriorate.
Nakul Mandan:
What about the other side of it, which is the founder also evolves in the journey. Mercury has come such a long way. The Immad that started Mercury versus the Immad who is today. So has the culture evolved with your personality?
Immad Akhund:
We end up assessing it probably once a year and changing it once a year.
Nakul Mandan:
So it is a living doc.
Immad Akhund:
It is a living doc. Change is difficult to some extent because you donât want to be drastically, âNow we believe in something completely different and opposite.â So what we tend to do is evaluate it from like, okay, the set of cultural things we have, are they serving us? Yeah, take that. We literally used to have one, like no meetings. I canât remember the exact phrasing, but sometimes the fastest way to get to a decision is a meeting.
Nakul Mandan:
Oh, was it no weekly meeting?
Immad Akhund:
No, it was a very anti-meetings-like phrase. It was like, âWe donât like meetings.â And it was fine. When thereâs 10 people and youâre all in a room, it doesnât matter. But it wasnât serving us and we got rid of that. And so thereâs some that you remove and then thereâs some that youâre like, âHey, we really want to encourage this behavior.â You have to also try to shift culture by creating something. So I think one of the more recent one was we were like, âFocus on outcomes,â and then we created a bunch of things to... We try to create a bunch of rituals around the culture as well. So itâs like weâre focused on outcomes. So what are the outcomes we focus on? How do we drive that? How do we encourage it and celebrate when people are doing the right thing? How do we put in the performance management process so someoneâs going to get promotion? Part of it is like, okay, maybe they did all the right things, but the outcomes didnât happen and how do we measure all of that?
So it has to be something that shifts over time. I think to some extent my personality changes over time, but not that much. I mean, I guess I started Mercury 80 years ago when I was 33. There hasnât been drastic changes in personality. I mean, I learn a lot, but yeah, the core stuff is still the core.
Nakul Mandan:
In Silicon Valley these days, thereâs a lot of talk about 996 constantly. Itâs gone towards fully about intense cultures and youâre on the other side. Youâve built a $650 million revenue business. Obviously you have scaled in a way that very, very few people in history have scaled businesses. Can you talk about what does the 996 culture get wrong? Whatâs your though about this?
Immad Akhund:
From day zero weâve had... We work hard, but itâs normally five days a week, whatever, 9:00 to 6:00 or something like that. We donât have a religion around it. I have a family. I already had two kids before I started Mercury, now I have three, and I really see entrepreneurship as a long-term thing. Iâve done it for the last 20 years. I want to keep doing it for another 20 years. So I really want to do it in a way that it fits into my life. I used to be a 996 person, not because I was like, âOh, I like 996.â It was just like I didnât have kids, I wasnât married. I woke up in the morning and I started... I was more than 996. I woke up at 9:00 and then I just worked till midnight being like, âMaybe you go out for dinner, maybe you donât, and then who cares if itâs the weekend because youâre not doing anything.â Thatâs basically how I ran the first few years of my startup life.
I think to some extent thereâs been a shift with AI where being young is no longer a problem in the sense that with B2B SaaS, if you were young and you donât have the experience to see B2B problems, it was hard to build a B2B SaaS company, which is the previous wave in the startup world. Now with AI, everything youâve learned is probably wrong and you have to kind of reinvent it.
Nakul Mandan:
Itâs liability problems.
Immad Akhund:
Yeah, itâs maybe even a liability. And also thereâs just a lot of people out there that just want to play with the technology that are young, that... I started startups when I was 22. Itâs the best time to do a startup. You have no obligations, you just need enough money to survive with the least amount of lifestyle cost. Yeah, thereâs a big wave of young people. So I think thatâs part of the shift that they donât care.
And I found personally when I was young and I was sitting on my computer all day, the reality is I wasnât productive all day. I was on Twitter and Iâm a pretty good programmer and thatâs what I used to do in my early startup days, but thereâs just only so much really productive in the zone programming you can do per day and the rest of the time youâre kind of pretending, sitting on a computer. So yeah, I think if youâre productive, like 9:00 to 6:00 and youâre just cranking away, I think you can be just as productive.
Nakul Mandan:
But your point is also that 996 is not good for the marathon of startups.
Immad Akhund:
Yeah, I donât think itâs sustainable. I mean, Iâd love... Iâm sure thereâs exceptions. But is it fun? I mean, is it fun not having a life? Thatâs the reality. I want to have fun. Itâs like lifeâs too short to not have fun while we do things. I have fun building startups. Itâs not a prison for me. Wow. Iâm just going to... And maybe itâs fun for some people to just... I know people that it is probably fun for them to do just that and only that, but I like having friends and like spending time with my family.
Nakul Mandan:
If I steel man their argument a little bit, one could say, âHey, the first few years of a startup, itâs so existential. If you donât put everything into those first three years, especially... 996 is just a phrase, but all in all the time, then the longevity will not even matter,â versus optimizing for, âHey, Iâm going to be an entrepreneur for 20 years, maybe even longer,â which is what youâve said. Would that argument hold, would that be a fair critique of, hey, the first three years, man?
Immad Akhund:
Yeah, I guess itâs okay. I mean, hereâs my counterargument to it. I genuinely want to know... Letâs just take one day in this personâs life. 9-9 is 12 hours. The extra two hours from 10 hours to 12 hours, was that really 20% more productive? And I will be shocked if it was 20% more productive. Itâs maybe 5% more productive. I think itâs just a reality of this. You just canât be that much more productive. Thereâs only so much productive time you can have in a day. Okay, so maybe youâre getting five... I guess youâre throwing a Saturday in there. So letâs say you get 10% more productive. If you have five employees and you hire a sixth employee, thatâs 20% more productive. Iâm not sure, theyâre working themselves to the bone. I think youâre probably a little less creative if youâre not doing anything else in your life.
Even when Iâm in the weekends and Iâm not working, Iâm thinking about Mercury. Itâs processing. I have my best ideas when Iâm on a flight or driving somewhere and Iâm like, âOh, why donât we think about...â Youâre grinding away, but actually sometimes these decisions or these ideas are actually worth 10x more than an extra few hours of work. Thatâs how a different fork can change everything. So that would be my counter to it, but at the same time, I think itâs fine. If people want to do it, I am supportive. I think itâs reasonable to say if youâre young, maybe that extra 10% is the difference and you donât have the money to go hire an extra person. I think thereâs lots of reasons why itâs completely reasonable and Iâm not here dictating on what people should do. I donât think itâs just obvious that thatâs the only way to do it.
Nakul Mandan:
The other aspect where I think you guys are slightly different is you still are a remote-first culture and Silicon Valley dabbled with remote-first. Actually during COVID years, it was like, âHey, talent is everywhere. Opportunity is not. Letâs get that talent.â But it bounced back pretty aggressively towards everyone in-person. What do you think of that dichotomy around remote-first versus everyone in-person? Because I do think you guys are somewhat different in this in 2026.
Immad Akhund:
I think when you are small and pre-product market fit, I would still recommend being in-person, and we were in-person until we were about 30 people and COVID hit. So from end of 2017 to 2020, we were in-person. I think later on thereâs a lot of advantages that come from being remote and the actual in-person advantages kind of die down. We have 1,200 people. You canât even fit them in one floor. So multiple floors. Probably most companies our scale even that in-person would have separate offices, et cetera. The really big advantage I find is talent. Often Iâm hiring an exec and Iâm like, âOh, wouldnât it be nice if this person was in San Francisco and I could see them often?â But I find someone amazing in New Jersey and Iâm like, âOkay, am I going to get this amazing person? Is this amazing person going to uproot their whole life to come to SF? Probably not.â And it ends up being a differentiator. If thereâs other companies that will not hire that person because theyâre in New Jersey and theyâre not there, then I can.
So I think the talent is definitely better and is definitely differentiated. I do think thereâs some downside, but it just requires more work. We create at the minimum three opportunities for people to meet in person every year. So we have an all hands where we bring the whole company together and each team gets two, and then on the leadership team, thereâs probably five or six times we meet in person and we try to make those very intense in-person moments. So when youâre in office, you can go days not talking to people, or whatever, whereas when we get together, itâs like youâre doing social events together, youâre working together and you do need that in-person touchpoint to create these kind of bonds and create connections and create ideas.
So yeah, I think the other pro on remote is, I think it is a little cheaper, especially in San Francisco. The cost of livingâs really high, the expectations are very high. Youâre competing against every whatever, and by now Anthropic, OpenAI, whatever it is. Yeah, I think on the talent side, itâs just a lot better, but it does require a lot of deliberate effort.
Nakul Mandan:
Yeah, on collaboration. I want to go towards your philosophy on recruiting and teams, but letâs go to the beginning. How did you hire your first 10 to 20 people?
Immad Akhund:
I think the super early stuff is really hard to hire. I remember after... So we raised a round from Andreessen Horowitz. So we were six million at a 23 million valuation, which-
Nakul Mandan:
At that time it was just you, or?
Immad Akhund:
It was me and my two co-founders who Iâd met at my previous company. Yeah, this is 2017. So that was a unusually large and successful seed fund, especially since we were super early. We hadnât ran much code and there was no traction or anything. And then I went to my friends that worked at Facebook and Google and all this stuff and there was nothing I could do to convince them to move. Because the problem is if you work at these big companies... If I was a seed stage startup now and I was trying to hire from Mercury, it would just be very hard because Mercury, the stock is relatively liquid and worth a lot and they pay a lot of money, whereas even if you raise $5 million or whatever, itâs very hard to sell them.
So I found the two main things we could do was like, A, go to people I knew. So actually five of the initial eight people, which is what we mostly built the initial product with, worked with me at my previous company and all of them still work at Mercury. And then the other thing is young talent, people who arenât proven. Itâs just much easier to even go to Stanford or even a smaller college or whatever and you hire someone straight from college. Those people arenât necessarily getting a chance at these kind of mega companies, but thereâs some great talent there and you just have to be really thoughtful by identifying it. Thereâs also another set of people are just very into small companies and those people tend to be pretty good as well.
Nakul Mandan:
How much of your first year did you spend on recruiting?
Immad Akhund:
One thing I always tell entrepreneurs is like, if you want to recruit, the only way to do it is if you make your number one priority, especially as a small company. No one else is going to do it for you.
Nakul Mandan:
Is there any other way? When you say if you want to recruit, every founder has to recruit. It has to be that way, right?
Immad Akhund:
I like phases when Iâm early on. So we raised that six million, I recruited up to a team of eight and then I stopped, and then we executed for a year and a half. I think we hired one other person then, but until we launched, we didnât hire anyone else.
Nakul Mandan:
So for those phases, itâs the number one priority for you.
Immad Akhund:
Yeah. It doesnât mean you donât do anything else, but I find recruiting a little boring, to be honest. Itâs like programming or sales is a little more fun, whereas recruiting is a bit of a grind for me, but itâs necessary. And Iâm always like, if I spend one moment of my time not recruiting... To go back to that previous thing where I said if you go from five people to six people, you become 20% more productive. Especially when youâre a small team, itâs very hard to beat the productivity youâll get from another good person. Why spend time doing something myself when I could hire that incremental person? Obviously at some point you shouldnât hire, which we could talk about when to stop. But yeah, you wake up in the morning, youâre like, âOkay, recruitingâs my number one priority. So whatâs there to do?â Job postings, trying to get referrals, emailing people that you know. Itâs a grind. You just got to wake up every morning, and hitting up people on LinkedIn, whatever it is, you just have to do it.
Nakul Mandan:
But during that phase, when you say itâs your number one priority, how much approximately your time is spending? 50%, 30%?
Immad Akhund:
About 50%.
Nakul Mandan:
Above 50%.
Immad Akhund:
Yeah. I mean, if you include interviews, if you go multiple roles, itâs very time-consuming.
Nakul Mandan:
For the first 10, 20 people, were you looking for something specific in terms of one or two traits that really mattered that might matter slightly less today, but those 20 hires, was there something like that?
Immad Akhund:
I wouldnât say itâs like the things that I donât look for now. And often people kind of glorify those early people at startups a little bit. I actually think we get better people now than we did early because people know Mercury and we can hire really amazing people, whereas it was hard hiring great people early on. Yeah. I mean, I think thereâs some attributes that are really essential when youâre early, but I would say theyâre still essential now. You want people that have really strong ownership mindset. Iâm always like, âHey, itâs not like Iâm not going to tell you what to do. You come here, you figure it out and you get it done kind of thing.â So you want people that have that kind of ownership mindset. I still want those people. I donât want people who are waiting to be told what to do. And then you want people that are very product-minded, that really care, and again, we still want that.
I think one thing that does happen over time is you end up having more specialists. If you want to hire someone in AML investigations, thatâs pretty specialist and we go hire people with a lot of experience to do that. Early on, you want much more generalists that can slot into different things very easily.
Nakul Mandan:
Mercury famously has this curiosity interview. You give somebody an hour to present a non-work topic to their team or the team they would be joining. What are you testing there?
Immad Akhund:
So I came up with this idea and itâs probably been 12 years or something since I came up with it, but basically Iâm an engineer and then I had to hire a sales team at my previous company and I was like, âThese salespeople are just so good at talking.â You come up with any fricking interview questions for salespeople and youâre like, âHow would you run a pipeline?â And they have these really beautiful answers to these things. So with engineering, I was like, âOkay, you just give them a coding challenge. Thatâs how you test if someoneâs good at engineering.â So I was like, âWhat is the coding challenge equivalent for sales?â And then I came up with this answer that I was like, âHey, present me a topic of your choice, spend 45 minutes preparing a presentation and you do it and weâll talk about it.â And it was an amazing filter for salespeople.
Iâd have these conversations with them, they were amazing. Iâd be like, âOh, this is the best salesperson Iâve ever spoken to,â and theyâd come in on these presentation things and they could not talk about anything interesting. And if you ask them questions, they wouldnât have thoughtful responses. And it turned out to be an amazing filter for salespeople. So since then I basically kept... I was like, âOkay, it works for salespeople. Why donât we do it for basically every role?â So we donât do it for engineers generally. Things that have a coding challenge or designers where you can really test a work sample and see what itâs like, I think itâs less useful, and honestly, the times Iâve done it with engineers, theyâre just too good at it anyway, because engineers always have some quirky thing that they know way too well. But for most other levels and roles... We do it a lot for exec hiring. Itâs amazing filter for that as well.
Nakul Mandan:
But what are you getting there? Is it that this person is multidimensional or that thereâs standard interview topics they know how to ace?
Immad Akhund:
So thereâs a bunch of elements and we have a whole grading score for it. So itâs not that you just sit back and listen to them. Itâs a much more conversational kind of thing. There is a basic thing, like are they interested in some interesting things and have they gone deep on those subjects? And most good people who are going to do a good job and at least my experience with people who are great in their job, they have weird things that theyâre really interested in and they know a lot about. Thatâs almost universal. I mean, obviously weâre biased with Mercury, but you probably have some weird thing that you know way too much about. I think it shows youâre curious, youâre interested, you want to learn things, and itâs the types of thing. In a startup, you need people who are curious, they want to learn things because thereâs always going to be something new and thereâs always going to be something.
So that is a strong filter. Thereâs also like, how do they react to... We try to intrigue them and push back and some people get real defensive about it. We want to have a culture where itâs like back and forth and we challenge you on ideas and say like... And then thereâs this kind of general product. If you think about the abstract of what is product thinking and product thinking is this being able to see a problem, come up with answers and solutions to it. So we have some classic questions that we do as counters. So weâd be like... If youâre really interested in, I donât know, pickleball, weâll be like, âHey, letâs say youâre in charge of making pickleball really popular in America, what would you do?â And itâs not quite build a product, but itâs kind of product thinking. And some people are just like, âOh, yeah. Iâd go do this and that in schools.â They have five or six things that come up with quickly. Other people just completely struggle and go blank on that kind of stuff and I guess we want the people that are good at it.
So we actually literally look at it in terms of every single cultural value we have and how do we have one question that tests against all of that? Itâs actually one of the best ways to tell ego as well. People who are high ego, they cannot handle pushback or any sort of challenge to an idea, whereas people are humble or much more like, âOh, yeah. Thatâs really interesting,â and they enjoy that kind of pushback.
Nakul Mandan:
Has there ever been an exception made in the sense that somebody who was acing their sales scores, whatever, the interview, their past background, on the curiosity interview, they did okay, maybe not bombed it because-
Immad Akhund:
Not for sales. I would never make an exception for sales. This is your core job. The presentation interview is 100% correlated with a sales job. We have made exceptions for some other things. I donât love making an exception ever really, but sometimes itâs like a very specialist skillset and this personâs got a ton of background on it, and yeah, maybe theyâre not a customer-facing role and they donât have to be that product-facing and weâve made exceptions. I donât love it. I personally would never make an exception, but I know some people in the team to have.
Nakul Mandan:
How are you ensuring that as the company has grown to 1200 people that the recruiting bar is still high? Because the first, I donât know... Till what time were you meeting almost every person?
Immad Akhund:
Yeah, I think around 180 I met everyone, and then since then you basically have to take the things that you did and try to productionize them. I mean, I donât want to keep talking about this presentation interview, but we set up a training course so other people can conduct the interview and we see how they do it and we score them on their ability to score the interview and then we take the different things weâre measuring and we turn that into a rubric that we judge in that interview process. So across all the interviews, we try to hold a bar in that sense. And then thereâs also performance managing on the other side. I think no interview process is perfect. Youâre trying to make a judgment after a few hours. So you just have to hold a bar on the other side as well and say, âOkay, they came in, how are they doing?â
Nakul Mandan:
Do you guys do a lot of back channel references? Is that a critical part of your process or not?
Immad Akhund:
Definitely for leaders and execs and things like that. Sometimes we end up doing it for lower level things. I donât find them super useful outside execs, but yeah, I donât think itâs necessarily critical.
Nakul Mandan:
Have any of your beliefs on recruiting and team building or even how teams are run evolved as Mercuryâs scales goes up? Youâve been an entrepreneur for a long time, but this is-
Immad Akhund:
Oh, yeah. By far. My previous company was 20 when we sold. One thing that was a complete game changer to me, which maybe itâs obvious to everyone, but it definitely wasnât to me, is after we raised the series A... Our series A was like 20 million raised. Someone, Josh from Gusto, he said, âHey, maybe you should get a recruiter,â because I was asking him for tips on hiring now that we had like... I was like, âOh, a recruiter.â And Iâd kind of had experiences with external contingency recruiters, which I basically never had a good experience with. So then we ended up hiring an internal recruiter and that was completely game-changing because we went from me and someone else in the team trying to directly find people from my network, et cetera, to someone whoâs just dedicated 100% of the time to finding the best talent. Complete game changer. Now we have a pretty big recruiting team of, I think itâs 15-ish.
Having a great recruiting team... And actually one thing we ended up doing is that first recruiter became our VP of recruiting, but she was really strong at sussing out culture. So often... And maybe I shouldnât say this on camera, but often people are kind of rude or dismissive of recruiters because I guess theyâre used to not taking them seriously, but thatâs the first touchpoint of the company. And if someoneâs rude or dismissive or whatever to the recruiter, thatâs a super bad sign. I guess in general, people tend to be very honest with recruiters and who they really are, whereas theyâre sometimes retaining. So it actually ends up being a really strong cultural filter for us.
But yeah, having an internal recruiter is a complete game changer. I often talk to people who are series A, sometimes in series B companies and they still havenât thought of this and always told them this. So if youâre working at all with external contingency recruiters and youâre hiring two people a month or something like that, itâs a no-brainer. Youâre going to save money and itâs going to be better hires.
Nakul Mandan:
It does seem to work better when the recruiter is treated even internally as a true partner by the VP of [inaudible 00:28:29], VP of sales and often it can be-
Immad Akhund:
There is literally no role at the company that works better if you donât treat them as a true internal partner, whether itâs compliance or sales or recruiting or whatever. I think one of the biggest mistakes people make at companies is not... If youâre going to have a role, you should make them a core part of what makes you successful, otherwise youâre wasting an opportunity.
Nakul Mandan:
Letâs walk through what does your week look like at todayâs scale, week as a CEO. So whatâs your Monday like and how much time are you spending on product, sales, marketing today through the week?
Immad Akhund:
Well, one thing I should plug is I got rid of one-on-ones. I actually listened a year and a half or so ago to Jensen Huang and he runs this company super funny, but heâs got like 60 direct reports he said. But one of the things he said is he got rid of one-on-ones. And up until that point, I was running one-on-ones with everyone that reports to me and a few other people and I was like, âWow. I donât find those meetings overly useful.â And then I talked to a bunch of other CEOs and theyâd also got rid of one-on-ones, not universally, but it was actually much more common than I thought.
So if you have mostly leaders and execs reporting to you, I think you should probably get rid of one-on-ones. So thatâs something Iâll go out there and say. But thatâs freed up a lot of my time and I actually ended up then increasing the number of direct reports I had because one-on-ones are just like... They take up a lot of time. Imagine if you have 10 reports and youâre doing weekly one-on-ones, thatâs one day wiped out, but they also take a lot of headspace as well. So anyway, I replaced all one-on-ones with group team meeting things with execs. So I have a series of things that we have a weekly all-exec meeting, then I have a biweekly, I think, GTM meeting, EPD execs, and then compliance and risk, legal, and then GNA. So exec meetings are on set of things we do and we try to get these into not status updates. I canât stand status updates. Itâs all like letâs go into meeting discussion topics around whateverâs the most important aspect of that team.
So thatâs a reasonable chunk of my meeting. On Fridays we do what I call exec workshop where itâs three hours long or something. Itâs kind of like an office hours where I will either go to a team and say like, âHey, I want you to come in and talk about what youâre working on for 30 minutes,â or a team will come to us and thereâs a channel, Slack channel for it, where theyâll say, âThereâs a new product launching, we want to show you what weâre working on, et cetera.â And all the execs are there, but itâs often a venue for me to see whatâs happening and opine on it and for teams to get decisions made at the exec level. Then a lot of the rest of it is Iâm still recruiting, especially if itâs execs, thereâs a lot of... Like this morning, I mean, today I think Iâve done three interviews so far. So yeah, recruiting is a big part of the rest of it.
Nakul Mandan:
How much of your attention is still on product and product innovation for Mercury?
Immad Akhund:
Yeah, in that exec product workshop, thereâs a ton of it. And then we do a ton asynchronously. So I instituted this two channels I canât remember how long ago, two, three years ago, but we call it the pre-shipped channel and the shipped channel. And basically anything thatâs often design-ready, like initial designs or sometimes itâs just like spec mode, goes to the pre-ship channel. Everyoneâs part of it, but it gives me a... Yeah, I can see what everyoneâs working on basically. And then the shipped channel is when it goes to production. So these channels are super busy because theyâre shipping things all the time and that lets me quite quickly see, and then if Iâm interested in something or have questions, Iâll either tell them to come to the exec workshop or Iâll just reach out to them. Thatâs my main direct involvement with product.
The other place where often comes up is if youâre launching something big, Iâll go test product out and give feedback. And then the other big thing is if there is new things we are working on... I have a lot of ideas. So often, but not always, often the ideas for a new product and weâve in the last, I guess three, four years, weâve gone very multi-product. So weâve got the core banking, but we also have credit cards and invoicing and bill pay, and we also just acquired a payroll company and thereâs a couple of other things that weâre working on. But those ideas often come from me and then thereâs a kind of process of me pitching them and seeing if an existing team can do it or instantiating a new team and then hiring against it, et cetera, that is also often top-down.
Nakul Mandan:
But do you think of your week as in different phases, Iâm guessing similarly like, âHey, Iâm attacking this massive problem that I need to solve,â or it kind of is a steady marketing product, sales, recruiting dial?
Immad Akhund:
Definitely thereâs always some big problem and I end up devoting quite a lot of time to it. So at the start of this year, I kind of took over as a CTO. Weâre doing a CTO search. If anyone goods out there, ping me. But as taking over for CTO, I ended up reorging the team a little bit and going in there and saying like, âOkay, what are the big problems? What can I make some quick wins with?â And so that was very time-consuming. Iâd say still pretty time-consuming. Iâm very involved in the engineering day-to-day stuff in a way that I wasnât before I took over. But this ends up happening every year that thereâs some department where I take over because someone leaves or theyâre not working out or we need to do a change. So I end up learning about all sorts of things in doing this. I think a big part of how I see my role, thereâs the steady state stuff, but then thereâs the chief problem solver/chief opportunity finder kind of thing and then you go very deep on that.
Nakul Mandan:
Actually, one topic Iâd love to get your take on is how do you know an exec is no longer the right exec? So beyond the initial culture for it where somebody came in, didnât work out, put that aside, somebody was great for two, three years, but they might not be good for the next two, three years. What are the early signals somebodyâs plateaued for the next two, three years?
Immad Akhund:
Iâd say honestly, it becomes pretty obvious by the time you make that decision and youâre normally like, âWow. I should have decided a while ago.â Hereâs some factors. Do they hire a really strong bench underneath them? I find that the best execs, you look one layer below them and youâre like, âWow. These people could do your job.â I think that not everyone, but they normally have two people that at least or maybe three people that youâre like, âWow. These are amazing leaders of the company.â Obviously you canât expect that day zero from an exec, but an exec thatâs done a really good job is both hired/facilitated really strong talent just below them.
Number two, I think the higher up a company you go in the chain, the less excuses you can have. Me as a CEO, I have no excuse. If somethingâs going wrong, itâs my fault. But I think itâs the same, one level lower. To go back to the, are you driving impact? I donât care how good your reports are. You could do these amazing presentations. I donât care. And generally every exec has very strong things that theyâre aiming for, whether itâs launching products or driving sales or whatever the metrics are that you would judge that exec on. And yeah, I donât think execs can have excuses. I think thatâs like the job is to deliver on that. So Iâd say those, two talent and delivery, are probably the two most important kind of measures where people start failing. And talent is often one where you can see people not scaling. They donât think ahead, they donât think what is needed, but same with outcomes. Theyâre not necessarily driving whatâs a 10x outcome. The shape of almost every org needs to be different, 10x bigger.
Nakul Mandan:
If youâre not doing one-on-ones, is there a risk of not catching that plateauing of that exec early on enough?
Immad Akhund:
I mean, Iâve never found one-on-ones help. One-on-ones is just an avenue for them to try to sell themselves. You need to create avenues where you have visibility into what that team is doing. So I guess I didnât talk about this, but most of our teams do a quarterly review kind of thing where they write a... We do docs, we donât do any presidency. Write a doc about whatâs your team doing well, the deliveries, what are you going to do next time, do some deep dives kind of stuff. So you get a sense from that. I mean, if itâs a core thing, like itâs product or sales or whatever, youâre seeing the metrics all the time. Are we delivering? Are we behind on deadlines? Are we driving enough sales? Yeah, I find that itâs my job to have that visibility, but not doing one-on-ones does not mean Iâm not interacting with them all the time.
Whatever the role is, there are tons of meetings that Iâm part of where Iâm seeing them perform, whether itâs the exec meetings or whether itâs the group meetings or whether itâs product meetings. Thereâs just so many places I see, especially the people that report to me, I see them. All of them, Iâll see at least once a week, if not multiple times. So yeah, itâs up to me to understand what theyâre doing and see if theyâre delivering on it. Yeah, Iâve never found one-on-ones with the way to do it anyway.
Nakul Mandan:
And then one of your operating principles is that I think the fastest or the most functional team is a 10% team. So you guys have kind of recreated that with an eight person product team, one PM, one designer, six engineers or something like that.
Immad Akhund:
Yeah.
Nakul Mandan:
How many such teams actually exist at Mercury today?
Immad Akhund:
Yeah. To some extent, maybe this is changing with AI, but this comes from my startup routes. Iâve always thought, wow, startups just seem way more efficient than a big company. And if you look at it... Mercury has 400 engineers, so in theory, we should produce 40 times more than a seed-stage startup. Obviously we have to maintain a base and blah, blah, blah. So then you have to go like, âWhat is it that makes these small teams more efficient?â I think thereâs small teams that have, I call them autonomous product teams, they have to have a overall remit where they can independently work where youâre like, âHey, you own this thing. This is your KPI that you own. This is your product surface area. Go talk to customers, go figure out your roadmap.â And maybe thereâll be some top-down, but that team is kind of driving it and owning it and feeling the responsibility for it. I think that is an ideal team. We have so many of these at this point. Letâs see, one, two times... At least 15.
Nakul Mandan:
And are these truly autonomous teams now? Where do they need your approval on product decisions?
Immad Akhund:
My approval, very little. The bit where theyâre not autonomous is when they need to work with each other and to do big things. We have one team thatâs on our Mercury personal product, which is our personal banker product. But yeah, the reality is sometimes they need to work with our risk disputes team or they need to work with our cards team and then thereâs some collaboration and for some teams they can really be independent. We have a invoicing team and theyâre failing, but even then they have to work with a mobile team. But anyway. There are overlaps and thatâs where they canât be fully autonomous, but they all have a real ownership or on a end result that theyâre driving at. And a lot of their roadmap is set internally.
Thereâs bits where at the end of last year we did go and say like, âHey, these are the big, big things we want to launch in 2026,â and in November last year, we set out a six item product calendar for this year, which weâre still working on. So for those teams, we were really a little more top-down and we were like, âOkay, this is a big initiative. What does it take to get this successful?â But even then somewhat the top-down is like, âWe are going to launch this.â We just launched an insights product, which helps you give insights on your... But how exactly it works, what are all the features is fairly up to the team to kind of...
Nakul Mandan:
Are they all reporting into a head of product or?
Immad Akhund:
Yeah. So we have two VPs of product. One is banking and risk, and the other is experiences, and they report to my co-founder whoâs kind of head of product and design.
Nakul Mandan:
You also said something around itâs dangerous if companies decisions are only made on metrics, and I think as an extension, I think you were talking about taste versus metrics. Correct me if Iâm wrong on that, but how does that play out in your mind? Why did you say that? And then I would love to dive deeper into that once...
Immad Akhund:
So I think thereâs some teams that are very metrics-driven and have to be. So a lot of our growth teams, youâre trying to drive conversion from sign up to application. And thatâs a complicated process because weâre a bank, but thereâs so many steps in that you can optimize and mostly you should be driven by metrics when youâre optimizing them. I think if thatâs all you do, youâre going to end up with these kind of local maximas of like, âOh, yeah. We optimized that and we increased conversion rate on that.â And I see some companies do that. I think to some extent Facebook, the app, Facebook lost their heart because they become so metrics-driven that theyâd never actually thought about the customer. I do think the other set of things are like itâs just harder to measure metrics on a lot of improvements to customer experience.
So we collect W9s. So if you want to pay a contractor, youâre supposed to collect tax information. And yeah, what are the metric we were optimizing for that? I mean, customers like it. It makes a huge difference to peopleâs lives because theyâd have to collect these over email and type them in. Itâs definitely better, but I just donât know what metric you would drive for that. I mean, obviously you want to launch that feature and people use it, but thereâs no overall company-level metric. I mean, eventually itâll drive retention and maybe customers will tell each other about it and then thereâll be more. Itâs not even about taste. I think building an amazing product that helps people, thereâs just not that much thatâs just pure letâs crank the wheel on optimizing metrics. So thatâs my main point. At the end of the day, the company needs to drive EBITDA. We need to drive profit and thatâs the only metric that really matters, but how you get there is a kind of...
Nakul Mandan:
Yeah, it kind of speaks to you and Mercury are both very product-led. I mean, Mercury is an extension of you. So youâre saying product, it has to be product-led, not growths-led.
Immad Akhund:
Yeah. Itâs like, how do we help customers? How do we solve their pains? A lot of that I donât think is a metric that youâre driving.
Nakul Mandan:
But then again, as youâve scaled to such a point, how have you thought about your product taste and product intuition scaling with the company? Because now itâs like Mercury started with a beloved product. Obviously it was beloved. Thatâs why it got product market fit. Now youâre so many products and youâre expanding that portfolio.
Immad Akhund:
I mean, firstly, itâs not my product taste most of the time. I think if it comes to me and Iâm the reason we... Obviously sometimes it is mine. Sometimes I see things that someone thought about at Mercury. Iâm like, âWow. I didnât even think we could do that,â or, âThat was a great idea.â And we encourage a very close collaboration between engineering, PM and design to come up with products and talk to customers. Thereâs no way thatâs going to come from me. Iâm not in the weeds on a very... Sometimes I am, but most of the time the designer working on the product has thought about it way more than I have.
So the main thing I do is when that slips up, Iâm going to complain a lot, not like in a, âOh, this team sucks.â But Iâll go, and not always, but I try to go pretty deep into products we ship. Obviously Iâm the user on Mercury, but I have my own fund on Mercury, my wifeâs business on Mercury. So I just test these things out and Iâm like, âOkay, is this good or not?â And then I give that feedback to the team. So in that sense, I think actually complaining is a pretty good skillset for... You have to try to put yourself in a mindset of like, I donât know anything about Mercury, but what are all the things that just would bug me, if I turned up at this product, Iâd be so annoyed that I have to click this button or this thing doesnât make any sense?
So I try to be this kind of mindset of I donât want to think too much. I just want to have an amazing experience. And when it doesnât feel like that, I can manifest that user, have a lot of sympathy for what that user would be. That I think is an important skill to have. But yeah, the product thinking and delivering a great product, you canât be doing that top-down. The team has to be driving that.
Nakul Mandan:
A similar question on scale. How do you ensure it remains a very performance-driven organization at this scale? Itâs just the quarterly reviews or trusting the [inaudible 00:44:56]?
Immad Akhund:
I guess it depends what you mean by performance-driven. What do you mean by performance-driven? Because thereâs a lot of elements to that, right? Thereâs like, are you driving revenue? Are you launching products and things like that?
Nakul Mandan:
I think itâs different for different teams. My question was more oriented towards... So obviously for some metric-driven teams like growth, marketing, sales, itâs much more metric-driven, but on product, on design, customer support, thereâs an intentionality and obsession with the outcome or impact. How are you ensuring that thatâs just percolating through the organization even as the company scales? Because stasis sets in with scale over time or apathy.
Immad Akhund:
Yeah, thatâs a good point. I think every team and concept at Mercury deserves its own kind of thought on how you do it. If you think about, letâs say customer support, you can be putting metrics-driven there. Youâre like, âOkay, what is the customer satisfaction? How many tickets are being processed? Are we reducing our meantime to get it to a ticket?â At the engineering, you also have metrics. Itâs like per request per week, or are you shipping on time to the calendar that you had set kind of thing. So I think thereâs the metrics side of it. I think second part of it is culture. Do people love what theyâre doing? Are they enjoying... When youâre small, again, people do things because they do things. Itâs not like thereâs an external, you have to perform in order... I think if people are liking what theyâre doing, theyâre liking the team, they understand why theyâre doing it and how theyâre helping customers. People want to ship things and they want to do things, I think at least the types of people we try to hire.
And then you do no matter what, have to have a bar where youâre willing to fire people that are not performing. I think if you donât hold that over time A player team gets a few B or C players and then theyâre like, âOkay, these people arenât working,â or they keep having to fill in for them and that can deteriorate the culture of the team. So you have to have that bar and you have to continuously be enforcing it.
Nakul Mandan:
Letâs talk about how AI is changing how Mercury is run. First of all, actually, how much of your own cadence has changed with AI? How deeply embedded is AI in your decision making or visibility into the organization?
Immad Akhund:
I use it a lot just for structuring my thoughts, thinking through problems, definitely seeing what everyoneâs up to, and then I try to get... I wrote some code actually, did some Claude Code stuff. I mean, Iâve done some personal projects, but I wrote some production code, which I hadnât done since 2021 actually, so that was kind of fun. But at the CEO level, itâs not necessarily been as drastic as other places. I mean, I have a literal Claude Coworker thing that helps me prepare for meetings and things like that, but I wouldnât say itâs been complete game changer. Obviously engineering team, itâs been a complete game changer, especially for new product development. I think existing, maybe some of our core banking core and payment stuff, it helps, but itâs not been like 10x. But yeah, if youâre doing something completely new greenfield, itâs amazing.
Weâve done a bunch of stuff to help encourage that. We have an AI enablement team that is working on tooling for the engineers, but also for the broader organization. So super useful there. And then on the rest of the teams, we encourage a lot of usage and then... Yeah. So whether itâs our finance team or our legal team... Our legal teamâs actually the biggest users of AI. We did a vote recently. Itâs like 80% of them say that itâs saved them 20 hours a week or something like that. Itâs been an insane use.
Nakul Mandan:
Have you guys adopted things like Decagon on customer support or some of these things?
Immad Akhund:
Yeah. So on the customer support side, itâs been really good. About 35% of our tickets are responded straight away. Weâre using Fin right now. Over time, weâre probably going to bring it completely in-house just because the next level of tickets, if we want to get it to 70%, you kind of have to go integrate into internal systems and at that point it might as well be internal. Yeah, super useful there. And then weâre launching a bunch of AI-related features where we launched a... So on part of it is connecting into AI. So we have a Mercury MCP and a Mercury CLI. So if youâre using Claud Code or Coworker or Codex or whatever it is, you can plug into Mercury and have workflows done.
Nakul Mandan:
As a customer?
Immad Akhund:
As a customer. The usage of that is going crazy. Thereâs lots of interesting things people are doing with it. You can analyze your finances or you can see all the pending payment requests and tie them into some of the workflow. You can generate invoices. Itâs like all the things you can do in Mercury and we keep adding more and more of those features into it. So thatâs been big. The second thing is we are adding more AI features directly into Mercury. So we launched this Mercury Insights thing where you can talk to your transactions, summarize them for you like, âWhy did my AWS bill change in the last month?â Or whatever it is. Thereâs lots of kind of questions you can ask around that data. And the more of Mercury you use, the better those answers can be. If you use us for credit card and banking and invoicing, you can ask a lot more.
And then weâre actually launching another thing later this year, itâs called Mercury Command, where itâll basically let you do a bunch of the workflows that you can do within Mercury directly from AI. Instead of having to learn all the commands or string together different actions, you can just be like, âHey, I need to pay my landlord 5K for the month,â and theyâll try to figure out what you mean and pop up something. Every action still needs affirmative action. We donât let the AI do the action, but we can basically automate a ton of your workflows directly from Mercury.
Nakul Mandan:
Have you seen your sales and marketing orgs also embrace AI very aggressively or not yet?
Immad Akhund:
Yes, but we went through this first phase and I think itâs probably got a lot better where we went and adopted some AI SDRs and it kind of sucked. And actually I think it probably still sucks. At least the AI STRs flood the field. So then the signal-to-noise ratio is way off. So you have to be different in order to stand out. But what we have found is enabling the salespeople with the right thing to say is a really powerful thing. Itâs like, who should I speak to? Someone just raised some money. Itâs like as youâre sending the email, itâs like, âOh, yeah. Congrats on the one million raised,â dot-dot-dot, or they just hired the first finance person. Thatâs a great point for them when theyâre considering new finance deals. So we find at least right now, and I think itâll change over time as AI and the sales team has been more useful as an enabler to internal people rather than a replacement. But yeah, we are hiring pretty aggressively on the sales side. So itâs definitely not freed up any headcount.
Nakul Mandan:
Has your recruiting criteria across various business functions changed towards testing for AI fluency or nativeness or not yet?
Immad Akhund:
Definitely on the engineering side, I think itâs almost a given that you need to be very fluent with it, and at this point, if youâre not, itâll be surprising. I wouldnât say itâs changed too much outside the EPD function, the engineering product design. I guess data also has become a much more AI field. So Iâd say case by case on the other ones. I donât know if the finance person needs to.
Nakul Mandan:
And you touched on some of the product things that are coming out, but how has AI changed taking a step back Mercuryâs long-term vision?
Immad Akhund:
I think the interesting thing about Mercury, and maybe this is true for other startups as well, I feel like if youâre on the right side of where historyâs going, then almost everything helps you. At least thatâs what Iâve noticed with Mercury, which for us, the right side is banking is going to be disrupted by FinTech and technology companies are going to really be the source of that disruption. Some of the stuff is obvious. Banking is going to be online and digital. Itâs going to be mobile apps and websites rather than bank branches. And one of the original thesis for Mercury was always like banking can do a lot more. Itâs not just the bank account, itâs going to be the whole financial suite. So everything thatâs happened... When COVID happened, that was a big boon for us because bank branches were closed and you had to come. AI is another one where itâs now the way you interact with banking and the tools you expect from banking and all the places you can use it in the back office is completely shifted.
Nakul Mandan:
Yeah, I highly doubt Bank of America as an MCP.
Immad Akhund:
Yeah. We are extremely well-situated for that because we are a technology company and thatâs just what we do, whereas what would be a threat to anyone else is a big boon for us because now itâs like our differentiator is now even further. It was already great, but now if you want to be able to connect it to your MCP, thereâs just no one else thatâs going to do that for you. Yeah, thatâs my high level take on it. Yeah, I think it means that the things that we do have changed and a lot of our focus has changed around it, but at the same time, itâs mostly accelerated the overall vision, which is still like, how do we deliver great products to our customers? How do we speed up their life and make it easier? How do we give them great insights? All of these things were in my seeds pitch. This is how I always describe Mercury. Itâs just like, now weâre going to do it with AI as part of that pitch.
Nakul Mandan:
You also recently got the approval to establish Mercury Bank, but your interface has always been a banking interface. So does that change much for the customer or is this more about the backend operations and maybe some downstream products you can unleash because of that, but the banking interface or the Mercury interface remains the same?
Immad Akhund:
So the interface will be the same. It will enable a few extra features. So weâll have Zelle support, which isnât something we have right now. Weâll have a lot more lending products that we can offer our customers, which to this day we have a couple, but we really minimize lending. So thatâs one thing. Thereâll be lots of smaller changes that maybe people wonât even notice, but we will have quicker money movement. Itâs just like if you control every step of it and we are trying to... When we do the Mercury bank integrations, we are going directly to the Fed and because we control every step of it, give you the biggest cutoff windows and we could just guarantee a lot more and those things will... We care about those little things even if customers sometimes donât notice. Economically itâll also be better for us.
But I think overall itâs also what makes sense for the long term for us, which is right now if you go to mercury.com, thereâs a lot of disclaimers on there because we canât call ourselves a bank and we have to make it very clear to our customers where theyâre actually banking. And that doesnât always engender the most trust from our customers. If you donât know about Mercury and you show up there, youâre like, âWhat is this? Iâm not getting a bank here.â It creates confusion.
I genuinely think if Mercury was a bank in a chartered sense, I donât know why someone would not use Mercury. Yeah, at this point I can say, okay, some people used to have some reservations about using a FinTech that has this pass through bank relationship, but yeah, I think itâll go a long way for trust. And trust has always been the biggest thing to get over with Mercury because I think our productâs always been the best, but do you trust this tiny startup? Thatâs what we used to be. Now weâre a big startup. But itâs a big trust thing. A bank account is not a low-trust service. Thatâs like something thatâs going to store your money and you have to really trust it.
Nakul Mandan:
On that note, did the SVB crisis have a big impact on you guys? Because trust at that time was fleeing towards the big banks and a lot of people were moving because whoâs too big to fail, better to park your money there? Did that month or so...
Immad Akhund:
It was actually net very beneficial to us. We gained two billion in deposits in that two week period and then after that point, our... Our graph went from... It was a step function change, but every month after that, we were acquiring 80% more customers and it grew at a higher gradient. And I think the reason that worked is we tried to say why we were different. We extended our FDIC insurance to five million within that week and we tried to say like, âOkay, donât take my word for it, your money is in these kind of safe places.â And I think that people really appreciated that. Obviously they were our main competitors, so they were still around, but people stopped choosing them as their primary, so they went to us.
I think in general, there was definitely some people that their lesson was like, âNow we need to go to a too big to fail bank.â But then there was other sort of people like, âHey, actually startups can be much more transparent about where the money is and keep it safe and itâs not necessarily...â SVB was supposed to be the too big to fail bank versus Mercury anyway. So I think it ended up being a net confidence boost to us, but that was a big part of it. We had to really deliver on showing why we could be trusted at that moment.
Nakul Mandan:
Would being a bank challenge some of the cultural norms from being a product-led company to, Iâm presuming, but correct me if Iâm wrong, on more compliance requirements, more back office buttoned-up aspects of the organization?
Immad Akhund:
So I think one of the reasons we applied for being a bank is the reality is at the scale weâre at, we are not exempt from regulations, we are regulated by proxy. If we have a new product launch, we have to go run it by our partner banks and we have to run every marketing update by compliance. So all these things are already true. Actually though, itâs almost worse because our programs instead of being dictated by us are dictated by third-parties. So I mean, thatâs the reality of any at scale FinTech that you are regulated. So thatâs one thing. Number two, our aim, and maybe this is naive of me, but naivety is the reason why we do anything is thereâs just not that many FinTechs at scale that become banks. Maybe two or something.
So I think thereâs a way to do it differently and I think I would be disappointed if we come back five years from now and I did everything that a bank does exactly how a bank does. We are not innovating and weâre not product-led and weâre not customer-centric and all of these things that... The reason we exist is because to some extent the banks failed at doing what we felt needed to be done. So yeah, I want to build a new type of institution and thatâs the opportunity. The contrarian thing is how do we become a regulated bank while building a product-led high velocity kind of company? And if we can achieve that, I think thatâs a huge TAM opportunity and thatâs whatâs fun about it.
Nakul Mandan:
So the other part that is changing about Mercury is that you started with the startup as a customer, but then last year, 73% of your business came from outside of Silicon Valley or outside of the startup ecosystem. Is that leading to change in how youâre thinking about product at all or you think itâs a natural extension for you?
Immad Akhund:
No, thereâs a bunch of changes, some non-obvious ones. For a long time, all of our onboarding was very startup-centric. I mean, a lot of it still is, but weâll ask you, whoâs your VCs? And weâll ask for your LinkedIn and all this kind of stuff. It ended up being like we werenât always collecting the right information and being able to onboard a normal SMB. So we did a bunch of changes actually start off last year. We were like, âOkay. Yes, it should work great for startups, but letâs also collect the right information and get other types of businesses onboarded quickly.â Then we also ended up making changes to our corporate credit card product where previously if youâre a startup and youâve raised a million dollars, you get a really nice credit limit. But if youâre a small business and you have 30K, youâll not get access to the card. So we made a new underwriting process where even smaller cash balancers will get a card, they just have to pay it back daily.
So lots of these little nuances. The product is horizontal. Itâs like every business can use a banking service. That bits exist, but all the little bits that make it perfect for startups donât always work for another type of business. So especially a cash-constrained business, which most businesses out there are. So the bunch of product changes and then thereâs a bunch of go to-market kind of changes. So on the website for a long time, it just said banking for startups. So if you showed up and youâre like... Even an eCommerce company, you might be like, âOkay. Well, this isnât really for me.â So how we talk about the product, where we market the product, the sales team, the partnership team, a lot of these things, again, it was more addition thing.
We still want to do all the things we do in the startup space and our startup market share has still grown in the last year and a bit, but we also want to go nail these other spaces. And the other spaces are huge. Our growth started accelerating. So last year we grew 50%. Weâre probably going to grow more than 50% this year on customer account, and thatâs just because the startup space, I mean, itâs big in terms of dollars, but itâs like in terms of number of customers, itâs not as big as the small business or SMB space.
Nakul Mandan:
Youâve mentioned you want Mercury to keep flourishing beyond you. What does Mercury need to become in the years to come or the decades ahead, which it is not today for that dream to come true?
Immad Akhund:
I mean, hopefully itâs already at the phase where if I disappeared, thereâs good bones there and it would keep going. I think the thing that a founding CEO can do is more disruptive things. We are going into payroll. Thatâs not an obvious thing for a banking service to do. I think this kind of reinvention where youâre willing to do new things and push yourself. Mercury Personal, us expanding from business to personal, I had a lot of pushback internally and I think most people now like it internally, but it was just not an obvious thing to do, and I just was like, âHey, weâll do it better. Itâs worth it, and these are my reasons.â But organizations are probably for reasonable reasons, donât like change and donât like new things as much as founders. I love change. I live for change. Change is the main thing I want in life.
I donât know what the answer is there. I think obviously other people have done a CEO transition where theyâve been instrument CEOs that also like that change, theyâre entrepreneurial, and that would be my hope. But Iâm only like 42. Iâm losing count ever... So yeah, thereâs still a lot to do at Mercury, so I guess we should plan for it, but right now Iâm happy to run it.
Nakul Mandan:
So letâs talk about the inner game. Youâve been a three-time founder. How is Immad today different than when you were on your first or second startup in terms of psyche management, how you approach founder life?
Immad Akhund:
Thatâs a good question. I think the biggest thing that you get over time is having confidence in yourself and your ideas. And when you first do a startup, at least I was very susceptible to other peopleâs ideas. I was like, âOh, maybe this is the way you do a startup. Maybe thatâs the way you do a startup.â And over time, A, you learn better ideas, but you also get confidence in your ideas and youâre like, âMaybe Iâll talk to Jensen Huang,â and go like, âOh, yeah. This is a good point. One-on-ones donât work.â But thereâs a lot of other things he said where you have to have your own opinions and apply them. And I think thatâs getting confidence in your own opinions, knowing who you are and going like, âOh, this is the types of people I want to work for, this are the types of products I want to build.â That takes a long time.
I would say by the time I started Mercury, I was 33 and I was probably in a better place to have most of that stuff set. Every year Iâm learning something new, but it doesnât tend to be these core dynamics of who I am and who you are does definitely influence in what type of business you build. Psyche management, it goes back to some of that 996 conversation. It was definitely like I had to learn what is a steady state for me. Something that I relearn over time is I canât spend too much time only doing things I donât enjoy. And thereâs some elements of being a CEO where you just have to go recruit or go solve a problem and itâs not that much fun, whereas I like engineering and product and talking to customers and if I donât give myself enough space to do that, I can definitely feel burnt out. So a lot of the psyche management goes back to burnout management. Itâs like, okay, what is the steady state for you to do this for a long time and have that energy that youâre bringing to the company?
And I definitely talk to some people who... Actually a lot of founders get burnt out. Yeah, thatâs not healthy for them or their startup actually. I think having that management of your own state and knowing what your needs are is important. Thereâs a bunch of elements to that. But yeah, the core of it is sleep well, meditate and exercise and I guess have a good family life.
Nakul Mandan:
Are there founder CEOs that you talk to when youâre stuck on an operating matter or a leadership matter that you go to repeatedly?
Immad Akhund:
It depends on what the matter is, and the tricky part of running a company at scale is you donât have that many peers that you can talk to. So weâve had some really useful investors at Mercury, like Zach from Plaid as an investor and Qasar from Applied is an investor. So that helps. And often if I have like, âIâm hiring a CFO, what should I look for?â That kind of question, itâs really useful. And then thereâs a few communities that Iâm involved with. So the one I joined a year and a half ago, itâs a YPO, but thatâs really useful. Itâs like a very high-performance CEOs in this group. We get together once a month and everyoneâs very honest and authentic in that group. So thatâs a useful place for me to get... Often I find with advice is itâs mostly a good avenue for you to talk honestly about so that you kind of... Obviously thereâs specific advice you get, but you also self-discover advice when you have an avenue for talking about it.
Nakul Mandan:
Has there been an instance recently and maybe the one-on-one Jensen thing where another CEO has been able to change your opinion or how you run the company?
Immad Akhund:
Yeah, there was a recent element where we are applying for this bank charter and I was trying to communicate to the compliance team around how do we want to do compliance in a regulated world, and I was getting stuck into very specific like, oh, how do we do data governance, and getting really in the weeds about it, and someone gave me this advice, I wonât say who, but really useful advice to go like, âHey, maybe instead of being in the weeds and trying to solve each problem one by one, maybe you need to write something about what is the culture of how we approach problems in compliance.â So I wrote a doc on this is the culture of compliance, and I went really deep into when should we think actively versus when we should just copy what are the expectations from other bank compliance things.
Maybe it sounds simple, but just that advice and writing that culture doc led on to a bunch of things that were super useful and really changed the way compliance worked at Mercury and really helped. But those little things, it was a blind spot for me. Afterwards, I was like, âOh, yeah. Of course. I shouldnât be going deep on every problem. I should be trying to solve the root of what is the culture around how we solve problems in compliance.â But sometimes you just need someone to give you that.
Nakul Mandan:
Is there a part of being a CEO that has become harder with scale?
Immad Akhund:
Thereâs different levels where you just have to change the way you do things. One thing that bugs me is Iâm a very direct and honest person and sometimes people take you too seriously.
Nakul Mandan:
Every word that you say-
Immad Akhund:
Yeah, exactly. Once I did this, I went into a Slack channel or something and I was like, âOh, yeah. Whatâs happening with this project?â And I genuinely just wanted to know whatâs happening with the project. The whole team was freaking out on the backend going like, âOh my God. Immadâs saying this projectâs late, and Immadâs annoyed about it.â So now when I have to say things, I have to really disclaim them. I have to say like, âHey, guys. I am not saying thereâs any problem, but Iâm just interested, and if thereâs an update, I havenât heard about it.â Just a reasonable number of disclaimers, which Iâm getting it better at, but itâs not my natural mode of operation. So yeah, thatâs a small thing that bugs me.
Thereâs definitely two or three phases as a CEO and you just have to operate differently. Thereâs the initial phase youâre still an IC, right? Less than 30 people, youâre probably just getting stuff done as a CEO, and then you have to move to a phase where youâre mostly managing managers and thatâs a different phase thatâs maybe from 20, 30 people, and then eventually you get to this phase around probably around 200 where youâre mostly managing execs. And most people would say that type of size and it varies a little bit, but those three jobs as a CEO are three different jobs. Theyâre not the same job at all, but none of them are inherently hard or inherently easy, but you have to... I think the bit that other people sometimes struggle with, theyâre like, âOh, they really like being an IC CEO.â Whereas Iâve always been like, âI want to build the biggest company CEO because I just think thatâs fun. I want to have the biggest challenge and the biggest impact.â So Iâm willing to really learn and try to be the best CEO of all three of those phases.
Maybe thereâs a fourth phase I havenât got there yet and all the 10,000 employees CEOs are like, âYeah, youâre just a strategic resource allocator or something like that.â Or maybe all companies will be small companies in the future. But yeah, itâs definitely a different job at those things, but I think a big part of being as good CEO is learning to do those jobs but also enjoy the jobs and also construct them in a way that you can enjoy them, because I donât know, I feel like if you donât enjoy what you do, youâre not going to do a good job with it.
Nakul Mandan:
I also heard you say in the podcast, I think it was a recent podcast where you talked about youâre at the bonus levels as a founder now. Youâve accomplished much more than what you set out to do. Can you talk about how the scarcity era Immad versus the bonus era Immad makes decisions differently, if at all, or plays the game differently?
Immad Akhund:
For a long time... I have a relatively poor background. My parents were on welfare and we moved to the UK with very little. So for a long time, I started as entrepreneur, I loved it, but I didnât have some grand ambition. I just wanted to survive. There was definitely a phase where it was just survival and I was fine with any exit, and when we sold our last company for 45 million, and that was my first 10 years of entrepreneurship, so it was a long time. I finally got to a level where I was like, âOkay, Iâm happy.â And then it was a kind of existential thing. I was like, âOkay, what do I really want to do?â And I considered being an investor. I eventually decided I really love doing a startup and I wanted to go for something really big. So I do think the mentality shift allowed me to go for something much bigger that I was willing to, especially in 2017, doing a new bank was not a thing you did.
I started the company and I was like, âMaybe this will take three years before I raise anything else or get to product market fit.â And I was really committed. I was like, âIâm going to spend three years doing it if thatâs whatâs necessary,â kind of thing. Thereâs no way you could take that three-year big swing if you have a much more scarcity mindset. So that was definitely a mindset shift for me. Iâm trying to think of what else. Iâd say then for a long time I had this other goal in my head, which is build a unicorn. Again, early on in the startup thing, everyoneâs like, âWow. Build a unicorn, get to a hundred million in revenue,â et cetera. And we got to a billion dollar valuation in 2021. So there was definitely at that point, which is I would say I really got into the bonus levels this game. I was like, âI donât even know what my next goal is,â at least in terms of the money side of things.
At that point, I would say I shifted a little bit to thinking, okay, how do I just build something that customers really love that can be as big as possible and really just disrupt banking? And so itâs a little bit more vague as a goal. Getting the first exit or the first unicorn valuation, et cetera, was a more definitive thing. Now itâs more open-ended. But yeah, I think each of those phases, you end up having more long-term goals and in some ways more long-term prep. I think this helps a little bit with your previous question around stress management. I think if you are always grateful for what youâve achieved and you do think about it as like, âOh, this is the bonus level,â otherwise you can be like... In some ways thereâs just so much more to think about today and thereâs more to stress about. Iâve tried to make it so the pressure goes down as success goes up, which is I think actually relatively rare.
Nakul Mandan:
I think itâs very rare, yeah.
Immad Akhund:
But it should be like that, right? If youâre more successful, why are you more stressed about it?
Nakul Mandan:
Is there something youâre still trying to prove to yourself though?
Immad Akhund:
I mean, thereâs always someone... Elon Musk is out there. Heâs running five billion-plus dollar companies, right? Thereâs always something if youâre somewhat competitive and thereâs always... But I think thereâs unhealthy, âI am not successful until Iâve squished everyone,â or something. So thereâs unhealthy competition and I think Iâm relatively healthy about it that thereâs people out there that I think have done amazing things and maybe I can try to achieve some part of that. But yeah, itâs not something that I stress about that I need to beat other people or something.
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com -
Zijn er afleveringen die ontbreken?
-
Cameron McCord spent 484 days underwater. As a submarine officer in the US Navy, he learned to run a reactor in a space where the crew sees you 24/7, even brushing your teeth, and where there is no "later" to sort out a disagreement. He still calls the Navy the single biggest influence on how he leads. After the submarine came Capitol Hill as a Navy congressional liaison, then early Anduril, then Saildrone, then a stint in venture at Lux Capital. Each one was a deliberate move toward starting a company.
â
That company is Nominal, a connected software suite that is changing how the world tests and operates hardware. Three years in, Nominal is valued at $1 billion, has raised $155 million in ten months, and counts four of the five largest defense contractors as customers. The team has grown from around 40 people to 170 in a year, across offices in LA, Austin, New York, DC, and now London.
â
In this conversation, Cameron walks through the operating playbook underneath that growth. Why he still interviews everyone and looks for people who are "three layers deep." What it means to "earn the right to stand the night watch." How he imports intentional ambiguity from early Anduril and a critique culture from the submarine. And where physical AI is actually going.
â
In this conversation with Cameron McCord:
00:00 Who is Cameron McCord?
01:34 What did 484 days underwater teach him about leadership?
10:36 How do you "earn the right to stand the night watch"?
16:28 What was so special about Andurilâs culture?
28:39 What is the "power of ambiguity"?
31:29 How does Cameron think about recruiting?
37:39 Can you recruit âkillersâ who are also low ego?
46:50 What was the broken old way of testing hardware?
50:50 How do you actually crack selling to the government?
55:12 How did Nominal land four of the five largest defense primes?
1:01:33 What does "physical AI" actually mean?
1:18:14 What is Cameron still working on in his inner game?
1:23:22 Quickfire: military movies, leadership books, and a favorite office?
1:25:07 What advice would Cameron give to his 25-year-old self?
â
Cameron's sharpest lines from this conversation...
On why the weakest link defines the team:
"Even if you're the person turning a wrench on a pump, if you don't do your job, the team could be at risk, because everything has to come together."
On what he recruits for:
"The biggest gift, and frankly the thing you look for in talent, in recruiting, is people that are fit to stand the night watch⌠a safe pair of hands."
On asking a team to do the impossible:
"It is always worth going the extra mile to draw the thread between the impossible task you're asking someone to do and the outcome for a customer."
On what it takes to win:
"Do one thing and do it really, really well⌠in practice, you need to do something 10 times better than the current way."
On the inner game of leadership:
"Being comfortable in your own head is so powerful, and itâs the biggest gift I can give to Nominal."
On a mantra from his submarine captain:
"Stress is the body preparing you for greatness."
â
Where to find Cameron:
⢠Nominal
⢠X
⢠LinkedIn
â
Where to find Nakul:
⢠Audacious Ventures
⢠X
⢠LinkedIn
â
Where to find Audacious Ventures:
⢠Website
⢠LinkedIn
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com -
Akshay Kothari has spent most of his time at Notion trying to put himself out of a job. His move, repeated for years, is to run at whatever the company's biggest unsolved problem is, build the system around it, hire someone great to own it, and then remove himself. In 2023 the approach left him, a co-founder of a multibillion-dollar company, with zero direct reports and a calendar wide open to think.
That instinct fits the company he helps run. Notion built itself like an art project: profitable, still largely owned by its founders and employees, raising money only when there was a real reason to, and kept deliberately small, around 1,200 people running a business at public scale. For years its core was a system of record, a modern editor and database that people genuinely loved and built their work inside. Then AI changed what people expect software to do, and Notion decided to reinvent the product rather than wait to be replaced.
In this conversation, Akshay walks through how Notion pulled that off: the stretch he calls the swamp of despair, when its AI agent failed four times before it worked; why the company stopped trying to make the model fit its product and started fitting the product to the model; and how it dissolved the line between its AI team and everyone else until there was nobody left who wasn't building with AI. The bet landed. AI moved from a defensive play to a real driver of growth, and Notion's growth rate has climbed for over a year, with the most recent quarter running about 50% above where it was a year earlier. As Akshay frames it, the business went from go-karting to Formula 1, and now the company has to rewire how it drives.
Listen on...
⢠Spotify: https://open.spotify.com/show/5KorHDryscDzLwPkWEgAH1
⢠Apple Podcasts: https://podcasts.apple.com/us/podcast/knuckle-up-with-nakul/id1893213920
Akshay Kothari is the co-founder and chief operating officer of Notion, where he has built and scaled functions from support and sales to marketing and finance, alongside co-founders Ivan Zhao (CEO) and Simon Last. Before Notion, he co-founded Pulse, a news-reading app that won an Apple Design Award, reached 30 million users, and was acquired by LinkedIn for $90 million in 2013.
Where to find Akshay Kothari...
⢠Notion: https://www.notion.com/
⢠X: https://x.com/akothari
⢠LinkedIn: https://www.linkedin.com/in/akothari
Where to find Nakul...
⢠Audacious Ventures: https://www.audacious.co/
⢠X: https://x.com/nakul
⢠LinkedIn: https://www.linkedin.com/in/nakulmandan/
Where to find Audacious Ventures...
⢠Website: https://www.audacious.co/
⢠LinkedIn: https://www.linkedin.com/company/audaciousventures/
In this conversation with Akshay Kothari...
00:00 Who is Akshay Kothari?
01:27 What were Notion's core founding principles?
06:20 Which early cultural principles scaled, and which broke?
08:22 How did Notion hire its first employees, and where did they come from?
11:48 How does hiring work now that the founders can't meet everyone?
14:35 Why does Akshay, as COO, prefer to have zero direct reports?
19:05 How do Ivan, Simon, and Akshay divide the work?
21:07 Does Notion's intentionality ever conflict with speed?
25:25 What should other founders steal from Notion's culture?
28:11 When did AI become a reason to rethink the whole product?
30:44 Why were the early AI years a "swamp of despair"?
36:05 How do you push AI across a huge product without losing the user?
39:25 Does Notion buy its AI DNA or build it?
40:44 Should Notion be afraid of OpenAI, Anthropic, and fast copycats?
46:58 What's hardest about the reinvention, and what does "meet the LLM" mean?
52:42 Is Notion AI-native in every function yet?
54:36 Are Notion's engineers still writing code, and how has engineering changed?
1:01:07 Once building is cheap, what's the new bottleneck?
1:02:39 How is AI reshaping sales, marketing, and support?
1:09:07 How many agents run inside Notion, and who builds them?
1:11:28 How has recruiting changed for the AI era?
1:13:48 What still worries Akshay about Notion's future?
1:15:22 Quickfire: admired founders, books, overrated AI advice, and Akshayâs superpower
1:19:42 What should a $50M pre-AI company do in the next 90 days?
Akshay's sharpest lines from this conversation...
On putting himself out of a job:
"Ivan calls me a stem cell. I go there and build something, then I get out of it."
On meeting the model where it is:
"You really have to stop trying to fit the LLM with what you have, and you have to fit what you have to the other."
On what AI is really for:
"The best companies are just raising their ambitions. It's less about efficiency and more about, can each person be way more productive?"
On just how much AI changed business:
"We cannot take the way we were go-karting and apply it to Formula One."
On disrupting yourself before the market does:
"You have to disrupt yourself to where the business is going. You can kind of do it to yourself before someone else does it to you."
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com -
Ashwin Sreenivas spent his childhood in India waking up at 5am and studying until 8:30 at night. History, geography, physics, chemistry, math. Every day, from 4th grade through 12th grade, through the Olympiads and the National Talent Search Exam. He says now, three years into building one of the more successful post-ChatGPT companies in the world, that all of that is precisely why what he does today feels almost easy. "I get to come in here and there's a lot of people and I'm having fun."
That mode is what Decagon runs on. The company Ashwin co-founded with Jesse Zhang in 2023 is now valued at $4.5 billion, has crossed 450 employees in three years, and works with some of the largest enterprises on the planet. The path there has in some ways been simple: don't theorize about where AI is going, talk to customers until the pain is unmistakable, build for that, ship, repeat. Decagon went from zero to $1 million in ARR with two co-founders and no employees.
In this conversation, Ashwin walks through what that means in practice. How Decagon operationalizes a single cultural priority: speed, even when it costs coordination. How they hire 450 people without breaking the bar. How AI has reshaped the IC engineer, the AE, and the VP of EPD. And why, after a year of running 6+ days a week, the thing he and Jesse would tell their earlier selves is: go faster.
Ashwin Sreenivas is the co-founder and CTO of Decagon, the AI customer concierge platform founded in 2023 that serves enterprise customers including Substack, Eventbrite, Duolingo, and Notion and is valued at $4.5 billion. Decagon Labs, the company's in-house model development effort, now powers around 90% of Decagon's model traffic. Before Decagon, Sreenivas co-founded Helia in 2019, an AI startup acquired by Scale AI a year later. He started his career as a strategist at Palantir Technologies in New York. Sreenivas holds a Bachelor's degree (2017) and Master's degree (2019) in Computer Science from Stanford University. Decagon raised $35M in 2024 and has scaled to over 450 employees.
Where to find Ashwin Sreenivas:
⢠Decagon
⢠X
⢠LinkedIn
Where to find Nakul:
⢠Audacious Ventures
⢠X
⢠LinkedIn
Where to find Audacious Ventures:
⢠Website
⢠LinkedIn
In this conversation with Ashwin Sreenivas:
00:00 Who is Ashwin Sreenivas?
02:10 How did Jesse and Ashwin decide what to work on at Decagon?
04:19 Why did they reject the top-down market-sizing approach?
13:16 What does Decagon give up to keep moving fast?
17:54 Does Decagon expect their team to be 6-days in-office?
23:33 Why didn't they hire a single employee until $1M in ARR?
27:12 How do you hire 450 people in three years without breaking the bar?
31:20 Are Decagon engineers even writing code anymore?
35:08 How does the IC engineer role change with Claude Code and Cursor?
36:32 Shipping in two days: how does EPD leadership change?
40:15 What are the two types of FDE, and which one do most AI companies actually need?
49:21 How will the human role at Decagon evolve over three years?
57:15 Why did Decagon build its own models in Decagon Labs?
1:00:50 What worries Ashwin most about Decagon today?
1:03:52 How does Ashwin manage his psyche while running this fast?
1:08:00 What hiccups has Decagon had that no one sees from the outside?
1:09:50 Quickfire: overrated advice, AI products, books, red flags
1:12:43 What would Ashwin tell his younger self about Decagon's journey?
Ashwin's sharpest lines from this conversation...
On what actually matters:
"If you build a company doing something that your customers care about, you can mess up everything else in a way and it doesn't really matter."
On founder market fit:
"If you pick the right market, the market will pull the problem and product out of the founding team."
On what Decagon traded for speed:
"The pace of building has changed so quickly. With AI, there isn't that much time for coordination. You have to give somewhere, and we gave for speed."
On packing desks tight:
"I specifically asked our head of workplace to get smaller desks so that people are packed even closer together, so that you can lean over and talk to an exponentially larger number of people."
On responsibility for AI-generated code:
"Use all the tools you have available so that you can move faster, but at the end of the day, you are responsible for the code you push and you should be prepared to defend it rather than say, 'oh, the AI agent wrote it.'"
On hiring red flags:
"It's not a specific flag, but rather that gut feeling of something's a little off and I'm not sure I want to pull the trigger on this."
On the normality of âfiresâ:
"I guarantee you every fast-growing company, probably without a single exception, they've had a thousand fires internally. Just normal. That's just how it is."
On the advice heâd give his younger self:
"Go faster, hire faster, build faster, get out to the biggest customers faster because the need is real, the market pull is real. You just need to go capture as much of it as quickly as you can."
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com -
Michael Grinich is a design-obsessed engineer who once spent days in a recording studio with an electronic musician crafting the perfect email notification sound. He now runs WorkOS, the $50M (accelerating) ARR enterprise infrastructure business powering nearly every major AI company you can think of, OpenAI, xAI, Anthropic, Sierra, Cursor. Michael has scaled the seven-year-old company to 100 people with no CRO, no VP of sales, and just three sales reps. In an industry where most CEOs default to hiring more executives, Michael runs WorkOS with senior ICs and a weekly operating cycle. In this episode, he unpacks the philosophy behind it all.
We also discuss:
⢠Why a great startup idea has to look bad first
⢠Why Michael subscribes to "Minimum awesome product" over MVP
⢠Micro-leadership over micromanagement
⢠How to "AI pill" your team
⢠Why senior engineers are the most impactful with AI
⢠The reverse Peter principle
Referenced:
⢠Billie Jean King
⢠Black Swan Farming
⢠Brian Chesky
⢠Cursor
⢠Founder Mode
⢠Ivan Zhao
⢠Model Context Protocol (MCP)
⢠Nat Friedman
⢠Paul Graham
⢠Peter Principle
⢠Seeing Like a State
⢠Twilio
Where to find Michael Grinich:
⢠WorkOS
⢠X
⢠LinkedIn
Where to find Nakul Mandan:
⢠Audacious Ventures
⢠X
⢠LinkedIn
Where to find Audacious:
⢠Website
⢠LinkedIn
Timestamps:
00:00 Introduction
01:33 From design-obsessed founder to enterprise infrastructure
04:20 Michaelâs year off and what made the WorkOS bet obvious
06:54 Why a great startup idea has to look bad first
09:46 Minimum awesome product beats MVP
11:09 The org with no CRO, no VP of sales, and one PM
13:29 Hiring for curiosity, not credentials
16:25 The "AI pilled" interview red flag
18:25 A week is 2% of the year
26:00 How WorkOS approaches brand
33:00 The future shape of engineering orgs
43:20 Why senior engineers benefit most from AI
44:45 Micro-leadership over micromanagement
49:10 Tough times in the early days
59:04 The reverse Peter principle
1:04:38 Quickfire: red flags, hires too early, and biggest fears
1:10:30 Michael's advice to their 25-year-old self
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com -
Bipul Sinha grew up in dire circumstances in India, made his way to IIT, and immigrated to the US in search of the American dream. By 40 he had become a successful VC at Lightspeed; he then founded Rubrik, today a $10 billion public company and one of the last decade's fastest-growing enterprise software businesses. Most founders look to mentors for guidance. Bipul's first principle is "nobody knows anything." In this conversation, he shares the mental models that got him here and how he's rebuilding Rubrik for the AI era.
We also discuss:
⢠The "state of intellect vs state of will" mindset shift
⢠Why "recruiting is like starting a religion"
⢠"Nobody knows anything" and how to use experts expertly
⢠"Product market fit is dead" and the S-curve stack
⢠"AdhogÄmÄŤ": only work on what worries you most
⢠Why time is the only real moat
⢠The "Maximal thinking" framework for dealing with uncertainty
Referenced:
⢠Bhagavad Gita
⢠Jiddu Krishnamurti
⢠Qasar Younis
⢠Upanishads
⢠Viktor Frankl
Where to find Bipul:
⢠Rubrik
⢠X
⢠LinkedIn
Where to find Nakul:
⢠Audacious Ventures
⢠X
⢠LinkedIn
Where to find Audacious:
⢠Website
⢠X
⢠LinkedIn
Timestamps:
00:00 Intro
01:47 State of intellect vs state of will
03:09 Bipulâs maniacal recruiting philosophy
07:06 Recruiting is like starting a religion
11:44 Will vs skill: you can teach one but not the other
14:28 AdhogÄmÄŤ: fighting mental downward slopes
16:03 Why Bipul thinks product market fit is dead
19:25 Nobody knows anything (and what that means for you)
23:06 Three questions that launched Rubrikâs AI transformation
32:04 âEither you go AI or you dieâ
33:42 There is only one moat
35:15 When Rubrikâs growth collapsed overnight
41:28 âMaximal Thinkingâ: How to succeed amidst uncertainty
45:12 Quickfire round: red flags, worst VC advice, more
47:29 Bipulâs advice to his 25-year-old self
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com -
Qasar Younis grew up on a farm in Pakistan, moved to Detroit as a kid, and worked at General Motors before landing at Google and becoming COO of Y Combinator. He then founded Applied Intuition, today a $15 billion company building AI for the physical world. In an industry where most people look up to tech founders, Qasar looks up to Sam Walton and Warren Buffett. Qasar is an N of 1 founder, and in this conversation, he shares his contrarian approach to company building.
We discuss:
* What truly makes a founder
* The âtwo exceptional indicatorsâ recruiting bar
* Why Qasarâs first 10 hires lived in a house together
* A simple framework for monthly performance reviews
* The âgolden age of small companiesâ
* How to operate with speed and intentionality
Referenced:
* Andrej Karpathy
* Applied Intuition
* Immad Akhund
* Kickstarter
* Peter Ludwig
* 18 Mistakes That Kill Startups
Where to find Qasar:
* Website
* Twitter / X
* LinkedIn
* Applied Intuition
Where to find Nakul:
* Twitter / X
* LinkedIn
Where to find Audacious Ventures:
* Website
* Twitter / X
* LinkedIn
Timestamps:
00:00 Intro
01:19 What really makes someone a founder
05:26 The company that almost became Kickstarter
08:12 The most common misread on feedback
13:40 Why most founders donât end up with the best team
19:45 How to pick a co-founder
23:38 Your first 10 hires are really your first 100
28:21 The case for hiring slow and firing slow
33:22 Red, yellow, green: how Applied gives monthly feedback
35:00 The role that knows whatâs actually going on in a company
40:01 How to operate with speed and intentionality
42:41 The three things Qasar spends time on
45:57 How Applied is driving AI adoption
52:06 The type of engineer Applied is now looking for
1:01:19 Why this could be the golden age of small companies
1:09:13 Quickfire: red flags, overrated advice, and superpowers
1:12:32 Qasarâs advice to his 25-year-old self
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com -
Frank Slootman is the only CEO in history to take three enterprise companies public: Data Domain, ServiceNow, and Snowflake. At their peak, the companies he led were worth over $200 billion combined. His playbook for building high-performance organizations, captured in his book âAmp It Upâ, has become required reading for CEOs. In this conversation, Frank opens up about the fear of failure that shaped his early career, why most CEOs tolerate mediocrity for far too long, and the moment he realized Snowflake needed a different kind of leader and chose to step aside.
We discuss:
⢠Why being a CEO is a confrontational job
⢠The âdrivers vs. passengersâ framework
⢠Why references matter more than interviews
⢠Why culture isnât about making people feel good
⢠How Frank faces his demons âfor breakfastâ
⢠How Data Domain survived year one on $3M and a product nobody believed in
⢠Why AI is a dislocation on the scale of the Industrial Revolution
Referenced:
⢠Amp It Up
⢠Data Domain
⢠Elon Musk
⢠Google
⢠Intel
⢠Peter Thiel
⢠Scott McNealy
⢠ServiceNow
⢠Snowflake
⢠Sridhar Ramaswamy
⢠Steve Jobs
Where to find Frank:
⢠LinkedIn
Where to find Nakul:
⢠Twitter / X
⢠LinkedIn
Where to find Audacious Ventures:
⢠Website
⢠Twitter / X
⢠LinkedIn
Timestamps:
00:49 Introduction
01:21 Why being a CEO is a confrontational job
03:51 Great people are hungry for hard feedback
08:19 Psychographic profiling: how Frank builds compatible teams
09:52 Drivers vs passengers: how to tell the difference
12:39 Why back-channel references beat interviews every time
16:19 âWhen thereâs doubt, thereâs no doubtâ
20:42 Inside Frankâs Tuesday operating cadence
22:27 The âgo directâ rule that breaks org chart politics
26:19 Why bigger goals force better plans
31:27 Standards are the real culture
38:17 The email Frank wrote every Monday for years
41:35 Advice for navigating todayâs volatility
47:25 Facing demons for breakfast at Data Domain
54:19 Why Frank fired himself as Snowflake CEO
1:05:19 Coming to Silicon Valley â10 years lateâ
1:07:59 Why AI is an industrial-revolution-scale shift
1:10:01 Frankâs advice to his 25-year-old self
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit knuckleuphq.substack.com
