On June 15, 2026, Salesforce agreed to pay $3.6 billion for Fin, an AI customer-service platform that doesn't charge per seat — it charges $0.99 for every conversation its AI actually resolves. **AI vendor pricing in 2026 is splitting into two real models, flat-fee licensing and pay-only-when-it-works outcome pricing, and which one actually costs less depends on your call volume and resolution rate, not on which one sounds more modern.**
What Is Outcome-Based AI Pricing?
Outcome-based pricing means a vendor bills only when its AI actually finishes the job — not for every seat, license, or attempted conversation. Fin defines a "resolution" as a conversation where the customer's issue was confirmed solved, not just answered, and it charges $0.99 per resolution with conversation-level transparency so a customer can audit what it's being billed for (Fin, 2026). Zendesk applies the same logic to its own AI agents: its pricing page states customers "pay only for customer requests that were successfully resolved by the AI agent, without any escalation to a human agent," with plan-level allowances and custom rates above that (Zendesk, 2026). Neither model charges for a seat sitting idle.
Why Did Salesforce Pay $3.6 Billion for a Pricing Model?
Fin — founded in Dublin in 2011 as Intercom, rebranded in 2025 — resolves an average of 76% of support volume end-to-end across roughly 12,000 customers, a rate the company says improves about 1% a month, and it counted more than 30,000 companies as customers at the time of the deal (Salesforce, 2026; Fin, 2026). Salesforce CEO Marc Benioff said Fin "brings proven agent technology, a deep commitment to customer success, and an incredible AI team that will complement Agentforce with powerful service agent capabilities" (Salesforce, 2026). Salesforce's own Agentforce platform had already reached $1.2 billion in annual recurring revenue the same quarter the deal was announced. Salesforce didn't just buy a customer base — it bought a pricing model with a track record.
Per-Seat vs. Outcome-Based AI Pricing: Comparison Table
| Pricing model | How it's billed | Who bears the volume risk | Best fit |
|---|---|---|---|
| Per-seat / flat license | Fixed fee per month regardless of call volume | You — the bill is the same whether the AI handles 50 calls or 500 | Predictable, steady call volume; simple budgeting |
| Outcome-based / per-resolution | Billed only for calls the AI actually resolves (e.g., Fin: $0.99/resolution) | The vendor absorbs cost on failed or escalated attempts, but your bill rises with volume | High, growing call volume; want cost tied to results |
| Skyline (flat build + flat monthly fee) | One-time build, then a predictable monthly fee scoped to your call volume — see /contact | Shared — the fee is scoped to your business up front, not metered per call | Businesses that want accountability without a bill that moves every month |
Is Outcome-Based Pricing Actually Cheaper?
Not automatically. Fin's own published comparison shows per-outcome pricing beats per-conversation billing when the resolution rate sits below roughly 80% — but the gap narrows, and can reverse, once resolution rates climb into the 85-90% range (Fin, 2026). Run the arithmetic yourself before assuming outcome pricing wins: at $0.99 per resolution, 500 resolved calls a month is $495 before any base platform fee; 2,000 resolved calls is $1,980. A flat-fee vendor charging the same or less per month, with no ceiling on volume, can beat that math the moment your call volume — or your AI's resolution rate — climbs high enough. Outcome pricing rewards a vendor that resolves calls well. It doesn't automatically reward you.
Who's Actually Paying This Way in 2026?
The market has moved further than most buyers realize. Futurum Group's 1H 2026 Enterprise Software Decision Makers survey found 43% of buyers now prefer consumption-based pricing, 27% prefer outcome-based pricing, and fewer than one in five still prefer the classic per-user license (Futurum Group, 2026). That's a reversal of the default assumption most business owners are still shopping with — that AI, like most software, comes priced per seat.
When Is Per-Seat or Flat-Fee Pricing Still the Better Choice?
Say it plainly: if your call volume is low, steady, or genuinely uncertain — a seasonal business testing an AI receptionist for the first time, or an office that gets 40 calls a week, not 4,000 — a flat monthly fee is easier to budget against than a bill that moves with usage you haven't measured yet. Outcome pricing rewards volume and a high resolution rate; if you don't have either established yet, you're paying for a bet you can't evaluate. A flat fee also protects you from a bad month — a marketing push, a review campaign, a seasonal spike — turning into a surprise invoice.
What This Looks Like for a Multi-Location Flathead Valley Business
A three-location dental, veterinary, or property-management group in the Flathead Valley evaluating an AI receptionist should run this math per location, not as one combined number. Three offices each averaging 300 resolved calls a month is 900 resolved calls total — at $0.99 each, that's roughly $891 before any platform fee, a number worth comparing directly against a flat-fee build scoped to that same volume before signing anything. The math changes fast with a handful of locations; asking a vendor to show it to you before you sign is a fair, and increasingly normal, question in 2026.