Product at Heart’s Post

My most valuable Claude workflow is turning an early product idea into something testable before it reaches engineering. That includes clarifying the problem, challenging assumptions, developing acceptance criteria, and creating a working prototype for stakeholder feedback. The real value is not faster documentation. It is, reducing ambiguity, removing friction between product, stakeholders, and engineering, and improving decisions before they become expensive downstream. It may also reduce the number of meetings needed to create alignment, creating a measurable cost benefit of its own. If you are in Orange County, I encourage you to attend the hands-on Claude Cowork workshop with OC Product Managers next week. I’m interested to see what other product leaders are learning, so please share your notes and takeaways afterward.

Our list of the top 33 AI Product Management books. The best AI PM books in 2026 are not about prompts... They are about product context. We reviewed 33 titles for the PM Lab. The strongest books focused on discovery, governance, delivery stability and adoption. Very few treated prompting as the core skill. The pattern was clear. Coding agents speed up execution. The bottleneck is everything before and after. What the team decided. Why the spec changed. What the code can actually support.

I care as much about the craft as I do about the business. A few weeks ago I shared Sana, my personal OS. The feedback went way beyond what I expected. So here is an update. 🙂 Last week at INSEAD, discussing AI with fellow product leaders, one conclusion kept coming back: the cost of building has collapsed. What has not collapsed is taste. Proof in the room: Mark Hull built his chief of staff behind a full web interface. Hugo Geissmann runs his headless, on Telegram, because he is always on the move. Same technical backbone. Radically different products. 👉 That gap is product taste, shaped by different needs. As a product builder, I want Sana to feel playful. Technology should serve the human. And I believe a terminal window is not the future of user experience, at least not mine. 😉 The little face in the video is Sana. A moon-faced companion that follows your cursor. A wink to G. Méliès and his film A Trip to the Moon, and to Notion's blinking characters (shout out, Notion is where Sana's memory lives). Under the playfulness, serious engineering. ⚙️ A persistent brain, a web app in progress, a headless worker running the scheduled jobs. A precompute pipeline digests my signals ahead of time (Slack, meetings, Notion, calendar): status questions become instant cache reads, only time-sensitive ones trigger a live run. And Sana learns from me: every correction feeds back into its memory and skills, so the noise drops week after week. Nothing ships without my green light. Most of it draws on principles Daniel Hertz from Anthropic shared: move repetitive work into deterministic scripts, keep orchestration outside the model, treat context as infrastructure. And keep the friction that is good. When making gets cheap, taste and care are the differentiators. What does your assistant look like, and why? ✨ #AIAgents #ProductCraft #BuildInPublic #PersonalOS #ProductManagement

You can read a conference program and still not know if it's for you. So here's a better signal: the people already talking about it. In the weeks before Productlab Conf26, product people across the community have been posting their own reasons to come, tagged #ProductlabConf. 🏆 And worth award them! Read together, they are a live map of where the craft is moving. Five operators, five angles: Julie Kamel: "AI will not save your product. It will expose it." AI amplifies the model you already run. Strong orgs get faster; weak ones fail faster, in higher resolution. Backlog administration is about to become very visible. Miruna Popa reframed the idea glut. Ideas were always plentiful. The real skill is experiment design: parallel tests, real success metrics, the discipline to stop when the data says stop. Her line: the big bets are easier to measure than the anxiety tells you. Alexander Tikhomirov, building Cope Pilot, on the aha most people miss: good principles aren't fairy tales, they come with concrete instructions. Deliver. Earn trust. Run as many experiments as you can to find product-market fit fast. Simple to say, hard to live. Harry James Coburn named it plainly: AI made it easier than ever to ship, and harder than ever to ship something that matters. Output is cheap. Outcomes are everything. He shows up for the honest conversations you can't get elsewhere. Subir Paul traced the job back to McElroy's 1931 memo: brand advocate, market analyst, business owner. As AI absorbs the admin, product returns to what it was always meant to be. Not a delivery function. None of them on a stage yet. All of them worth listening to. That is the point of Productlab Conf26, September 15-17 in Berlin. The program starts the conversation. The community shows you where it's actually going. Who else has been posting? Add them below.

NavoPM's upcoming roadmap isn't just about new features — it's about reshaping how AI fits into the full product management workflow. Here's what's coming and why it matters for PMs navigating an AI-first world.

Tomorrow at 12 PM ET! AI is changing product management, but what does that actually mean for how product teams work? Join Jeff Keyes as he breaks down the shift from Build to Learn to Build to Earn, the new operating model for modern product teams, and the skills PMs need to thrive in the AI era. 📅 July 16 | 12 PM ET Reserve your spot before it's too late. 👇 https://lnkd.in/gE44HyEk 

Summary In this episode, Jay Stansell interviews Dave West from Scrum.org about the evolving role of AI in product management. They explore how AI acts as a teammate, its limitations, and its impact

🚀The next challenge for AI-native teams isn't finding tools. It's integrating them into something coherent. Our last In Product Meetup 'The End of Hand-Offs: AI Native Teams' hosted in collaboration with AutogenAI brought together the views of leaders all working in different contexts and facing different challenges 💜 🚨One theme shared among the panellists regardless of environment was the importance on building a shared language. Eugene Kuznetsov Product Lead at Autogen AI shared the importance of shared language not just between people - between people and the systems they're working with. ➡️His team built a library of historical decisions. Every time they spot an inconsistency they want to change, they update it. It becomes a living skill - something the AI can query, something engineers and PMs can both reference. ✅The result: less time spent re-explaining context. Reviews and communication still matter, but the foundation is more stable. He also made a point that's worth sitting with: five months ago, a specific problem in their space had no good solutions. Now there are five tools trying to solve it. The next challenge isn't finding tools - it's integrating them into something coherent. The teams that build that integration layer, even informally, will move faster than the ones still shopping for the perfect tool. 💜 #inproduct #ainative #productteams

🚀The next challenge for AI-native teams isn't finding tools. It's integrating them into something coherent. Our last In Product Meetup 'The End of Hand-Offs: AI Native Teams' hosted in collaboration with AutogenAI brought together the views of leaders all working in different contexts and facing different challenges 💜 🚨One theme shared among the panellists regardless of environment was the importance on building a shared language. Eugene Kuznetsov Product Lead at Autogen AI shared the importance of shared language not just between people - between people and the systems they're working with. ➡️His team built a library of historical decisions. Every time they spot an inconsistency they want to change, they update it. It becomes a living skill - something the AI can query, something engineers and PMs can both reference. ✅The result: less time spent re-explaining context. Reviews and communication still matter, but the foundation is more stable. He also made a point that's worth sitting with: five months ago, a specific problem in their space had no good solutions. Now there are five tools trying to solve it. The next challenge isn't finding tools - it's integrating them into something coherent. The teams that build that integration layer, even informally, will move faster than the ones still shopping for the perfect tool. 💜 #inproduct #ainative #productteams

AI in product management sounds great on paper. But what actually happens when teams adopt it? The gap between vendor promises and on-the-ground reality is where most PMs live—and it's where the real story is. Here's what teams are actually seeing...