by Stefan Wolpers|FeaturedAgile and ScrumAgile Transition
TL;DR: The AI Delegation Audit Webinar, October 6
Join me on October 6 for the AI Delegation Audit webinar. You will learn how to run a 45-to-60-minute recurring check that reveals whether delegated AI work still meets its required quality standard, still uses the right model, at the right cost, can still be stopped, and hasn’t quietly become more autonomous than intended. (Rogue AI is no longer science fiction, isn’t it?)
TL; DR: Guide to Agent ROI — Food for Agile Thought #559
Welcome to the 559th edition of the Food for Agile Thought newsletter, shared with 35,351 peers. This week, Chandana Asif and colleagues find that human oversight, not tokens, defines agent ROI, so redesign the workflow. Cleo Lant pushes you to disrupt yourself first, and Sangeet Paul Choudary warns operational excellence can accelerate irrelevance once scarcity moves. Taylor Belrose denies AI any moral status, while Bill Gates wants new institutions for a shift that substitutes for cognition. Also, Jason Knight and Barry O’Reilly ask you to expand judgment, not output.
Next, Stephanie Leue argues roadmap overrides come from invisible trade-offs, not from Sales holding power, and Mike Fisher shows where blindness ends: the sea squirt digests its own brain once it settles, much as companies defund customer research. Tim O’Reilly, answering Ted Chiang, calls AI a medium that rewards craft, and Addy Osmani situates that craft in intent and architecture. Also, Johanna Rothman narrows attention further, running one Retro experiment at a time.
Lastly, Chad McAllister and Brooke Rennison show BMW routing ideas from 40,000 employees through stage gates, prototypes, and university partnerships. Anthropic’s release of 250,000 Claude conversations to outside researchers finds people directing and editing, with friction improving results, and Sachin Rekhi warns those individual gains evaporate without a shared platform. Finally, Gustavo Razzetti traces conflict to agreements nobody made, while Mark Graban asks who received your last ten speak-up reports, and what happened.
TL; DR: 10 New Product Owner Interview Questions on AI
“How do you use AI in your work?” If that is still one of your Product Owner interview questions, you are screening for tooling fluency in a role that lives or dies on product judgment. Every candidate has an answer; LLMs make sure of that. None of those answers tell you whether the person can decide what is worth building when building is no longer the hard part.
These ten new Product Owner interview questions are designed to expose the gap. Each one has a seductive wrong answer that sounds smart enough to pass a surface-level interview. The strong answers require something a blog post cannot teach: lived experience with the judgment calls that the Product Owner role now demands, as three shifts justify the addition: the collapse of code generation cost, the rise of product operating models that rename existing practices without rewiring, and the quiet fact that AI is already reshaping product decisions in organizations that made no deliberate decision about it.
Thesis: When building is no longer the constraint, every Product Owner interview question must test the candidate’s judgment about what is worth building, not their fluency with the tools that build it.
TL; DR: Transformation Failure — Food for Agile Thought #558
Welcome to the 558th edition of the Food for Agile Thought newsletter, shared with 35,366 peers. This week, Nigel Thurlow and Stephanie Leue tackle leadership: Thurlow analyzes patterns of transformation failure, showing that leaders seek change only when they hold power, while Stephanie Leue proposes that steady presence, not control, builds authority. That maturity matters because, as Roman Pichler warns, bolting AI onto broken practices accelerates dysfunction. Andreas Horn reveals how agentic AI costs spiral unchecked, John Gruber calls Anthropic’s text watermarking a sacrifice of clarity, and Aaron Horwath asks what happens when AI strips knowledge work of meaning.
Next, Eddie Pratt warns that PMs who use AI solo build “reasoning silos,” while David Pereira proposes that decision-making, not building, is the real bottleneck. Pavel Samsonov agrees: velocity without judgment ships net-negative software faster. Governance lags, as VB Staff reports that 21% of enterprises lack cost controls despite running an average of three orchestration platforms, and Zvi Mowshowitz finds Anthropic’s safety case weaker than advertised. Also, Rudrendu Paul and Sourav Nandy urge B2B sellers to redesign for AI agent buyers.
Lastly, Jenny Wanger proposes that influence grows from trust and a readiness to change your own mind, a theme Ryan Murphy extends: good managers coach rather than control the room’s mood. On the tooling front, Paweł Huryn compares four AI prototyping platforms for PMs, and Avi Chawla walks through SpaceXAI’s Grok Bot, which features persistent cloud agents. Finally, John Cutler warns that “return on tokens” risks becoming the next proxy trap, echoing story points.
Your organization counts AI tokens, seats, and pilots, but can anyone name a single decision those numbers actually changed? Tokenmaxxing is only the symptom; five old Agile Laws explain the cause, and each one comes with a test you can run this week. There is no need to reinvent the wheel with AI transformations and learn the hard way what the veterans of other transformations already figured out.
Thesis: Tokenmaxxing is the vanity metric of pushing low-value work through an AI tool solely to inflate usage metrics. Tokenmaxxing emerged in 2026, when large technology companies began ranking employees by token consumption on internal leaderboards. The behavior is rational for the individual but useless for the organization because tokens measure input rather than outcomes. The five Agile Laws in this article explain why organizations keep making this mistake and what to measure instead.
TL; DR: AI Watermarks — Food for Agile Thought #557
Welcome to the 557th edition of the Food for Agile Thought newsletter, shared with 35,378 peers. This week, Anthropic details the AI watermarks and C2PA metadata Claude now attaches under the EU AI Act, while Dror Poleg shows machines spot each other through word statistics that every writer carries. Detection remains messy, and so does judgment: Marty Cagan revisits Benedict Evans on problems AI never solves; John Cutler swaps prioritization frameworks for 12 tension prompts, and Vasuman suggests counting automated work rather than AI adoption based on vanity metrics. Also, Nigel Thurlow traces delay to variation rather than to people.
Next, Johanna Rothman counts running tested features, not activity, and pushes teams to finish aging work before inventory eats money. Ant Murphy goes further: value appears only after delivery, so prioritization interrogates confidence; remember ‘thinking in bets?’ Then, Itamar Gilad warns AI produces work nobody can judge, blurring roles and inflating certainty; Tim O’Reilly unpacks Drew Breunig’s prompt debt, which traps teams on old models, while Sunil Pai’s Cassandra agent watches Slack and speaks only when consensus looks wrong and rocking the boat seems appropriate.
Lastly, Matthew Hodgson notices agents spend by the second while budgets renew yearly, so he wants a per-agent cap now, while Meryem Arik finds similar waste in inference bills. Mark Levison keeps the human ledger: GenAI-triggered human skill loss is a choice, so pick which skills must stay sharp. Finally, Sebastian Ankargren, Joel Persson, and Mårten Schultzberg show LLMs replace A/B test users only under unverifiable assumptions.
