Operator note

Why product, go-to-market and pricing must be rebuilt together

May 2026.6 min read

Audience: CEOs and chief revenue officers running a SaaS-to-AI-native transition.

Most software companies treating AI as a transition programme have built three workstreams. Product reworks the roadmap to add agents and copilots. Sales redesigns the motion to handle longer enterprise cycles. Finance considers a pricing reset, usually next year. The three move at their own speeds, with their own owners, on their own slides. The CEO calls this disciplined sequencing. It is the operating mistake that is making most AI transitions fail.

The first instinct is to call this a coordination problem. It is not. Coordination assumes the three workstreams are independent and need to be aligned. They are not independent. AI changes value capture before it changes anything else, and value capture is a pricing variable. The pricing model dictates the unit economics. The unit economics dictate the sales motion that can be funded. The sales motion dictates which buyer is in the room. The buyer in the room dictates what product the company is actually selling. Run product, go-to-market, and pricing as sequential workstreams and the company will reach the pricing decision after it has already overinvested in the wrong product for the wrong buyer.

The market evidence is now visible. MIT’s Project NANDA found that 95% of enterprise generative AI pilots are failing to deliver financial return despite roughly 30 to 40 billion dollars in committed enterprise spend (MIT NANDA, 2025). McKinsey’s State of AI 2025 puts the same number from the buyer side: 88% of organisations use AI in at least one function, but only 6% qualify as high performers where AI contributes more than 5% of EBIT (McKinsey, 2025). The gap is not model quality. It is operating-model rewiring. Workflow redesign, McKinsey reports, has the biggest effect on whether a company sees EBIT impact from generative AI at all (McKinsey, 2025).

The mechanism is unfamiliar to anyone trained on the first growth curve. AI introduces inference cost as a variable. For every million dollars in AI product revenue booked in 2026, roughly 230,000 dollars walks out the door as inference cost before payroll (SaaS CFO, 2026). Bolting an AI assistant onto an 80-dollar seat adds roughly 15 dollars of direct variable cost and drops gross margin from 80% to closer to 65% (SaaS Magazine, 2026). The seat that powered SaaS unit economics is now structurally unprofitable as the carrier of an AI feature. The fix is not a cost programme. It is a different pricing model, which forces a different sales motion, which forces a different product packaging. Salesforce’s Agentforce ships at roughly two dollars per conversation; Intercom’s Fin at 99 cents per resolved interaction (Bessemer AI Pricing Playbook). Both are outcome-priced, both pass inference cost through, both require a sales conversation framed around outcomes the buyer can verify, not seats the buyer can authorise.

The seat-based default is now visibly retreating. Seat-based pricing fell from 21% to 15% of companies in 12 months; hybrid models surged from 27% to 41% (Monetizely, 2026). Gartner forecasts at least 40% of enterprise SaaS spend will shift to usage-, agent-, or outcome-based models by 2030 (Gartner via Monetizely, 2026). 92% of AI software companies now use mixed pricing models (Bessemer, 2025). The buyer is changing too. EY notes that the agentic AI buyer is shifting decision rights from IT to business-function leaders who can underwrite outcomes (EY, 2026). Mayfield’s 2026 CXO survey reports most enterprise technology leaders are reallocating budget from legacy vendors to AI-native providers (Highspot citing Mayfield, 2026). The procurement door has moved. The sales motion that worked at the IT door does not work at the function-leader door.

The discipline that works is to run the three as one programme with one accountable owner, one operating cadence, and one decision log. Call it a value-capture programme rather than an AI programme. Its starting question is not which features ship next; it is which outcome the customer will pay for, what unit that outcome is priced in, what gross margin the unit can carry after inference cost, what motion can sell it inside the buyer’s procurement reality, and what product packaging makes that motion repeatable. Each of those answers constrains the others. Sequencing them takes the constraint out, which is why sequencing has stopped working.

A programme structured this way looks different on day one. The chief revenue officer and the chief product officer share a single P&L for the AI portfolio, with the CFO calling the unit-economic gate. Pricing decisions cannot ship without sales-motion sign-off, because the motion has to be funded by the margin. Product packaging cannot ship without pricing sign-off, because the package has to be sellable at the price. The three meetings collapse into one weekly forum. The CEO chairs it. This is unpopular. It removes the autonomy each function had under the SaaS playbook. It is also the only known sequence that has produced AI products with durable economics. The Bessemer cohort that hit 60% gross margins shares this property: pricing, packaging, and motion were designed together, not in sequence (Bessemer State of AI, 2025).

The strongest objection is that an integrated programme moves slower than three parallel ones. In the planning phase it does. In the value-capture phase it is two to three quarters faster, because the company stops shipping AI features that pricing cannot carry and pricing changes that sales cannot sell. The second objection is that this is consulting reorganisation. It is not. It is a temporary operating overlay that runs until the AI portfolio either earns its own operating leverage or is killed. Most companies will need it for 18 to 24 months. The ones that already have it usually built it after a failed sequential attempt.

The leadership test is short. Ask the chief product officer what pricing assumption the next AI release is sold on. Ask the chief revenue officer what gross margin the motion is designed for. Ask the chief financial officer what motion the pricing is funding. If the three answers do not interlock, the company is running three workstreams, not one programme. It is making the operating mistake that the first cohort of failed AI transitions made. The rule of thumb is the same one the SaaS transition produced 15 years ago, restated for this curve.

Pricing, motion, and product packaging move together or they do not move.

Sources

go-to-market, pricing, product, operating-model, CEO

Back to writing