Operator note

A CEO's 90-day AI-native reset agenda

May 2026.6 min read

Audience: CEOs and senior operators of B2B software companies in the first quarters of the second growth curve.

Most AI programmes in B2B software fail at the same moment: somewhere between the third and the sixth month, when the feature backlog catches up with the operating model and nothing has moved on the P&L. The agenda below is the operator’s version. Ninety days, in three 30-day blocks, designed to move pricing, margin, and capital allocation before the next board cycle. It is concrete enough to act on this week. It is senior enough to stay on the CEO’s desk. It assumes the diagnosis is already correct: the company has entered the second growth curve, and the SaaS playbook is not enough.

The operating tension

CEOs who attempt this reset usually start in the wrong place. The default move is to commission an AI strategy document, run an enterprise-wide tooling audit, and convene a cross-functional working group. Three months later, the deck is detailed, the working group is busy, and nothing has changed on the operating dashboard. The reset is paused for a strategy.

The false diagnosis is that the company needs more analysis. It needs a sequence. The five resets, in the right order, before the analysis. Pricing exposure, gross margin, go-to-market motion, capital allocation, roadmap. Anything else is theatre.

Days 1 to 30. Diagnose pricing exposure and product exposure

The first 30 days are diagnostic. Two questions, one each from finance and product, both run by the CEO directly.

Pricing diagnostic. Map every product line against three pricing models: seat, usage, outcome. Calculate the share of new ARR and renewal ARR exposed to seat-based pricing. The benchmark to test against: per-seat pricing collapsed from 21 percent of SaaS to 15 percent in the twelve months to early 2026, and 83 percent of AI-native SaaS now bills on usage (Monetizely, 2026; Stormy AI, 2026). Any seat-priced revenue above 50 percent of the book is a renewal exposure. The diagnostic does not need to be perfect. It needs to be honest enough to focus the next 60 days.

Product exposure audit. Walk the top ten customer workflows. For each, mark whether the product owns the workflow end-to-end, supports it, or feeds it. Mark the workflows where AI agents in the market are now resolving the work without a vendor in the middle. This is the workflow obsolescence risk: where AI collapses or absorbs the workflow a product currently owns, leaving the product on the wrong side of the control point. Five customers, one hour each. The patterns surface fast.

Output by day 30. A one-page exposure brief. Seat-priced ARR exposure, top three workflows at obsolescence risk, top three workflows with the largest pricing-model upside.

Days 31 to 60. Stress-test the go-to-market motion and the capital allocation

The second 30 days are interventional. Two stress tests, one each on go-to-market and capital.

Go-to-market stress test. Run a deal-quality review on the last 50 closed-won and closed-lost deals. Mark the median sales cycle, the win rate, the average committee size, and the share of deals that closed at outcome-based pricing. Compare to current benchmarks: 84-day median cycle, 22 percent SaaS win rate, 11.2 stakeholders for deals above 50,000 dollars, two to four weeks of added procurement load (Prospeo, 2026; Martal, 2026; Corporate Visions, 2026). The gap between observed and benchmark is the go-to-market debt.

Sales debt accrues silently. It pays out at the worst time, usually a quarter before a board review.

Capital allocation review. Pull the last four quarters of operating expense by line. Mark the dollars going to AI feature shipping, AI infrastructure, and AI-adjacent enablement. Mark the marginal dollar for the next quarter. The honest test: if AI is being funded out of slack inside engineering and marketing, the company is running an AI programme. If finance owns an inference budget, an AI-adjusted gross margin is tracked monthly, and one workflow-collapse bet has a concentrated allocation, the company is running a reset.

Output by day 60. A two-page operating brief. Go-to-market debt quantified, AI capital allocation visible, one bet identified for concentrated investment.

Days 61 to 90. Spin up the integrated reset

The last 30 days are constructive. The reset moves from diagnosis to operating cadence.

One. Reprice one product line. Convert one product line from seat to hybrid or outcome pricing, with a 90-day pilot window and a renewal cohort identified. Do not announce the full pricing strategy. Move one line. Gartner forecasts 40 percent of enterprise SaaS will move to outcome-based billing by 2030, but the operators making the move now are doing it because the cost structure compelled it (Bessemer Venture Partners, 2026). One line is enough to test the renewal model.

Two. Set the AI-adjusted gross margin target. Recalibrate the gross margin target for the AI revenue line specifically. ICONIQ data puts the average AI product gross margin at 52 percent, with inference at 23 percent of revenue (Monetizely, 2026). Pick a target in the 55 to 65 range and report against it monthly. The board should see the new line by the next reporting cycle.

Three. Concentrate capital behind one workflow bet. Move 20 to 30 percent of the marginal AI budget onto one workflow-collapse opportunity identified in the day 1 to 30 audit. The other 70 to 80 stays distributed. The company is not abandoning the feature backlog. It is signalling, internally and externally, where the second growth curve is being underwritten.

Four. Replace the standing AI roadmap review with an AI economics review. Four numbers, monthly: AI-adjusted gross margin, inference cost as a percent of AI revenue, pricing exposure at next renewal, outcome-priced share of new ARR. McKinsey reports 5.8x average ROI on production AI within 14 months, but only 29 percent of executives see significant returns, and 79 percent of organisations report adoption challenges (McKinsey, 2025; Writer, 2026). The four numbers above are how to close that gap inside the operating model.

Five. Write the board paper. Two pages. The diagnosis, the four monthly numbers, the one concentrated bet, the renewal cohort, and the decision being asked of the board.

What this is not

This is not an AI transformation programme. There is no working group, no centre of excellence, no enterprise change framework. Those structures absorb attention and produce slides. The reset is concrete because it stays on the operating dashboard.

The leadership implication

Three changes survive the 90 days. The standing AI roadmap review is replaced with an AI economics review. One product line is repriced. One workflow bet is funded. Each of those is a small change. Together they shift the orientation of the company from feature-shipping to economics-resetting. That is the difference between an AI strategy and AI roadmap theatre.

The rule of thumb

If the next board paper does not contain four numbers (AI-adjusted gross margin, inference cost as a percent of AI revenue, pricing exposure at next renewal, outcome-priced share of new ARR) and one concentrated bet, the 90 days have not been used. Add another 30, and run the same five tests.

Sources

CEO, AI-native, pricing, operating-model, capital-allocation

Back to writing