Half of ServiceNow's new business is no longer seat-based. The seat is dying.

For thirty years, business software has been priced on a single, convenient fiction: the value of software is proportional to the number of people who log into it. The seat was never a great proxy for value, but it was a stable one, easy to count, easy to forecast, easy to true-up at renewal.

Last week, on ServiceNow's Q1 earnings call, that era got its clearest obituary yet. CEO Bill McDermott reported that active seats grew 25%, a number most SaaS companies would celebrate, and then noted that roughly half of the company's net-new business now comes from non-seat-based pricing: tokens, consumption, connectors, infrastructure. One of the largest enterprise software companies in the world is telling investors, on the record, that its growth no longer runs through the seat.

Why the seat is breaking now

Three forces are converging on the same pressure point.

AI decouples value from user count

When an AI agent resolves a ticket, drafts a contract summary, or runs an outbound campaign, no human logged in to do that work. The value happened; the seat did not. As more of a product's value is delivered by the software acting rather than a person clicking, per-seat pricing systematically undercharges for what the product does and overcharges for who watches it do it.

Productivity gains shrink seat demand

The quiet second-order effect: if your product makes each customer employee twice as productive, your customer needs fewer employees in that function, and eventually fewer seats. Your best product work becomes your worst renewal conversation. Benchmarks tell the story of where this ends: median net revenue retention across B2B SaaS has compressed to roughly 101%, with SMB-focused companies dipping below 100%. The expansion engine that seat pricing relied on, more hires means more seats means automatic growth, has stalled across the industry.

Buyers have new comparisons

Your buyer now sees agent products priced per conversation, per resolution, per task completed. Once a CFO has bought anything priced on work delivered, a per-seat line item for software that "assists" invites an unwelcome question: what did we actually get for these seats? The utilization audits we keep writing about are the buy-side version of this same shift.

What to do if you sell software

Map your value events. Somewhere in your product, discrete things happen that customers would recognize as valuable: a document processed, a workflow completed, a lead enriched. List them, instrument them, and rank them by how directly a customer would connect them to money. That list is your future price metric menu, whether you act on it this year or not.

Move to hybrid, not to extremes. The pattern emerging among companies that have navigated this well is a platform fee that secures predictable baseline revenue plus a consumption or outcome component that captures upside. Pure usage pricing terrifies CFOs on both sides of the table; it makes your revenue volatile and your customer's costs unbudgetable. A committed-plus-overage structure gives both sides a forecast.

Rebuild your comp plan alongside your price metric. This is the step everyone skips. A sales team paid on booked ARR will resist consumption pricing with every fiber of its being, because consumption revenue lands after the deal instead of inside it. If your pricing changes and your comp plan does not, your new model will die in the field, not in the spreadsheet.

Watch your NRR composition, not just its level. Retention driven by seat true-ups is decaying structurally. Retention driven by consumption growth is the durable kind. Two companies with identical 105% NRR can have opposite futures depending on which engine produced it.

What to do if you buy software

The same shift is leverage on the buy side. Vendors mid-transition are unusually flexible, and there are three moves worth making this year:

• Stop paying for shelf seats. Pull utilization data before every renewal. The era when nobody checked is over, and vendors know it.
• Ask every vendor for their consumption option. Even when it is not on the price list, it increasingly exists. For workloads with variable volume, metered pricing usually beats the seat math.
• Negotiate caps and floors. If you accept usage pricing, cap your downside. If you keep seats, negotiate the right to convert mid-term. Optionality is cheap right now because vendors want the reference migrations.

Choose your transition while it is still a choice

Pricing model transitions look gradual from the outside and sudden from the inside. The seat will not disappear; access still has value, and plenty of software will stay per-user for years. But growth has already moved. When a company of ServiceNow's scale sources half its new business outside the seat, the default assumption of the industry has flipped, and every pricing conversation your team has for the next five years happens in the shadow of that flip.

The companies that redesign their revenue model while seat revenue is still healthy get to choose their transition. The ones that wait for NRR to force the issue will make the same changes with less leverage, less runway, and less credibility. Choose early.