The Cannibalization Playbook: How CPOs Win the AI Transition

Most CPOs I talk to in 2026 are treating the AI shift as a soft pivot. We're "evolving the product line." We're "adding agentic capabilities." We're "running a hybrid pricing model." We're "letting the old and the new coexist."

It isn't a soft pivot. It's a cannibalization decision, and pretending otherwise is the most expensive mistake a product executive can make right now.

This is the playbook I've been running, the one I've watched competitors and peers run badly, and the one almost nobody in the public conversation will publish because it's career risk to say out loud. I'm going to say it out loud.

The short version

The soft pivot is a fantasy. Per-seat revenue does not gradually migrate to outcome revenue. It splits into two business systems that fight each other inside the same org chart. Sales gets paid on one model and the new product earns on the other. CS is staffed for the legacy product and underwater on the new one. Roadmap reviews become turf wars. Customers see two motions and trust neither.

The honest move is to set a sunset date on the legacy product, invest all product innovation in the agent-native successor, run the legacy for cash for 18-24 months, and tell the board the gross margin trough is coming before it arrives. Every company that has won a platform shift in the last fifteen years (Adobe Creative Cloud, Microsoft 365, Salesforce post-Slack) ran a version of this. Every company that lost ran a "soft pivot."

This essay is the seven-decision sequence I use, the political map for getting it through, the revenue cliff math you owe the board, and the one CFO conversation that decides whether the transition works. End to end.

The two systems problem

Imagine you run a $30M ARR SaaS business with healthy 110% NRR. You launch an agent-native version of your product. The agent product is priced per outcome and starts at zero. Three quarters in, you have $1M in agent ARR and your team is exhausted.

Look at what's actually happening inside the org. The legacy product still drives 96% of revenue, so the CRO is still pushing seat-based renewals. Sales comp is still tied to ACV, which is a seat math. CS is staffed for ticket volume that scales with seats. The agent product needs entirely different sales motions: it sells to a different buyer, has a different sales cycle, and earns variable revenue your CFO has never modeled. Engineering is being asked to maintain the legacy and ship the new product simultaneously, and the math says they can do half of each, badly.

Then the legacy product starts to soften. NRR drops from 110% to 102%. The CRO insists on more seat sales, which means more demos, which means more sales engineering, which means engineering attention pulled back to the legacy product. The new product slows down. The competitor that was a year behind catches up.

This is the two systems problem. It is the default outcome of the soft pivot, and I've watched it happen at three companies in three industries in the last eighteen months.

The two systems problem ends one of two ways. The legacy product wins, the new product gets quietly killed, and the company gets displaced from below by a startup that didn't have the legacy to defend. Or the legacy product loses, the company eats a 30-50% revenue drop in one bad year, and the CPO gets fired for "cannibalizing the business."

There's a third way, but it requires deciding to do it.

The decision: sunset, refresh, or split

Three options when the AI inflection arrives at your product line. Each has a profile and a price.

Sunset. The legacy product is end-of-lifed on a published date 18-24 months out. All product investment redirects to the successor. Existing customers are migrated, churned, or harvested for cash until sunset. This is the right move when the legacy architecture cannot scale to support the successor's quality bar, when the legacy customer base is the right ICP for the successor, and when the company can afford the gross margin trough.

Refresh. The legacy product is rebuilt with AI capabilities while keeping the same pricing and positioning. Customers experience the change as upgrades, not a new product. This is the right move when the legacy unit economics still work, when the buyer has not changed, and when the AI capabilities augment rather than replace the underlying job-to-be-done.

Split. The legacy product and the successor live as separate product lines under one company. Different teams, different sales motions, different P&Ls. Eventually one of them gets divested or wound down. This is the right move when the buyer is genuinely different (legacy sells to IT, successor sells to Operations) and the legacy customer base is too sticky to sunset cleanly.

Most CPOs default to refresh because it's the least scary. In 2026 it's increasingly the wrong answer. Refresh works when AI is a feature improvement. It fails when AI is a business model replacement. If your successor product is priced on outcomes and the legacy is priced on seats, you are not doing a refresh. You are doing a sunset and pretending it's a refresh, and your customers know.

The honest test: can you publish a single roadmap that contains both products without lying? If yes, refresh works. If no, you owe yourself the harder conversation.

The seven-decision sequence

Once the call is sunset (and most of the time it is), here's the order the decisions need to come in. Out-of-order is the most common reason I see this fail.

1. Decide the sunset date

Pick a date 18-24 months out. Less than 18 months and the customer migration becomes a fire drill. More than 24 months and the team's urgency dissolves and engineering attention drifts back to the legacy product.

The date is published internally first, externally later. The date is the forcing function. Every decision downstream gets evaluated against "does this still work given that we end-of-life on March 31, 2028?"

2. Reorg into Successor and Maintenance

One product team owns the successor. A smaller maintenance team runs the legacy. The maintenance team has a different mandate: keep the lights on, fix critical bugs, support migration tooling. They do not ship features. They do not own roadmap.

The split is brutal because your strongest engineers want to be on the successor and your most loyal engineers want to maintain what they built. You have to make calls neither group will love. Lay the calls out as part of the sunset announcement, not as a surprise three months later.

3. Rewrite sales comp before anything else moves

Sales is the function most likely to torpedo the transition unintentionally. They will keep selling what pays, and what pays is the legacy product on the legacy comp plan.

Rewrite the comp plan in Q1 of the transition. New deals on the legacy product earn 50% of historical comp (you want the salesperson to think about the outcome alternative). New deals on the successor earn 150% of equivalent ACV (you want them excited about it). Renewal accelerators on the legacy disappear. Migration accelerators on the successor are introduced.

This is a CRO decision, not a CPO decision, but the CPO has to be the one explaining why the comp change is necessary. Otherwise it gets watered down by Q2.

4. Pre-sell the board on the gross margin trough

The board will see gross margin compress from 80% (mature SaaS) to maybe 55% (early Service-as-Software during ramp). This will look like a disaster on the standard SaaS dashboard. It is not a disaster. It is the expected shape of the transition.

Walk the board through the math before it shows up on the financials. Show them: legacy revenue declining at X%, successor revenue growing at Y%, blended gross margin trough at Z%, expected recovery to W% over 24 months. Get them to commit to the trough as a planned investment, not a surprise.

The CPOs I've seen lose their jobs in transitions like this lost them not because the transition failed but because they didn't pre-sell the trough. The board saw the margin compress, panicked, and pulled the plug halfway through.

5. Communicate to customers in cohort waves

Not a press release. Not an email blast. Cohort waves.

Wave 1 (Month 1): the largest enterprise customers get individual calls from the CPO and CRO. The message is: we are sunsetting on date X, here's the migration path, here's the pricing, here's what we will do to make this clean. They get bespoke commercial terms if needed.

Wave 2 (Month 3): mid-market customers get a structured email plus a webinar. They get a self-service migration tool. They get incentives to migrate before date Y.

Wave 3 (Month 6): the long tail gets the public announcement, the documentation, and a clear deadline. Self-serve migration only.

Most companies invert this and announce publicly first, which means the largest customers find out by reading a blog post. That single decision can cost you 20% of your strategic accounts.

6. Track the migration as one number

The CEO meeting and the board meeting need to track the same number every week: percent of legacy revenue migrated to the successor. Not as a side metric. As the number.

This forces the team to stop debating whether the transition is happening and start asking why it isn't happening fast enough. It also keeps the legacy team and the successor team from blaming each other for the trough; both can see the same percentage moving.

7. Plan the post-sunset reorg before sunset

Here's the one most CPOs miss. After sunset, you no longer need a maintenance team. You no longer need the salespeople comped on legacy seats. You probably don't need the same CS structure. The post-sunset org is leaner and shaped differently.

Plan that reorg in Month 12, six months before sunset. Communicate it transparently to the maintenance team: here are the roles that exist after sunset, here's how to interview for them, here's the timeline, here's the severance for those who don't transition. Treating people humanely during a sunset is both the right thing to do and the only way to keep the rest of the org motivated.

The political map

The seven decisions above don't move themselves. They move because a coalition forms behind them.

The coalition has five seats: CEO, CPO, CFO, CRO, CCO. Each one has a reason to fight the transition and a reason to support it. Knowing both is how you build the coalition.

CEO. Will support if the option value math is clear and if the board is pre-sold. Will fight if the transition feels like CPO empire-building. The way to win the CEO is to walk in with the seven decisions already structured, make the option-value argument, and ask "do you want me to lead this or do you want me to escalate to the board?" CEOs who hesitate at that question are not going to back you in the trough.

CFO. Will support if the gross margin trough is bounded and modeled. Will fight if the revenue model is variable and the forecast goes from "predictable" to "depends on customer volume." The way to win the CFO is to model both scenarios, propose hybrid pricing for the successor (committed minimum + per-outcome overage) so there's still a forecast floor, and offer to co-author the board narrative with them.

CRO. Will fight hardest. Comp is the lever. Renewal economics are the second lever. The way to win the CRO is to give them the upside math on outcome pricing (in many cases, 2-3x the SaaS price for the same customer) and to rewrite comp in their favor for successor deals. Make them the hero of the new motion, not the casualty of the old one.

CCO. Will fight if the CS team's metrics are tied to seat-based usage. Will support if the CCO can see the CS organization redesigned around outcome quality (which is the new job). The way to win the CCO is to give them a vision of CS as the team that owns outcome SLAs, dispute resolution, and customer migration. Make CS strategic again.

CPO (you). Will fight yourself, because cannibalizing the product line you spent five years building feels like an admission of failure. The way to win yourself is to write the option-value memo. The legacy product was the right answer for its time. The successor is the right answer for the next time. Both can be true. Loyalty to a product is not a virtue when the world has changed.

The coalition takes 90 days to assemble. Most of the work is one-on-one conversations, not meetings. By the time the seven decisions hit a leadership meeting, the calls have been made.

The CFO conversation

There's one conversation that decides whether the entire transition works. It's the conversation with the CFO about the gross margin trough.

The math is straightforward. Legacy SaaS runs 78-85% gross margin. Service-as-Software runs 55-70% in the first 18 months because inference cost is meaningful (15-25% of revenue) and infrastructure overhead has not yet been amortized across volume. The blended margin during the trough is somewhere around 60-65%, depending on the speed of the migration.

That number will look bad in any board deck that benchmarks against pure-SaaS comps. It is not bad. It is the expected shape of a Service-as-Software transition. The pure-SaaS comp set is the wrong benchmark.

The CFO conversation has three parts.

Part one: walk through the trough math. Show the curve. Show the recovery (most companies stabilize at 65-72% gross margin after 18-24 months as inference costs fall and volume scales). Show the comparison to peers who attempted this (Sierra hit $100M ARR in two years on outcome pricing; Intercom moved Fin to per-resolution at $0.99; HubSpot dropped to $0.50 per resolved conversation).

Part two: agree on the new comp set for board reporting. The right comp set for a company in transition is other companies in transition, not pure-SaaS leaders. Some of these comps are public. Some require footnoted analyst write-ups. Get the CFO's commitment to retraining the board on the comp set during the next quarterly review.

Part three: agree on the leading indicators. The board cannot wait for revenue to recover to know the transition is working. They need leading indicators that show the successor is healthy: outcome volume, customer migration percentage, gross margin per outcome, agent quality scores, NPS on the successor, expansion revenue from migrated customers. Define these together with the CFO and put them in the board deck before the trough hits.

If the CFO is not bought in by the end of these three conversations, the transition will not work. Find out before you publish a sunset date.

What I have learned by getting this wrong

I have run versions of this transition multiple times. As SVP at Salesforce, where the cannibalization tension between Marketing Cloud, Quip, and Slack was a daily experience. Once at SOCi during a positioning shift. Now at Smartcat, in real time, as the company transitions from translation SaaS to expert-enabled agents.

Three things I got wrong in earlier transitions that I will not get wrong again.

One: I waited too long to publish the sunset date internally. I thought I was protecting the legacy team by keeping the sunset informal. What I was actually doing was preserving their ambiguity, which kept them holding on, which kept the transition slow. Publishing the date is a kindness, not a threat.

Two: I let the CRO modify the comp plan halfway. The first version was strong. The CRO pushed back. We compromised. The compromise meant salespeople kept selling the legacy because it still paid better in some scenarios. Three quarters of slow successor growth could be traced to that one comp decision.

Three: I tried to make the post-sunset reorg humane by being vague about it until late. Vagueness is not kindness. People want to know what their lives look like after the transition. Plan the reorg early, communicate it transparently, give people time. The teams that knew their next role moved with energy. The teams that didn't know became ghosts.

The cleanest version of this transition I have seen, anonymized, ran 21 months from sunset announcement to sunset day. Started at $30M ARR, dipped to $24M at the bottom of the trough, recovered to $42M at sunset, hit $58M six months later. Gross margin compressed from 81% to 61% at the trough and recovered to 71% by month 30. The CPO who ran it kept their seat. The board sent a thank-you note.

The dirtiest version I have seen ran 36 months, ended at the same revenue as it started, lost three executives, and the company was acquired in distress. Same industry. Same customer base. Same starting position. Different decision sequence.

What to try this week

If you suspect your product line is sitting in the soft-pivot fantasy, run this exercise on a single page before Friday.

1. Write down the legacy product line. Annual revenue, gross margin, NRR, dominant pricing model.
2. Write down the AI-native successor (or the one you would build if you had to start today). Pricing model, expected unit economics at scale, the buyer it sells to.
3. Compare the buyer columns. If they're the same buyer, you have a sunset on your hands. If they're different buyers, you have a split on your hands. If the AI-native version is a feature additive to the legacy, you have a refresh.
4. If the answer is sunset, draft the seven-decision sequence on the same page. Don't get fancy. One bullet per decision, one sentence per bullet.
5. Show the page to your CFO. Just the CFO. Ask them what would have to be true for the gross margin trough to be acceptable to the board.

Their answer is your next month of work.

You don't have to publish the playbook this quarter. You do have to know whether you're running it. The companies that win this inflection are running it. The companies that lose are still talking about soft pivots.

---

The Cannibalization Decision Tree, the CPO Coalition Map, and the Board Narrative Slide Outline I use with operating teams are at falkster.com/toolkit. The pillar handbook chapter that goes deeper on the decision framework lives at /handbook/cannibalization-decision-framework. The companion essays on dual transformation and pricing migration are linked from /cpo.

Further reading

• Adobe's Creative Cloud transition postmortem (Stratechery archive)
• Microsoft's perpetual-to-subscription playbook (Ben Thompson)
• Sierra reaching $100M ARR on outcome pricing (The Information)
• Sarah Tavel on "selling the work, not the tools" (20VC)
• HFS Research: The CPO Mandate

Related

• The Dual Transformation Operating Model, the operating rhythm for running both clocks.
• The Pricing Migration Sequence, the 18-month quarterly playbook.
• The Cannibalization Decision Framework, the pillar handbook chapter.
• /cpo, the executive-strategy index.