Brex President & CPO on embedding Product team into sales process to accelerate revenue growth | Karandeep Anand | E251

In this episode, Carlos Gonzalez de Villaumbrosia interviews Karandeep Anand, President and Chief Product Officer at Brex, one of the fastest-growing fintech companies in the world.

Founded in 2017, Brex quickly became the go-to financial services partner for most Silicon Valley startups, with 80% of Y Combinator startups using their platform. This rapid growth led to over $1.2 billion in funding and a $12.3 billion valuation. Brex now serves a wide range of companies, from startups to large enterprises, including over 150 public companies with a combined estimated market cap exceeding $2.9 trillion.

Karandeep's impressive career includes 6.5 years at Meta, where he led a team of 3,000 people and served 200M+ businesses globally, and 15 years at Microsoft, heading product management for Azure's application and developer platforms. At Brex, he has been instrumental in transforming the company's Product Management culture and expanding its product into a comprehensive financial platform offering corporate cards, treasury solutions, and software for travel and expenses in one place.

In this episode, we explore Brex's evolution from a startup-focused company to serving enterprise customers, the integration of AI in financial products within a highly regulated environment, and the restructuring of product teams for efficiency and innovation. We also discuss Karandeep's unique perspective on product-market fit and the importance of continuous innovation in a rapidly changing market.

What you'll learn:

• Karandeep's approach to evolving Brex from a startup-focused company to serving enterprise customers.
• How AI is integrated into financial products in a highly regulated environment.
• Strategies for restructuring product teams to drive efficiency and innovation.
• The importance of PMs spending more time on customer calls and listening to customers.
• A third go-to-market motion mixing PLG and SLG: Assisted Product-led growth.
• Why product-market fit is a journey, not a destination, and the importance of continuous innovation.

Key Takeaways:

• AI Integration: Karandeep highlights the use of AI not just in products, but also to reduce change management costs for customers.
• Product Team Restructuring: He discusses the transformation of Brex's product team, emphasizing the role of PMs as ""mini-CEOs"" for their product lines.
• Customer-Centric Approach: Karandeep stresses the importance of PMs being directly involved in sales calls and customer interactions.
• Continuous Innovation: He argues that product-market fit is not a static goal but a continuous journey in today's rapidly changing market.
• Go-to-Market Strategy: Karandeep introduces the concept of "assisted PLG" as a hybrid approach between product-led and sales-led growth.

Social Links:
- Follow our Podcast on Tik Tok here
- Follow Product School on LinkedIn here
- Join Product School's free events here
- Find out more about Product School here

Credits:
Host: Carlos Gonzalez de Villaumbrosia
Guest: Karandeep Anand

Transcript

Carlos González | Product School 00:03:04 Welcome to the show, Karandeep.

Karandeep Anand | Brex 00:03:06  Thanks for having me here, Carlos.

Carlos González | Product School 00:03:08  So you were a bartender before the legal drinking age?

Karandeep Anand | Brex 00:03:14  I hope I'm past that statuary  limitation of when I can be charged for that. But in India, when I actually started working at a friend's bar when I was 21. And it's interesting that in India, the legal drinking age actually is 25, which is like the weirdest thing ever. So before I was 20, not even 21. And a friend opened a new bar in Hyderabad and I was like, I need a job in the evenings just to go hang there and mix drinks and serve drinks and people like are, you are a little too young for this. And I'm like, as long as I don't get caught, it's all fine. So yes, I learned mixing drinks before I could taste them myself.

Carlos González | Product School 00:03:50  Okay, now you're in the US. It was a long time ago, so I think we're good.

Karandeep Anand | Brex 00:03:55  We should be okay.

Carlos González | Product School 00:03:56  So before you joined Brex around three years ago, you spent six and a half years in meta, and before that, 15 years at Microsoft. So I'm very curious to know what is your what is the different definition of winning in product management at those different companies?

Karandeep Anand | Brex 00:04:11  Indeed, I think Carlos and I was a huge fan of your podcast as well. I think you've probably already seen so many different perspectives on what product management is. I think is probably one of the most bastardized version of definition of a job. Role is like product management can be anything and everything which the company needs, and there's no clear person to do that job. So engineers write code. Designers love figuring out interactions, and Figma and Pixels and everything else kind of falls on the product management discipline, and it's a very broad discipline as a result. And I think, no, no better example than this, as where Microsoft used to have program managers is what they would call the product managers back in the day. They would do everything from write right specks, birds, birds as we know them today. And they would own making sure the bird gets released to marketing and then the sales. And they would coordinate the entire lifecycle of shipping the product. And hence the PMS of Microsoft are very different. there used to be a lot more coordination and very technical focus. And when I move to, meta, half the PMS were not even technical. In fact, best of the PMS were not technical at all. Fiji, who now is a CEO of Instacart? She's a journalist and she was the PM for video, and she was the PM for Facebook app. And then she became the GM. And we have so many examples of people who were very, very successful product visionaries who are absolutely no idea of, like, anything, tech as a background, nor were they, you know, mired in the details of how do you ship a product and take it to launch. Obviously they were smart enough to figure it out, but the main thing was the definition of what a product does is very simple. Your job is to figure out how to win in companies in your enterprise SaaS like Azure and Microsoft. It was all about technical, deep technical execution. In companies like meta, the definition changed very quickly to figure out how to get to product market fit and scale very rapidly. Consumer product. And in some cases, it was all about growth. In some cases it was about monetization, but a long way of answering a question, Carlos, which is the, you know, all the PMS I mentor and people are like, oh, what should I go learn about being a PM? I'm like, are you passionate? Can you learn new things and can you figure out how to win? If you can do that, you'll be a pretty good PM. so the the definition of PM effectively is figure out what your company and culture needs and needs most help with. You are the de facto bridge between customers and what will get built, and your job is to go make sure everything that's coming in the way of winning. You're removing the roadblocks. If you if you start with those first principles, then whether you focus on growth, monetization, bird, bird, all that gets sorted out pretty quickly.

Carlos González | Product School 00:07:01  But let's talk about Brex. You. Your official title says president and CEO, I think you are the first ever CEO. I know that's also a president. So what's the difference?

Karandeep Anand | Brex 00:07:12  You know, building a previous example of my job is to win. And when I joined Brex after meta and my last team in meta was like 3000 people, so like, hey, you're going to go run this team of like 25 people. I'm like, I love that because you can go back into building the product. And hence, you know, I joined the CPO. But two years into building the product and rebooting, you know what we need to be as a Brex? Because Brex started as a corporate card for startups. So we think of 2 or 5 founders in a garage using Brex cards to now like, you know, 50,000 people, enterprise companies using Brex. You had to evolve the entire product. So a job was to Redefine what part management means, what the part strategy roadmap vision should be. And that was what Brex needed three years ago to win. Last year we realised that we've built a pretty fantastic product, but the now next job is to how do you scale go to market? So as a result, I ended up expanding my my job to not just our own product. And typically most product people go and take on engineering. I went on and took on sales, marketing, customer success, because I figured that the next big S-curve for Brex will come by scaling go to market. And given you know how much of PM cycles get spent on customer and customer conversations, it's actually a very natural fit to go from CPO becoming a president and owning the customer lifecycle from selling to marketing and value proposition, positioning and customer success. Then we normally take it for granted, like most people think, like, oh, CPO should become CTOs. And I'm like, no, if you're amazing PM and you talk to your customers, you can actually be a pretty good go to market leader as well.

Carlos González | Product School 00:08:54  So you mentioned Brex started as a corporate card for startups. I also live in San Francisco and I remember being Brex everywhere, every single billboard. Right. It was literally the Silicon Valley darling. And now it's an entire platform with many more products. So love for you to guide us through that journey as a product leader and how you were able to figure out kind of what to build next.

Karandeep Anand | Brex 00:09:14  Great question. I think a lot of people think strategy is something that you sit in an ivory tower and you come up with like amazing, great ideas and like Boof, you build a product roadmap and strategy. that's that's one way to do it. It's a pretty bad way to do it, in my opinion. the right way to do is you have customers, you have one product, you have customers. Go talk to them. Don't ask them what they need. Go talk to them. Watch what they're doing. See where the pain points are, or see what happens when they grow up. See what happens when they wind down the business. See what happens when they merge with another company And automatically you start seeing a lot of patterns on what we today call as adjacencies. So most people are looking for adjacencies to their core business. Oh, Brex started as a corporate card. We didn't have to do a meeting strategy session to go figure out, like the next value will come from banking. We watched 50 founders, and all these founders are like, we love the fact that we can go to X.com and sign up for a credit card in five minutes. Boom, I'm ready with a credit card. And they were still driving, spending the entire weekend to go to their local bank or, you know, Silicon Valley bank and going opening a bank account. And it was like, oh my God. The next adjacency is not no rocket science. Next adjacency is if a founder is trusting us with their own financial service need, which is credit card. What are the other adjacencies we can solve for and make their life easy? And the reason why I double click on that example in this level of detail is sometimes we think strategy has to be like, you know, you need management consultants to do this Tam analysis and like do this Swot analysis. Know best strategies come by talking to customers and seeing where the market is headed. And it's far easier. And this you know this is an example of when your target audience is the same. So in this example it was founders. We did one job to be done which is credit cards. We did the second job to be done, which is banking and bill pay and sending ACH and wires and very became very, very easily became a portfolio of products but for the same audience.

Carlos González | Product School 00:11:14  So I hosted the co-founder and CEO at gusto on the podcast, and he shared something similar, when they started the company, they were targeting founders and they focused on processing payroll, which is also a category that is not very, sexy for a founder like myself. And, and he talked about how they were looking at how customers were hacking his own product as a way to influence what to build next, versus, as you said, you know, bringing a management consultant or overcomplicating a strategy, trying to trying to figure out what someone wants in reality. You already have people telling you with their actions what they really want.

Karandeep Anand | Brex 00:11:51  Precisely. Actually, you know, I love that example because the first example, Castle, I talk about how Brex went from cards to banking to bill pay was us talking to customers and finding needs. But I'm going to take another example. At meta, we did exactly what the customer guys did, which is we looked at how are people hacking our current products? And one great story we have is the the way marketplace was built, which is a very successful billion user plus product was not because we're trying to build marketplace. It started because we looked at how people were hacking the community groups that we had, like normal groups for families and friends engagement and started doing, selling and buying of items in the local communities and were like, wow, why would people do commerce on a social platform? And from that was born the idea of a buy sell group, which is a special type of group. And we just constantly kept the experience. And then all of a sudden we're like, let's bring all of this traffic into its own marketplace tab. So to your point, there are multiple ways of getting to adjacency strategies, and they don't have to be complex. As long as you're willing to talk to your customers, look at data and look at how they're hacking your current product to find what are the next obvious jobs to be done that you can solve with what you already have working?

Carlos González | Product School 00:13:00  So as you think about the next big iteration for your product, especially within the AI era, I noticed that you launch AI assistant and you are also launching other AI powered tools. But in general, like how do you go about incorporating this technology within a highly regulated environment?

Karandeep Anand | Brex 00:13:18  Indeed, I think, this joke that the day finance people can trust AI to touch money is a day I would have genuinely arrived and truly arrived in our lives because people are very comfortable with self-driving cars, but for some reason they don't trust AI touching your money. But on a more serious note, obviously traditional ML and traditional AI is what under par was Brex for the last seven years. What has changed, Carlos pretty heavily is with Genii. There's a bunch of use cases which are just impossible to do in the past, but now they become possible. So the two frameworks that we use when thinking about AI and how do you apply AI in product development now is in one of these two categories. Category number one is normal human tasks, which can be automated with the help of context like finding expense reports, collecting receipts, adding memos. These are like easier tasks that you can, you know, as you've seen with customer support automation, bot automation, a lot of that can be taken care of relatively easily because it understands context. The second category is actually more fun. The second category is tasks which are impossible to do but can now be done. They're still hard, but they can be done and they're possible. And I think that's where the finance use cases or a non-deterministic model is what becomes very interesting. I'll give you one example here. At the end of the month, finance teams are required to close their books and they have to reason over lots of different data, different transactions, GL codes because if they get it wrong, they can get audited. They have to file taxes on this. So this has to be highly deterministic with very low error rate. And instead of saying oh we will build these fancy models and JNI will solve the closing of books. We did exactly the model that Tesla used for autopilot. The model is you will be in control. So first time you're doing your self-driving. You keep your hands on the steering wheel. Then you can let this hand go after a while and 30s later, you're like, oh, okay, you know, the car can drive for 30s. I can trust it. Three minutes, it can trust it. And now we are a point where Waymo can actually drive me around San Francisco. And similar to that, what we did was we reasonable lots and lots of billions of records of data to prompt the accountants and the finance teams on all the rules they can apply automatically to close their books, based on the reasoning that we've done looking through hundreds and hundreds of millions of transactions. And first time you get 70% right, they're like, okay, I accept, accept, accept next time 90% right. And you get to a point where they're like, okay, I can trust it. Tell me what the five exceptions are. And we get to a point where they can gradually, you know, it used to be impossible to consider. The finance team will use AI to close books. We're getting to a point where they're like, it's possible, it's hard, it's possible, and they build trust. So the reason why, you know, this sharing for these two models is AI being used for normal automation. That's going to be table stakes. In fact, we won't even think about AI. That's like, obviously I should not be doing normal humans. How do you apply AI for things which are impossible to even consider is where a lot of value is being generated now. And I think that's the challenge for us, for our people, is to go figure out what are those impossibly hard tasks that are now possible because you can reason over in cinema, large data very quickly.

Carlos González | Product School 00:16:34  And switching gears a little bit, talking about your team, I think you had a quote that said that if a PM is not used in spending 20 or 30% of their time using AI, they should be worried about their role.

Karandeep Anand | Brex 00:16:49  I think there's definitely this meme going on. That part management is dead. And if I look at it honestly, there is some truth to that statement, which is a lot of PMS, especially in the last, you know, 5 or 7 years of excessive hiring. A lot of people became PMS. And I mentioned earlier, and if the work became coordination of work, I'll follow up with engineers, I'll follow up with design, I'll go through customer call recordings. Most of the things are gone, Carlos. That's not the job. In fact, that should never have been a job. But you could not do it earlier, so somebody had to do it. And people quickly forgot that all those were inputs that the PMS had to process to actually come up with the strategy, come up with the value proposition, come up with what next move to make. That was part management. Everything else was cost to doing part management. When the cost of part management goes down to zero, only the real PMS will survive. And I know it sounds scary, especially, you know, when I mentor younger new PMS. I'm like, do you know what it takes to really be a PM and like, do you have an opinion? Do you have creativity? Can you think strategy? Can you think out of the box? Because everything else that you're not doing out of the box is going to get automated. So I'll give you an example. We you know, we had average day eight hours of a PM. Let's say they're spending four hours going through a lot of data, a lot of calls and recordings and user feedback and user forums and user voice and free tools and stuff. You know, you start saying now, most of those 4 to 5 hours can be compressed into 20 minutes. And guess what? You don't need as many PMS to go through that kind of task anymore. So what are you doing with the rest of time in coming up with, with either much higher velocity and much higher iterations on what you can do uniquely. And that's really important to consider, which is, if you're not already using AI to assist, to cut down those four hours of busy work and hard work into 20 minutes, and using your remaining extra time to think creatively and to think strategically and solve through next generation customer problems. You are probably going to get eliminated in that role change that's coming very, very rapidly for all of us. So yes, I know it sounds very doomsday, but actually not doomsday. That's how productivity kicks in, is when you use assistant assistance from technology. And if people and PMS today are not living and breathing AI tools and cutting down the busy work by 90%, they are not spending the cycles on the right things.

Carlos González | Product School 00:19:24  I mean, and I think it's a way to be consistent with what you are selling to your customers, right? If you are selling their productivity and savings, then that should be a way to also implement that internally. In fact, I noticed that you had to rightsize your your product team. a few few years ago. So I would love for you to tell me more about that experience and where you are today.

Karandeep Anand | Brex 00:19:44  Yeah, I think we made that pretty difficult call in January this year where we are half the size of the team where we used to be, last November. And initially people are like, oh my gosh, you have a large surface area. We have a global fintech product with, you know, expense management, travel there. Like you have such a five product lines. You serve everything from two people to 20,000 people. There's no way with half the PMS you can keep running and scaling the business. The reality is, we've grown faster this year, shipped more products this year, and shared more features with higher customer satisfaction than we did in the last two years. Why? For two reasons one, I think leaner teams and efficient teams are generally moved faster. But second, the amount of information velocity that kicks in with assistance from these tools. We did not compromise on that front and we use heavily. Tools like glean to summarize what's happened. Instead of reading $200, lean summarizes it at the end of every call. No one sitting through, you know, going to gong calls, gong, Gong call, send us summaries on like here are five things we go look at every single tool. We enabled AI to be far more productive, to help us get to a point where all of the busy work that was overloading the PMS gets reduced into half. So half the PMS can not just do double the work. Can they actually do three times the work because they're now thinking on more creative, important problems? So I think we have a very breakfast of very living case study of how running leaner with the help of technology, especially AI, can fundamentally unlock value and scale.

Carlos González | Product School 00:21:17  So tell me more about how you're structuring your your team, especially now that you also oversee go to market.

Karandeep Anand | Brex 00:21:22  There's still clean function. So I have product management as a function. I do have sales, pre-sales marketing and customer success as independent functions. I think one thing that changes with the fact that they all are part of my org is there's a very tight connection and a very fast feedback loop between the that's building to the seller or the marketing person, which is right, you know, building value proposition and the seller. That feedback loop close gets close very clearly, but the organization is very clean. I have an amazing CRO responsible for revenue and all selling, he runs all of our sales and, and demand teams marketing, which runs all the value proposition brand campaigns. And we have the product management team, which is comes up with the core product areas. But big difference, Carlos, from last year to this year is we used to have PMS assigned to features and technologies and customer segments and all that is gone. RPMs are like mini CEOs. I have five lines of business, between cards, banking bill payment, which is accounts payable, travel. And I have exactly 1 p.m. for each one of them as opposed to having multiple PMS features. It doesn't matter 1 p.m. which looks at their product line as a business. They look at acquisition, retention, revenue, restrictions churn like a full, full. You know, CEO does it for the business. They do it for the area. And, you know, a lot of people in the industry get very triggered when we use PMS as a mini SEO analogy. But the reality is that's exactly what a PM is supposed to be doing, is looking at a holistic picture of what I'm building. Is it selling? Is it successful? What do you need to do to make sure I'm creating value and capturing value? And I think that's my expectation from a PM is to be able to do that well. And you don't need a lot of PMS for it. A handful of smart PMS can do that.

Carlos González | Product School 00:23:13  In fact, I heard you say that you get PMS involved in sales calls too, right?

Karandeep Anand | Brex 00:23:20  Oh my gosh, my secrets are leaking. Yes. So now we have this, you know, every time we're building a new product, one of the things which I, you know, always irked me and frustrates me is the time it takes for a PM to then explain to the product marketing person who writes the sales enablement material, to train the salesperson to sell a product, that just needs to die. So we have a rule which says that the first 25 customers are sold by the PM itself. So building a new feature. The PM is the one who's actually on the sales call explaining why what they built is actually much better than what customers have. So forcing the PMS to actually think about value proposition and pitching and positioning and pricing is a very good way to make sure that we're not building technology for the heck of building technology. We're building technology because it genuinely creates some value. And through those calls of selling customers, they also understand how to capture that value through monetization. So once they get 25 customers to say, yes, I can buy this, I can sell this. luckily, I don't have to pay my PMS any sales compensation. So that's good. So they won't get any quota on that one. But they have to go sell because then I have confidence that the sales team can do this at scale. So sales and marketing are scale functions. They should not be the testing functions. PMS are the ones who should be accountable for finding the right PMF and getting to a point where they feel confident that this can be repeated.

Carlos González | Product School 00:24:51  So how do you find or train PMS to develop that type of muscle? Because I agree with you, GTM is becoming more and more relevant within PMS, but it's also not the obvious background, right? Like a lot of us come from an engineering background.

Karandeep Anand | Brex 00:25:05  That was probably the hardest transformation I've gone through with my PM team, and we had to do a lot of that at the beginning of this year. So, you know, we come breakfast of deep fintech, we build, you know, very, very, very complicated global financial services. And a lot of our previous PMS essentially used to work behind the scenes on risk and fraud and modeling and underwriting. And very quickly we realized that if that's what they do and they're not spending a ton of time with the customers and selling and figuring out what to build next, we will both be not using PMS correctly. But more importantly, that's not where value will get created for. So what we did earlier this year for almost four months was we completely took away any responsibility that PMS had on coming up with the engineering backlog, designs and interactions to our engineering and design leaders and PMS were responsible exactly for one metric. You need to grow adoption of your product. The minute you change that. They're like, wait, wait, wait. I don't not control adoption because I'm not going to sell. I'm like, sounds good. Why don't you start selling and just changing the top line metrics, saying your job is adoption, not even revenue. Next will be revenue. Start with adoption forces, all alignment on what they need to prioritize and not prioritize. So it took us, you know, 4 or 5 months of retooling the entire discipline to be heavily outcomes driven from revenue as a lagging indicator of adoption. But everyone is gold on adoption, and that very quickly started changing what they spend time on, who they talk to. They start spending a lot more time with the go to market peers as opposed to the engineering peers, because now building the next feature is not about sexy. What's sexy is to move the chart up and to the right.

Carlos González | Product School 00:26:50  Love that so much. well, let's talk about go-to-market. because now that you oversee both sales and product, so you might have a motion or a slow motion. So how do you go about combining those two in a productive way?

Karandeep Anand | Brex 00:27:05  the holy grail of PLG and SLG. I think one of the hardest parts of Brex job at Brex. We had a very successful company because we built the first corporate card for startups, and there are hundreds and, you know, thousands of stars that we sign up every month, every quarter. And that all had to be paid. And two years ago, when we started building the enterprise business, we had to go build an entirely new SLG muscle as well. And now we're at a point where, like, wow, you can serve two people in a garage, a founder, and you can serve an enterprise, global enterprise with like 50,000 employees globally through sales. We serve everyone in between that spectrum very quickly. The idea of PSG and SLG breaks down, and what emerges is essentially a humans in a loop plg motion, which not only accesses SLG but also makes PSG a lot more human. And what this looks like is we actually don't have two motions. We have three motions, we have pure plg, which is to sign up for a credit card transfer bank account. Boom, you're good to go. And the product assists you with onboarding and activation. Then you obviously have this massive enterprise deployment, which is a sales rep is talking to you. People flying help you deploy the software and configure and roll it out. But bulk of our growth comes from what we call lite human assisted plg, which essentially is still a sales rep involved. But sales rep does two calls to the customer. Everything has to be done heavily by the product being as frictionless as possible from onboarding, KYC underwriting. All of that is tech assisted and assisted as opposed to that being human processes. And that is the hybrid model that we finally got working at Brex as well. So we're a company with three models Plg, SLG, and this light assisted Plg still run by sales team and sales supporters for it. But it's fantastic to see we can go from 2 to 20,000 companies and everyone in between. With that increasing level of automation in the product to not just acquire but also retain and cross-sell customers.

Carlos González | Product School 00:29:20  I can imagine. When you started targeting founders of startups, there was an opportunity to become that first product that they needed for credit card or other financial needs. But as you move up market and start targeting large enterprises, that has been around for a long time. They might have some solutions prior to you having that conversation. So for for that, that type of product that it tends to be very sticky, even if it's not good. It's just so hard to change. How do you go about kind of penetrating that account and potentially expanding it?

Karandeep Anand | Brex 00:29:52  I love that I love that question so much because when the, you know, most PMS at least think like, oh, all I have to do is build the feature. And if my feature is better than the the competitor's feature, people will buy me. But as you described, the cost of using your product is not just your prior. The cost of using part is the cost of removing the old product, removing all the old material and muscle behavior, and then deploying a product. So you have to create n times more value to justify change management, cost switching, retraining and redeployment. And that is way harder than just building another feature. So the question is how do you do it? there's a there's obviously, you know, the you have to especially in today's world, given how cost conscious every company is, they really have to see a value in moving to a new solution. So one of the things we had when we were building Brex expense management was the market has 6,080% penetration with Concur and American Express. And the challenge for our teams, for our teams was if your product is not literally ten times better, why would I ever ask somebody to go rip out an entire concurrent deployment in place of the Brex? So we started with a ten x better product experience as the North Star. So my first goal was like it has to genuinely be that much more valuable. The second part was we need to take the customers on a journey of, you know, landing and expanding so they're not paying all the change management costs upfront. And that's the trick most people forget, is to reduce the friction cost. You can actually start and use a land and expand as a, as a core motion versus an in place. And that's a good way to to reduce the friction required to try out your product. The third and most fun part is, you know, I can disrupt you, but you can use AI to disrupt your competitors as well. So what we used to we go into accounts and normally what would take us six months to understand change management. We can use AI to automatically generate user onboarding material for that company. We can automatically use AI to reconfigure the connections to send data to us. We use AI to, you know, repackage what they have in terms of API connections. All of that can significantly get reduced thanks to amazingly LM powered AI solutions now. So constantly thinking AI doesn't have to only be in a part. Yeah. Can actually be leveraged heavily into change management cost reduction as well. So all those three things which is techniques, better product land and expand versus rip into place. And then using AI to reduce change management cost have are three strategies which have worked exceptionally well for Brex to go win like very, very large enterprise companies. And in fact this morning we were celebrating. We added a 1/75 public company running on Brex. Well.

Carlos González | Product School 00:32:42  Yes. What what a good time to have you on the podcast, you have so many hot takes. I need to bring this one up. You said: product market fit is dead. Tell me more about it.

Karandeep Anand | Brex 00:32:55  Oh my gosh. yes. You know, this is controversial, as it should be. The 101, like the 101 off product training starts with thou shalt find product market fit, and then you will scale. And. Great. Sounds like a, you know, sounds almost like Bible. So let's let's listen. What does that really mean? Right. The very definition of product market fit means you have built a product, or you've identified a problem which you can solve in a large enough market, which is hence a market in a way where users are willing to pay a price for it. Very simple definition of PMF. And you know, this logically makes sense. The only problem, Carlos, is there is no definition of market left anymore. The market changes every day, every week, every month, and it's a constantly evolving, rapidly changing cycle. So if you assume that I have a static market, you would assume you have a static product. When the market is changing, the buyers are changing, the buying criteria is changing. And imagine this is pretty high where they are. The market changes way before we understand what's going on. The notion that you have you found PMF and then you can keep scaling. That's just bull. So sorry for very very strong language about that. But I think which is the single biggest reason I see most companies fail is they assume they've achieved PMF and hence they start sleeping on the wheel. PMF is not a destination anymore. PMF is a journey, and you have to constantly keep evolving because what you thought was PMF last week, last month, last quarter know there are ten new competitors, five new ways to do the same thing. And with AI and the momentum of building products, the AI PMF is going to become a very, very rapidly shifting goal. So hence anyone who's still sitting on saying, oh, I have found PMF in this segment with this core audience, and now let's just keep scaling. They're in for a pretty rude shock. So if you do not wake up every day morning saying very good chance my PMF or my product today is dead, and I need to go back to first principles to saying, hey, do I still have relevance? Am I still the best thing out there? Am I actually still solving the same need which existed, I thought a month ago, or six months ago, or a year ago? And am I still charging the right price for it to justify why I should be charged. You will not have PMF for too much longer. So hence, saying PMF is dead is a very blunt way of saying if you don't wake up and understand that the market shifts way faster than your product shifts, you will always be out of PMF. So your job as a PM now is to make sure that in the equation of product market fit, product innovation is outpacing market's expectation. That is what the new world of PMF will look like.

Carlos González | Product School 00:35:43  So when you are building a multi-product organization, I can imagine there might be some business lines that are more mature that we can justify. They have more traction than others, but you still have conviction in some new ones. So how do you go about protecting those initiatives or products that might still not move the needle on an absolute basis? But but they are part of your future.

Karandeep Anand | Brex 00:36:04  I love that. I love that because it's so easy to run out of patience when you have a big business and you have small incubation businesses, because every incremental down on a big business will generate big yields in the short term. And then and then it's very easy to move investment between, an incubation versus like a fully established business. But I think that's exactly why companies die. You know, you can call it a classic innovator's dilemma. You can call it adjacencies playbook, whichever playbook you like. The investment portfolio has to think about 17, 2010. Or you could do 60, 30, ten depending on on the company and the domain you're in. But effectively, if you're spending anything more than 60 to 70% on your core business and not earmarking 40% for where the business will come from a year from now, two years from now, three years from now, very good chance that you will be out innovated. So what we do is we have these incubation businesses. So 70% is core business heavy investment in making sure that we continue to scale what we have today with constant innovation 20% are obvious adjacency. So the example I took was we had cards when we were banking or we had cards and we built expense management. The core business was 70% was card, 20% came for expense management. But we also had 10% of moonshots. So while I was building expense management and we knew the market will very quickly evolve to also have travel expectation of the same cards, one card that can do cards, expense management, travel bill pay and 10% of team. Even though my 20% was already starving for resources, we kept 10% of the team ringfenced to travel because we knew that customers expectations will change way faster than we can start building a new product line, and it made us go slower on the 20% investment versus 30. But it made sure that at the end of the year, we had all the three products working in unison really, really well. And the way we used to evaluate those teams, which is to your question, how do you protect those investments is we're not looking to bring revenue from the new investments. All I'm looking for are signs that customers are finding value they're retaining, they're using, and they would sign up and there's adjacency. So you start looking at attach rates, early adoption and churn as the indications of value versus revenue and users and stuff. Because that's not a proxy of value for an incubation business.

Carlos González | Product School 00:38:30  So what's your next moonshot after travel.

Karandeep Anand | Brex 00:38:33  that's confidential because we're a very competitive market. But safe to say that in the world and fintech world that we are in, three things are constant. One, customers will constantly keep expecting more and more global products versus fintech. As you know, Carlos is very country specific. You have to do one country at a time, but most customers don't have the patience to wait for like 85 solutions in 85 countries. They want one global. So so a lot of our innovation is happening in making sure everything works globally. Second, we talked a lot about AI that will change fintech more dramatically than anything we've seen before. Because finally, we can reason over with high deterministic numbers over financial data, which was impossible to do in the past. And that will change how people use financial products. So you can expect a lot of announcements in that frame. And then third and all, the most fun one is continued innovation for the end user, because my ultimate vision is every employee in every company can make the best spend decisions for their company because they have perfect information on how every decision is up to the company's value. So we have this campaign called Make Every Dollar Count. That's what Brex is trying to do. And a lot of the innovation that we do is to make sure every individual employee is empowered to make every dollar count in the company.

Carlos González | Product School 00:39:51  Well, thank you so much, Karan. It's been a pleasure to have you on the podcast and learn more about how you're thinking about building products.

Karandeep Anand | Brex 00:39:57  Always fun catching up with you, Carlos. Thanks for doing this.