Essay 01
The Cannibalization Playbook: How CPOs Win the AI Transition
The honest CPO move when AI replaces your SaaS: set a sunset, kill the comp asymmetry, pre-sell the gross-margin trough. The seven-decision sequence.
The Product OS · the CPO seat
Own the whole outcome.
The decisions behind the AI-era product organization: one loop, one owner, the org, the margin, and the board story.
What this seat owns
The whole product outcome, from the customer problem to adoption, and the org, the margin, and the board story that go with it.
When building is cheap, the outcome is decided by the loop from customer to adoption. One loop needs one owner. The CPO owns the whole product outcome, and product, design, product marketing, and forward deployed engineering run it together. See the whole OS.
- 01 · leadsCustomerField signal from deployments, support, and usage, read daily. Measured in hops and days from finding to fix.
- 02 · leadsSpecThe problem, the behavior, the claim, and the eval that says it works. Measured by the eval pass rate.
- 03 · leadsBuildAgents and Product Builders take the spec to working software. Measured by the repair rate in use.
- 04 · leadsStoryPositioning and proof that buyers, sellers, and AI assistants read, shipped in the same release as the build.
- 05 · leadsAdoptionThe one outcome number moves, or the loop runs again.
Start here
The argument, in 4 essays.
Written for product leaders. Read them in order, or start with the question on your mind.
Essay 02
The Dual Transformation Operating Model
Two clocks, one product org. Run a legacy SaaS on a six-week rhythm and an agent-native successor on a one-week rhythm without tearing the team apart.
Essay 03
The Pricing Migration Sequence: An 18-Month Quarterly Playbook
Per-seat is dying, but it pays your payroll. The quarter-by-quarter playbook for moving $20M ARR to hybrid outcome pricing without crashing NRR.
Essay 04
Context Is King. Written Context Is Rented.
Everyone is racing for the context layer. The quadrant nobody occupies is empty for a structural reason, and importing your wiki will not fill it.
The handbook
12 chapters for this seat.
How to do the work, chapter by chapter. Every seat starts from the same core.
- Core · Why This Exists
- Core · The AI Product Operating Model
- Core · Engineering Builds the Substrate, Not Features
- Core · The Eval Is The Spec
- Core · Kill the Roadmap
- Core · The Impact Loop
- 01The CPO 30/60/90The operational week-by-week plan for a new CPO: four audits, a trust ledger, a coalition map, and a day-90 readout the board will quote. With templates.10 min read
- 02Strategy From Signals, Not SlidesThe annual strategy deck is a memorial to a meeting. Run a one-page living strategy doc, updated weekly with the signals that could change your beliefs.8 min read
- 03Dual Transformation: Running Two ClocksTwo cadences, three talent categories, six CEO scoreboard numbers, three rituals. The operating model that runs legacy and successor cleanly.8 min read
- 04Gross Margin Is Your Job NowCost per successful action is the new primary PM metric. If you don't own it, your CFO will kill your product before your customers do.7 min read
- 05The Investor and Board Narrative for AI-Era SaaSGross margin compression, NRR redefinition, the new rule of 40, the right comp set, and the four CFO sub-agreements that make the trough boring.4 min read
- 06The Cannibalization Decision FrameworkFour diagnostic questions, three operating modes (sunset, refresh, split), the political map, and the seven-decision sequence for the AI inflection.9 min read
- 07Trust, Safety, and the Guardrail as a Product DecisionEvery guardrail is a product decision. The PM who outsources it to legal gets a product they didn't design and a customer experience they wouldn't approve.8 min read
- 08Incident Response Is a PM RitualAn incident is a customer telling you the truth about your product, loudly, all at once. Stop letting engineering listen alone.7 min read
- 09Kill the Status MeetingThe status meeting exists because nobody trusts the dashboard. Fix the dashboard once. Stop paying the tax weekly.7 min read
- 10Your Weekly PlaybookWhat a week looks like when you run the full handbook. The week is the rhythm, the loop is same-day, and you run it as often as hot signals land.11 min read
- 11Hiring the Builder PMMost PM hiring loops test skills the role no longer needs. Hire by what they can ship, not by how they talk about shipping.8 min read
- 12The Skill Stack: What PMs and CPOs Must Learn NowThe 12 skills that decide PM and CPO careers now, organized in five layers from judgment to storytelling, with a self-test and learning path for each.10 min read
The running series · New parts every Thursday and Saturday, part of The Product OS
The CPO Transition
Part 7
Atlassian Says Its FDEs Aren't a Services Business. Put the Field Under the CPO.
Atlassian says its FDEs speed customer request to delivery. HubSpot's agent resolved 2M issues. The field's signal is product work. Put it under the CPO.
Part 6
Canva Merged Its Product Org. Finish the Job: Put the Whole Loop Under the CPO.
Canva merged product, growth, and platform into one org to end fragmented ownership. Finish it: design, PMM, and FDE report to the CPO too.
Part 5
Anthropic's S-1 Has No Gross Margin Line. Your AI Product Needs One.
Anthropic's S-1 books compute as opex and shows no gross margin. Where inference sits, and who keeps a 40% price cut, are now CPO decisions.
Part 4
Anthropic Runs 17 Engineers per PM. The Ratio Is the Wrong Question.
Anthropic runs about 17 engineers per PM and junior tech hires fell from 27% to 9%. The CPO's org question is spans, layers, and who builds the bench.
Keep reading
More arguments.
- Field Report: 90 Days Inside a SaaS-to-Agents TransitionWhat we believed, what we did, what broke, what we changed. An anonymized after-action report from the first 90 days of a $30M ARR cannibalization.
- Field Report: What Broke When We Killed Our Per-Seat TierMigration drift, an unexpected churning cohort, a unit definition that broke at scale, dispute volume at 12x projection. After-action from the sunset day.
- Why "Soft Pivots" Always Become Two Companies in One OfficeEvery CPO running a 'gradual transition' to AI ends up running two companies inside one org chart. The five patterns and how to escape them.
- Continuous Discovery Doesn't Scale for AI-Native ProductsTeresa Torres' continuous discovery is the right answer for human-centric SaaS and the wrong answer for agent products. The punctuated discovery alternative.
- Outcome Accountability Is a Luxury Good. Measure Direction.Outcome-driven roadmaps assume 6-12 month measurement cycles. Agents iterate ten times a week. The dual-cadence direction-metric system that closes the gap.
- The Triad Is Dead. Pods Are Dead. Agent Departments Now.Forrester still talks about human-with-agent teams. Wrong unit. Agents form their own departments and interface with human departments on exception.
- The Jevons Cliff in Outcome PricingInference costs drop 50% per year. Your $0.50/resolution price is a moving target. Hold or pass? The framework nobody has published until now.
- Cagan's Definition Is Right. The Weights Are Wrong.Cagan maps Benedict Evans's three product skills onto his risk framework. The skills are right. Their AI exposure is wildly different, and that is the story.
- Should customer success report to the CPO?Half of it. AI splits customer success into two jobs. The commercial half, renewal, expansion, and the buyer relationship, stays with the CRO. The adoption half, onboarding, the support agent and its transcripts, and the engineers who fix live deployments, reports to the CPO next to the forward deployed engineers, because each piece changes what the product does. HubSpot said its Customer Agent resolved over 2 million support issues in Q3 2026, which makes support a product surface, and Atlassian said in November 2026 that its FDEs are not a services business and speed up the customer request to delivery.
- What does Canva's one Product Org mean for who reports to the CPO?It means fragmented ownership is the AI-era failure mode, and the fix is one owner for the whole loop. Canva is merging product, growth, and platform into one Product Org in a restructuring that touches about 22% of the company, roughly 1,271 people, because, as Melanie Perkins's memo puts it in Startup Daily's reporting, divided work left different teams responsible for many pieces of the same work. Finish the job: product management, design, product marketing, and forward deployed engineering report to the CPO, while brand, revenue, platform reliability, and finance stay where they are.
- Where should AI inference cost sit on the P&L?Put the cost of serving customers in cost of revenue, split by feature, and leave training, evaluation, and internal experiments in operating expense. Otherwise gross margin flatters the product and nobody who designs the feature sees what a successful action costs. Anthropic's S-1, filed October 26, 2026, books $8,945 million of compute and infrastructure as operating expense against $16,286 million of first-half revenue and reports no gross margin; Tomasz Tunguz derived 85 percent from a footnote, while subtracting the whole compute line gives about 45. Where one cost sits moved the answer 40 points.
- Should product orgs still hire junior PMs when AI does the entry-level work?Yes, deliberately and in smaller numbers, paired with seniors. The entry rung is thinning everywhere: Anthropic's junior share of new technical hires fell from 27 percent to 9 percent (October 2026), and Aakash Gupta's 2026 report shows L1 fell from 17 to 11 percent of Levels.fyi PM reports. Every future senior PM and CPO starts as an L1, so a product org that hires no one at the bottom has no senior bench in about six years. Atlassian chose the opposite, weighting hiring toward juniors and seniors with fewer in the middle, because juniors use AI more and seniors catch low-quality output.
- Who owns the eval harness, the CPO or the CTO?Split it. The CTO owns the platform: scaffolded environments, the model gateway, the harness that runs evals, deploy isolation, spend metering, and the enforcement of guardrails. The CPO owns the bar: what the eval set says good looks like for a customer, the guardrail policy for what the product may do or spend on a customer's behalf, the graduation criteria, and the cost-per-outcome target. Both sign the gate to production. The split matters because agents made code cheap and left verification expensive: in an October 2026 survey of 300 engineering leaders for Undo, 79 percent said code is faster and releases aren't, and 35 percent of generated code reached production before the team understood it.
- What should a CEO hold a CPO to on AI?Four things, and none of them is a roadmap. A growth measure named before each AI bet ships, because hours saved, velocity, and adoption tell a CEO that product got faster, not that anything a customer cares about moved. Cost per outcome, reported next to the outcome. A straight answer on how product's own work has changed, since the October 2026 Wharton and GBK Collective study found only 16 percent of Product/Engineering departments have AI built into core processes, the lowest of any department. And one bet killed every quarter, with the reason. The CPO's side of the deal is one sentence per AI bet: the customer outcome, the evidence, the cost per outcome, and what was stopped to fund it.
Part of The Product OS
One loop. Five seats.
This seat runs one part of the loop from customer to adoption, under the CPO. See how the other seats connect to it.
See the whole Product OS