Product School

💡Introducing Product School PRO – Learn more skills and save $2,000 💡 PRO is our new offer that gives you access to any two certifications within one year – so you can build the combination of AI skills that fits your goals and your career trajectory. Plus, you will also: -> Save $2,000 compared to purchasing two individual certifications. -> Get access to on-demand content. -> Get access to the alumni community. -> Showcase Member status on your LinkedIn profile. Coming soon: Member-only AI workshops to complement your learning. Single certification is great for becoming better at your current position. PRO is great for learning a whole suite of AI skills and setting yourself up for a promotion 🚀 Sign up today to ensure the best year of your career to date: prdct.school/4t5Z2mW

The most important takeaway: Old scarcity thinking guarded engineering cycles – agents demand an abundance mindset: Spend the tokens. Build both prototypes. Learn in production. Treat agents as teammates in discovery, definition, delivery loops. Split big teams into subsets with agents filling in – that's how you move, iterate, and ship faster.

"AI is not magic. It doesn't make problems go away. We're just shifting where the bottlenecks occur." @BrentBarkman, VP of Product at @MiroHQ, gave a memorable presentation at ProductCon London, explaining how a unified workspace where humans and AI agents collaborate as teammates is the future of productivity Watch the full talk now: youtube.com/watch?v=0bw0W9…

That's a wrap! Thank you for following us 🧠

We relaunched our Substack Newsletter! Want to get the new podcast episodes straight into your inbox every week? Subscribe here: productschool.substack.com

This week in Product School 👇

Diya Jolly built monetization at YouTube, led product at Okta, and now runs product + tech at @Xero (4M+ customers) On #TheProductPodcast with @villaumbrosia, she shares PM fundamentals that hold from consumer to enterprise, plus her AI strategy for regulated SMBs Watch the full episode: prdct.school/4lSUFJr

Xero CPTO on Building an Agentic AI Platform to Manage Multiple Agents x.com/i/broadcasts/1…

Apply via this form: docs.google.com/forms/d/e/1FAI…

Volunteer Applications Now Open for ProductCon New York 2026! 🗽 If you want to be a part of the leading AI Product conference, this is your chance! As a volunteer, you'll get: ✅Free Access: Attend the conference sessions and get to hear insights from the industry experts ✅Premium Networking: Connect with top product leaders, speakers, and fellow attendees between sessions and at the official after-party ✅Food & Drinks: Enjoy complimentary breakfast and lunch Don't miss this opportunity to be a part of something amazing! Act fast – positions are filling up as you're reading this ⭐️

There's a PM making decisions about your product right now, that you never hired and can never fire. And they're talking to your customers millions of times a day. It's AI. @p_fabrikantov, SVP of Product at @semrush, calls it your Shadow PM. The middle of the funnel has disappeared. Old world: user searches → clicks 10 links → compares on Reddit → weeks of deliberation → decision. New world: user asks AI → AI picks one brand → user confirms → buys within hours. Your positioning is no longer fixed. It's fragmented across systems you don't control. You can't optimize for "AI" as a single channel any more than you could optimize for "the internet." Three things to do right now: 1. Audit per model, not per "AI" Run 50 core queries across 4–5 platforms. Track whether you're mentioned, whether it's accurate, and who's winning instead of you. This is your new competitive landscape. 2. Optimize content for humans but make in machine-friendly Most AI crawlers don't render JavaScript. Your JS-heavy landing pages are invisible to them. Unblock AI crawlers. Server-render your content. Structure docs in short-answer-first chunks. 3. Win in earned media 77% of brand citations in AI come from third parties – Reddit, G2, LinkedIn, YouTube, industry reviews. Seed these sources with use cases, methods, and real outcomes. AI-referred users have 27% lower bounce rate: they arrive with the right expectations already set by your Shadow PM. You can't plug off your Shadow PM, but you can plug in your product and your knowledge base. 📺 Full talk from ProductCon London 2026: youtu.be/8LlYrAGb56k?si…

The full conversation with Vanessa is now available on our website: prdct.school/4772zJc

3 most important things we've learned from #TheProductPodcast with @vlaurenlee and @villaumbrosia: On the shifting role of PMs: The job is no longer about writing airtight specs and hardening every edge case before engineering touches it. It's shifting toward something harder to define – and harder to teach. Judgment. Courage. The ability to make decisions with incomplete information and bring teams with you. On how Shopify evaluates AI quality: When you're building non-deterministic features, your old definition of "good" (the PRD) doesn't work anymore. Vanessa's team built an internal LLM judge – where PMs define the dimensions of quality, then rate conversations, then use those ratings to train the model on what good looks like. The PMs spend more time teaching the system than they spend writing specs. On courage: She had a "screw it" moment early in her career – walked into a meeting with the CEO before ever being in a small room with him, armed with domain knowledge, a hand-drawn diagram, and a clear point of view on an API versioning decision that entire teams were afraid to raise. And she turned "we're not doing it" into a roadmap in 45 minutes.

Shopify VP of Product on Building a $100B+ Agent-Led Commerce x.com/i/broadcasts/1…

Sign up for the Unlimited Membership and enjoy one year FREE of Replit Core: #our-plans" target="_blank" rel="nofollow noopener">productschool.com/certifications…

3️⃣ “Day in the life” simulations for new personas Use Agent 4 to simulate workflows for a new persona you want to serve, including tools, touchpoints, and failure modes Prompt idea: “Create: 1) a simple web app that represents a ‘day in the life’ for [persona] using our product, 2) three scenarios (happy path, time pressure, system failure) with click-through flows, and 3) a short internal training deck that explains these flows to sales and support.” Then, walk stakeholders through each scenario live and capture where the flows break or feel unrealistic.

Proud partner moment: @Replit, a tool that takes PMs from idea → design → runnable app in hours, has just raised a $400 million Series D at a $9B valuation 👏 AND, they've just released Agent 4, the latest version of the AI engine, which helps you stay in a creative flow while it handles the implementation. It's included in our Unlimited Membership, and here are 3 ideas to test its new capabilities: