Kapitus

Fanboy moment - meeting your legend to discuss the work you have been doing. Tom Davenport is one of the most cited management thinkers in the world on Analytics, AI and process re-engineering & innovation. 26 books, 300+ articles, HBR's most frequently published author, the person who put "competing on analytics" on the map, and most recently the author of Agentic Artificial Intelligence. One of the top three business and technology analysts in the world. His work was what got me enamored with data science 15 years ago and played a significant role in my decision to pivot toward it. Meeting him in person, and finding him completely down to earth, made it even more special. 🔹 Thanks to Lars Reinkemeyer , I got a chance to sit down with Tom and Lars to discuss what operational AI transformation actually looks like on the ground. We covered a lot of ground. How process intelligence was foundational before any AI deployment. What our team has put into production across underwriting and collections. How ML models, GenAI models, and deterministic automations work in tandem. What changes after every rollout: policy, product, process, and people. And how we monitor all of it and make sense of what to do next. At one point he asked about the combination of Six Sigma discipline and technical implementation capability and noted it was rare. It was a sharp observation from someone who is an expert on both of them. I think that rarity is precisely the problem most AI transformations are trying to solve without naming it. When you deliberately build both disciplines into a focused pod, the implementation results speak for themselves. 🔹 Towards the close, Tom complimented the clarity of our approach and assessed it as on par with some of the best organisations in the world that have built this capability. (PS: And to make the fanboy moment complete — Tom and Lars reiterated in the comments what they mentioned during the meeting. Yipee!) That kind of validation from someone who has seen the inside of hundreds of transformations globally is reassuring in a way nothing else can replicate. It motivates us to double down on execution without second guessing ourselves. 🔹 Thanks to Lars for facilitating this opportunity. And more importantly, thanks to the Kapitus leadership — Ben Johnston , Arun Narayan , Thomas Robina , Steven Podhorzer and Murali G — for their unwavering trust and sponsorship. None of this happens without that. And thank you both for the picture - a moment I will not forget. 🔹 Beyond the fanboy moment, one thing worth taking away: For operational AI transformation, you cannot vibe or prompt your way to process clarity. Build the full stack process-intelligence layer first (link in the comments) PI → AI. The order matters.

A $20T commerce economy will break application‑based SMB lending. McKinsey sees global commerce scaling toward ~$20T, with $3–$5T mediated by AI agents. That pace doesn’t wait for: • Applications • Historic financials • Point‑in‑time approvals In an embedded, agent‑driven world: ➡️Capital is pulled, not requested ➡️ Risk is continuous, not periodic ➡️ Credit is dynamic capacity, not a fixed loan Underwriting shifts from permission → live capital control. Lenders who underwrite behavior in real time will scale with commerce. Those underwriting yesterday’s risk won’t. #Underwrite behavior. Not history. #Kapitus #SMB #EmbeddedFinance #SMBLending #Underwriting #Fintech #AgenticAI #DigitalCommerce #FutureOfFinance

I'm honored to be a panelist next Tuesday at KBRA's Consumer ABS Breakfast Briefing titled "Credit Fundamentals Across ABS". I'll be discussing small-business ABS, which is a testament to Kapitus's leadership role in the industry and to our frequent issuances in the capital markets. If you are there, please stop by and let's chat (no tomatoes, please, lol).