Pricing is the first AI-native lever

The first growth curve in B2B software rewarded one behaviour above all others: sell more seats to the same customer, then sell them more modules. Pricing was almost an afterthought. The list price was a number on a slide; the real economics lived in seat counts, upsell motions, and renewal mechanics. The second growth curve will not reward that behaviour. AI changes what the buyer is paying for and what the vendor is paying out, and pricing is the only lever inside the company that translates those two changes into recurring revenue structure.

Every board that anchors on feature velocity while leaving the pricing model alone is compounding margin compression instead of relieving it.

What AI changes in the value chain

The economic mechanism is simple once stated. AI introduces a marginal cost into a business that used to have almost none. It also collapses the workflow steps a customer was previously willing to pay for in seats. Both effects land in the same place: the pricing model. Seat-based pricing under-monetises usage that creates real cost and over-monetises usage that no longer requires a logged-in human. Outcome-based pricing solves the value side but, without a usage floor, leaves the vendor underwriting compute on the customer’s behalf. Usage-based pricing tracks cost cleanly but breaks the predictability that boards and underwriters built the first growth curve on. The pricing model is not a downstream consequence of product strategy; it is the bridge between AI cost economics and recurring revenue quality. Get the bridge wrong and the rest of the operating model fails to compound.

The downstream consequences

A pricing reset is not a marketing exercise. It rewires the company. Go-to-market changes first, because the conversation is no longer about seat counts and ROI calculators but about workflows owned, units of outcome priced, and usage envelopes underwritten. Sales compensation has to follow, or reps will keep selling seats because that is what their plan still pays for. Customer success becomes a margin function, not just a retention function, because the difference between a profitable and an unprofitable account is now driven by usage patterns inside the contract. Product roadmaps inherit a sharper test: each feature is either margin-accretive at the new pricing model or it is not, and the answer is knowable rather than asserted. Capital allocation tightens. Investment in features that demo well but cannot be priced at positive contribution stops earning room on the roadmap.

The strategic choices now required

The board agenda for a CEO running this transition has five entries. First, decide the pricing architecture before the next product release cycle, not after, because every release ships into the old model until the new one is set. Second, instrument cost-to-serve at the workflow level, so the pricing model can be tested against real margin rather than illustrative slides. Third, redesign sales compensation in the same quarter as the pricing change, or the field will run two playbooks at once and execute neither. Fourth, set a clear policy on grandfathered seat-based contracts, because the renewal cohort is where the new model is either reinforced or quietly diluted. Fifth, replace the question the board has been asking, which is some version of how fast are we shipping AI features, with a different one: are we monetising the workflow we now own.

A roadmap is not a strategy, and feature velocity is not a moat. Pricing is the lever the second growth curve actually rewards. The companies that move first will compound margin while their competitors compound activity.