Salesforce just started charging per resolution instead of per seat — the 2026 pricing rewrite every CRM buyer needs to understand

For two decades, buying business software came down to one number: how many people need a login, times a monthly price per login. That's per-seat pricing — the default for CRM, ERP, HR tools, almost everything a business has ever paid a subscription for. It made sense when the thing generating value was a person sitting at a desk using the software. It stops making sense the moment the thing doing the work is an AI agent that doesn't log in, doesn't take a lunch break, and doesn't correspond to a seat at all.

That's not a hypothetical anymore. Salesforce — the biggest CRM company in the world — restructured its Agentforce pricing this year to bill by the resolved case instead of the seat. Zendesk and Intercom made the same move for their AI support agents. Sierra, the customer-service AI company Bret Taylor co-founded, only bills when its agent actually finishes the job — resolves a ticket, saves a customer who was about to cancel, or closes an upsell. None of this is one quiet experiment: Bessemer Venture Partners, tracking pricing across more than 200 AI vendors, found pure per-seat pricing fell from 21% to 15% of the market in twelve months.

If you're comparing a CRM, ERP, or HR system for your business right now — a name-brand platform or something built for you — the quote in front of you might not be measured in the same unit as the one next to it. One vendor's '$79 per user' and another's '$2 per resolved case' aren't different prices for the same thing; they're different bets on how your business actually uses the software. Getting that wrong means either paying for seats nobody logs into, or getting an unpredictable bill the month usage spikes.

The logic break is straightforward once you see it. Per-seat pricing assumes the thing paying for value is a human who logs in — more people using the tool, more value, more seats, more revenue for the vendor. An AI agent inverts that: the better it works, the fewer human seats a business actually needs, so a vendor charging per seat gets paid less the more useful its own AI becomes. Sierra co-founder Clay Bavor put it plainly on CNBC in July 2026 — AI agents break the per-seat assumption because an agent doesn't log in the way a person does, so Sierra bills only when its agent fully resolves a conversation, retains a customer who intended to cancel, or completes an upsell ^[5].

This isn't one startup's pitch. Bessemer Venture Partners tracked pricing structures across more than 200 AI vendors and found pure per-seat pricing fell from 21% to 15% of the market in a single year, while hybrid pricing — a base seat fee plus usage charges on top — rose from 27% to 41% over the same period. The firm's finding that should worry any vendor still selling flat per-seat: companies that price to outcomes are growing 1.5 to 3 times faster than the ones still clinging to the old model ^[1].

Buyers are pulling in the same direction, not just vendors. Futurum Group's 1H 2026 survey of enterprise software buyers found 43% now prefer consumption-based pricing and 27% prefer outcome-based pricing — fewer than one in five still want the classic per-seat model, and the survey noted seat-only vendors are already getting disqualified from deals before they reach a demo ^[2].

The clearest name-brand proof of the shift is Salesforce itself. Agentforce launched at a flat $2-per-conversation rate and was restructured this year into a pay-per-resolution model — $2 charged only when the agent autonomously resolves a case, nothing for a failed attempt or one escalated to a human, alongside a separate $500-per-100,000-action Flex Credit option that works out to roughly ten cents an action. Zendesk made the equivalent move, charging for successful AI-driven resolutions instead of bundling the feature into its per-agent license, and Intercom applies a comparable resolution-based price to its Fin AI agent ^[3][4].

None of this makes outcome-based pricing automatically cheaper — that's the part worth being skeptical about. A low resolution rate can make 'pay only for what works' cost more per real outcome than a flat seat ever would have, because every failed attempt still ties up the agent's time even when it doesn't show up on the invoice. That's also why hybrid pricing is winning the adoption race over pure usage or outcome pricing: it gives the vendor predictable revenue and the buyer a cost floor they can actually budget against ^[1][3].

What actually matters for a small or mid-size business shopping for a system in 2026 isn't picking a side in this debate — it's reading the contract closely enough to know which meter you're on. A per-seat quote and a per-resolution quote for what looks like the same CRM aren't comparable numbers until you've run your own usage pattern through both: how many people actually touch the tool daily, how many AI-handled interactions you'd realistically generate a month, and what happens to the bill in your busiest month, not your average one.