The Death of Per-Seat Pricing: Why B2B Software Is Being Forced to Sell Outcomes, Not Logins, in 2026

For thirty years, the entire B2B software industry rested on a single, almost invisible assumption: that you pay for software by counting the people who use it. Seats. Logins. Named users. It was clean, predictable, and easy for a CFO to model. Add a hundred employees, add a hundred licenses. Renew, expand, repeat. The seat was the unit of value, and the seat was the unit of revenue, and for a long time nobody thought to question whether those two things should ever come apart.

Then AI agents arrived and quietly broke the assumption in half.

An AI agent does not log in. It does not consume a named-user license. It does not map to a row in your HR system or a line in your headcount plan. It does work — sometimes the work of one person, sometimes the work of fifty — and it does that work at three in the morning without a seat assigned to it. When the thing producing value no longer corresponds to a human user, the price tag stapled to "users" stops making sense. And once buyers notice that the value and the bill have decoupled, the most durable pricing model in enterprise software history starts to look like an accident of history rather than a law of nature.

For Revenue Leaders, Product and Pricing Executives, CFOs, and GTM Teams trying to understand why the per-seat model is unraveling — and how to monetize software when the "user" is increasingly a machine.

This article is about the structural collapse of seat-based pricing, the messy hybrid era we're living through right now, and what it takes to price software around results in a market where the buyer has finally started asking what, exactly, they're paying for.

The model that is quietly dying

Start with the raw direction of travel, because the shift is no longer a forecast — it's already showing up in the numbers.

The share of SaaS companies relying on pure per-seat pricing fell from 21% to 15% in a single twelve-month stretch, a six-point drop that would have been unthinkable in the subscription industry's first two decades. Gartner now projects that by 2030 at least 40% of enterprise SaaS spend will shift toward usage-, agent-, or outcome-based pricing, with seat-based revenue share continuing its decline from 21% toward 15% over the same period. Looking nearer term, Gartner has also estimated that 70% of businesses will prefer usage-based pricing over per-seat models as the agentic era takes hold.

This is not a fringe experiment confined to AI-native startups. Roughly 85% of SaaS companies now use some form of usage-based billing, even if only as a component of a larger package, and 47% of SaaS companies are actively exploring or piloting outcome-based models. The pure-seat vendor, once the default, is becoming the exception.

The headline for revenue leaders is uncomfortable but clarifying: the unit you have built your entire forecast, comp plan, and expansion motion around is depreciating in real time. The questions that follow — what replaces the seat, how fast, and what it does to your net revenue retention — are now operational questions, not strategic musings for an offsite.

Why agents, specifically, broke the math

It's worth being precise about why AI is the trigger, because plenty of people have predicted the death of per-seat pricing before and been wrong. Usage-based pricing has existed for years. What's different now is that the economic logic of the seat has been severed, not merely challenged.

Per-seat pricing works when value scales with the number of humans doing the work. More salespeople using a CRM, more developers using a code tool, more agents in a support center — each additional human extracts roughly proportional value, so charging per human is a rough but serviceable proxy for charging for value. The seat was never really about the login. It was a convenient stand-in for value delivered.

AI agents sever that link. When an autonomous agent resolves a thousand support tickets, the value created has nothing to do with how many human agents are on the team — in fact, the better the AI performs, the fewer humans you need, which means the better the software works, the less a per-seat vendor gets paid. That is a pricing model running in reverse: success for the customer becomes shrinking revenue for the vendor. No CFO at a software company can defend a model where doing the job well cannibalizes the bill.

The buyer side sees the same arithmetic from the opposite direction. Deloitte predicts that up to half of organizations will direct more than 50% of their digital transformation budgets toward AI automation in 2026, with agentic AI investment potentially reaching three-quarters of companies. Buyers spending that aggressively on automation are not going to keep paying for seats that automation is actively eliminating. The moment a procurement leader can say "we deployed agents and cut our license count, but our bill stayed flat," the per-seat contract becomes the thing they renegotiate first.

The hybrid era: where everyone actually is right now

If you read the breathless takes, you'd think the industry had already leapt from seats to pure outcomes. It hasn't. The reality of 2026 is messier and more interesting: hybrid pricing — a fixed base plus variable consumption — has become the de facto industry standard, with roughly 41% adoption, making it the single most common model and the dominant transition state for enterprise renewals.

The reason hybrid won the interim is simple and worth internalizing: it solves the problem both sides actually have. Vendors need revenue predictability to run a business and satisfy investors; buyers need cost predictability to budget. Pure usage-based pricing terrifies the CFO who just wants a number for the annual plan, and pure outcome-based pricing terrifies the vendor whose revenue would swing wildly with customer behavior they don't fully control. A base subscription that guarantees a revenue floor, paired with consumption or outcome overages that capture upside, threads that needle. As one survey of the 2026 landscape put it, buyers want "pricing predictability with flexibility, not a blank check" — and hybrid is the structure that delivers it.

Buyer preference data backs this up rather than contradicting it. In Futurum's first-half 2026 survey, 43% of buyers said they prefer consumption-based models and 27% favor outcome-based structures — meaning a combined majority want their bill tied to value, but a large bloc still wants the predictability that consumption (rather than pure outcomes) provides. The market is voting for value alignment with guardrails, not for the abolition of all fixed costs.

The practical takeaway: if you're redesigning your pricing in 2026, the destination is almost certainly hybrid, not a clean swap from one pure model to another. The teams getting this right are building a base-plus-consumption architecture and treating pure outcome-based pricing as a premium option for specific use cases, not a wholesale replacement.

What "outcome-based" actually looks like in the wild

Abstractions are easy to nod along to; the real test is what vendors are charging for. The leading edge of the market has already moved from "per user per month" to "per result," and the examples are instructive precisely because they're not theoretical.

In customer support — the first category to feel the full force of agentic automation — vendors are now billing per resolution rather than per seat. Intercom's Fin AI agent charges $0.99 per resolved conversation, and competitive pressure is already compressing that figure, with HubSpot reportedly dropping its equivalent to $0.50 in 2026. Zendesk now prices AI agents at roughly $1.50 to $2.00 per automated resolution rather than selling AI seat licenses at all. The value metric is no longer "how many support reps have access" but "how many problems did the software actually solve" — a unit the buyer can directly tie to deflected headcount and faster response times.

The most revealing case study, though, is Salesforce, because it shows just how unsettled this all still is. Salesforce has shipped three distinct pricing models for its Agentforce product in roughly eighteen months. It launched with $2 per conversation (a conversation being any interaction within a 24-hour window). Then in May 2025 it introduced Flex Credits at $0.10 per action, with packages sold at $500 for 100,000 credits. Then in late 2025 it added a per-user Agentic Enterprise License Agreement starting around $125 per user per month, reframing the seat as a wrapper that now includes a "digital workforce." Crucially, Salesforce didn't kill the earlier models — all three now run simultaneously.

The Agentforce saga is the whole industry in miniature: even the largest, most sophisticated vendor on earth cannot yet decide whether to charge by the conversation, the action, or the seat — so it charges by all three and lets the buyer pick. If you've been agonizing over choosing the one perfect pricing model, take some comfort: the market leader hasn't either. The winning near-term move is optionality, not certainty.

The CFO's revenge — and the value-metric problem

There's a deeper reason the per-user model keeps clawing its way back, visible in Salesforce's return to a seat-based wrapper: finance leaders still need a number they can budget. A pure pay-per-action model where the bill is a function of a thousand micro-events is operationally honest but financially terrifying for a buyer trying to forecast next year's software spend. This is the "tollgating" anxiety Deloitte has flagged — the fear that every interaction with the software becomes a metered toll, turning a predictable subscription into an unpredictable utility bill.

This is why the hard part of outcome-based pricing isn't philosophical — it's the value metric. To charge for an outcome, both parties have to agree on what the outcome is, how it's measured, and who's accountable when the agent gets it wrong. "Per resolved conversation" works because a resolution is observable and roughly attributable. "Per dollar of revenue influenced" or "per qualified lead generated" gets murky fast: attribution is contested, the vendor doesn't fully control the result, and disputes over measurement can poison the relationship. The vendors winning at outcome-based pricing have done the unglamorous work of defining a value metric that is observable, controllable, and hard to argue about — and they've resisted the temptation to price on outcomes they can't actually influence.

For your own pricing strategy, that translates into a sharp filter. Price on outcomes only where you can measure the result cleanly and credibly take responsibility for it. Everywhere else, charge for consumption with a predictable base. Outcome-based pricing is a scalpel, not a sledgehammer, and treating it as a universal replacement is how vendors end up in unwinnable measurement fights with their best customers.

What this means for net revenue retention and the expansion motion

The strategic stakes here go well beyond the price list. The per-seat model didn't just set prices — it shaped the entire go-to-market engine. Land-and-expand worked because expansion meant selling more seats into an account as it grew its headcount. Customer success teams were measured on seat adoption. Sales comp was built on seat-count growth. When you pull out the seat, you pull a structural support out from under all of it.

In an agent-priced world, expansion is no longer about adding users — it's about increasing consumption or driving more measurable outcomes. That's actually a healthier dynamic when it works: revenue grows when the customer gets more value, automatically, without a renewal negotiation to add seats. But it inverts the customer success mandate. The CS team's job shifts from "drive seat adoption" to "drive usage that the customer experiences as valuable" — and if usage stalls because the product isn't delivering, revenue contracts just as automatically. Net revenue retention becomes a direct, real-time readout of delivered value, with nowhere to hide.

This is the double edge that revenue leaders need to sit with. Usage- and outcome-based models can produce spectacular NRR when the product genuinely creates compounding value — consumption rises on its own. But they remove the cushion that seat-based contracts provided, where a customer kept paying for licenses they barely used. In the new model, underused software doesn't quietly generate revenue — it generates churn. The forgiving slack in the seat-based system is gone.

The road ahead

The honest summary is that the industry is in the middle of the messiest pricing transition in its history, and anyone selling you a tidy answer is selling you something. The seat is dying as the universal unit of value, but it isn't disappearing — it's being demoted from "the model" to "one option among several," often surviving as a predictable wrapper around consumption and outcomes underneath. Hybrid is winning the interim because it's the only structure that gives vendors a revenue floor and buyers a budgetable ceiling at the same time. And pure outcome-based pricing, for all the headlines, remains a precision instrument for the specific cases where value is cleanly measurable, not a wholesale replacement for everything that came before.

What's no longer in dispute is the direction. With per-seat's share falling from 21% to 15% in a year, Gartner projecting 40% of enterprise SaaS spend shifting to usage, agent, or outcome models by 2030, 85% of SaaS companies already using some form of usage-based billing, and the market leader running three pricing models at once because it can't yet pick one, the era of counting logins as a proxy for value is ending. The agent broke the proxy, and there's no putting it back together.

The mandate for revenue leaders is to start the transition on your own terms before a buyer forces it on you at renewal. Audit where your price is still tied to seats that automation is quietly eliminating. Identify the one or two outcomes you can measure cleanly and stand behind. Build a hybrid architecture that protects a revenue floor while letting your bill rise with the value you deliver. And rebuild your expansion motion around consumption and results rather than headcount, before the headcount you've been billing for disappears. The vendors who do this deliberately in 2026 will set the terms of their own repricing. The ones who wait will have the terms dictated to them — by a procurement leader holding a contract for seats that no longer log in.