Kevin Henrikson

Demoed our voice AI to a urologist this week. I played the hard calls. Prior auth lookups. Referral checks. Insurance verification mid-call. He stopped me. "Do you have a simple one where a patient's just trying to make an appointment?" I was showing off. He wanted the 90%.

7:42 pm: appointment rescheduled by voice. 10:05 am: prior auth approved through the portal. 10:11 am: co-pay balance collected by text. And everything else on your list. Your front desk closes at 5. Your patients don't stop calling. Pretty Good AI handles the work in between and writes it all back to athenaOne.

Half of the companies on the athenaOne marketplace joined after ChatGPT launched. Not athenahealth employees telling me that; data. I sat down with Max Alsgaard-Miller, who spent eleven years inside athenahealth before spending the years since tracking 550-plus marketplace partners and publishing quarterly reports on who is actually winning. Max walked me through where the marketplace actually came from: Jonathan Bush and Clayton Christensen's more disruption, please experiment to invite athenahealth's own competitors onto the platform instead of fighting them. That bet is still paying off years later. We got into why Chrome extension and RPA-based integrations stay fragile by design, why athenaOne's 800-plus APIs reward companies that go deep on one EHR instead of claiming to support all of them, and what actually happens when athenaOne builds a feature your startup already sells. Then we pulled up his live Q3 numbers together. AI phone agent connections nearly quadrupled quarter over quarter. A revenue cycle management surge was hiding behind one partner's churn. And most marketplace ratings tell you a lot less than the sheer volume of recent reviews does. CHAPTERS 00:00 – A decade inside athenahealth, then a career tracking the marketplace from outside 02:42 – The origin of the athenaOne marketplace, Jonathan Bush, Clayton Christensen and more disruption please 04:55 – Why half of marketplace partners joined after ChatGPT launched 07:44 – Why Chrome extensions and RPA are fragile by design 10:28 – Navigating 800 plus APIs without breaking a practice's workflow 13:48 – What happens when athenaOne builds the feature your startup sells 17:50 – athenaOne owns the ledger, startups own the edge 20:43 – Why going deep on one EHR beats claiming to support them all 23:48 – The origin of the Athena Collector Rules Engine and why RCM startups are challenging it 25:49 – Q3 data, AI phone agents like Pretty Good AI explode and RCM hides a quiet surge behind one partner's churn 31:32 – Why marketplace ratings do not mean what practices think they mean 35:41 – Where the marketplace goes next: RPA, regulation, and the race between payer and provider AI

AI is getting better at solving problems. It’s also getting better at finding ways around the rules. This week on Founder Mode, @jasonshafton and I dig into what happens when AI agents become powerful enough to do things their creators didn’t expect. The technology is moving fast, and the opportunities are enormous. But as agents take on more responsibility, founders have to think harder about security, control, and what makes their businesses worth building. The goal isn’t to slow down innovation. It’s to move fast without losing control of what you’re building. CHAPTERS 00:00 – Cold Open 00:43 – Jason’s First Marathon and the AI Agent That Ordered His Carbs 04:42 – 2026: The Year AI Agents Hacked Real Companies 06:18 – How Worried Should We Be? 07:35 – Inside the Incidents: RubyGems, Gemini, Medicare and More 12:09 – Emergency Brakes: Who Gets to Pull Them? 15:40 – Will the AI Labs Eat Your Business? 19:28 – Is the Age of the Consultant Over? 20:49 – The $23 Million Navier-Stokes Heist 23:52 – Has a Robot Run a Marathon Yet? 24:50 – Founder Mode Top Five

I've spent the last 18 months watching people interact with AI tools. The single best predictor of who's winning isn't technical skill. It's not budget. It's not team size. It's curiosity. There are three types of people in every organization right now: Type 1: The Overwhelmed They know AI is important. They're vaguely anxious about it. But they don't know where to start, so they wait. "I'll figure it out when things settle down." Things aren't settling down. Type 2: The Dabbler They've tried ChatGPT. Thought it was cool. Used it a few times, then went back to how they used to do things. They mention it in meetings. They're not using it daily. Type 3: The Obsessed The moment a new tool drops, they've already tried it. They're running agents, testing builds, breaking things on purpose. Every week they find two or three things that save them hours. They're not waiting for permission. The gap between Type 2 and Type 3 is compounding every month. Because every tool you adopt early means you're a year ahead of the people who adopt it late. The engineers running Cursor two years ago are miles ahead of the ones just now learning Claude code. Curiosity isn't a personality trait you either have or don't. It's a practice. Go try the new thing. Break it. Ask "what else can this do?" That question is worth more than any technical credential right now. Which type are you right now, and which type do you want to be?

Company culture isn’t what you say you believe. It’s what you’re willing to stand behind when it gets difficult. This week on Founder Mode, @clairewaves joins @jasonshafton and me to share what 12 years inside Google and YouTube taught her about culture, leadership, and the gap between a company’s promises and its actions. She helped tell Google’s story, believed in its mission, and later helped organize a walkout when that story no longer matched what she was seeing. It’s a reminder that building a great company isn’t just about creating a mission people believe in. It’s about making sure your decisions still reflect it as the company grows. CHAPTERS 00:00 – Cold open 00:41 – Is culture load-bearing or just marketing? 04:07 – Selling meaning as a perk 05:42 – Ghostwriting Larry Page and the cure-death era 08:35 – Google+ and the corporate bonus 10:29 – When the story cracked: Creative Lab 13:18 – YouTube and the brand vulnerability briefing 15:54 – Was advertising the real mission all along? 18:10 – Marketing to executives instead of the public 21:31 – Maven, the walkout, and why workers went quiet 25:21 – Would she organize the walkout again? 28:08 – The subway ride home 32:06 – Where to find Claire and Don't Be Evil 33:13 – The Founder Mode Five