The third Agentforce pricing model in two years: $2 per autonomous resolution. Why the meter moved into the definition of a resolution, the four questions that belong in the contract, and what to do with stranded Flex Credits.
The presenter in this briefing is an AI generated avatar. The research, figures, and guidance are real, produced by Redress Compliance analysts from our consulting engagements and market network.
Two dollars per autonomously resolved case is Salesforce's third Agentforce pricing model in under two years. Claire explains where the meter really moved: into the definition of a resolution. Four questions belong in the contract, plus audit rights, a rate hold with volume tiers, and a conversion path for stranded Flex Credits.
The full narration, section by section. Click a section heading to jump the player to that moment.
Two dollars per resolved case. That is Salesforce's new price for Agentforce, and it sounds like the simplest pricing you have ever been offered. So here is the question that decides whether it is a good deal: what exactly counts as a resolution, and who gets to count them?
Some facts first. On June twenty fifth, Salesforce announced pay per resolution pricing for its Help Agent: a flat two dollars when the agent resolves a case autonomously, end to end, and nothing when it escalates to a human. It is the third Agentforce pricing model in under two years, after per conversation pricing and Flex Credits. The machinery behind it comes from two acquisitions, m three ter for metering and Intercom's Fin for outcome pricing.
And the context matters. Agentforce has crossed one point two billion dollars of annual recurring revenue against a forty five billion dollar base, and the stock is down roughly a third this year. A vendor under that pressure needs adoption logos and consumption growth. That is buyer leverage, and it is unusual leverage to hold against Salesforce.
Now the substance. Outcome pricing is genuinely better aligned than paying per conversation or burning credits. You pay when the machine actually does the work. But the meter has not disappeared.
It has moved into the definition of a resolution, and that definition is where this deal is won or lost. Four questions belong in the contract, in writing. What counts as autonomous when the agent used a human authored macro along the way? What happens with partial resolutions, where the agent solved half and a person finished the case?
What about cases that reopen two days later? Is that still a paid resolution? And what about deflections, where the agent pointed at a help article and the customer simply gave up?
now. Then secure the mechanics. Audit rights over the count, so your team can reconcile what was billed against what actually happened. A rate hold on the two dollar unit for the full term, with volume tiers, because a price that has moved three times in two years will move again.
And if you committed to Flex Credits under the previous model, those credits are stranded by this one. Convert them or true them down. Do not pay twice for the same work.
One more thing. Keep every Salesforce AI commitment short, with conversion rights to whatever model comes next, because on current form there will be a next one. Model instability is not a reason to avoid the product. It is a reason to refuse long rigid paper, and a vendor experimenting with its own pricing cannot credibly demand it from you.
The bottom line. Pay per resolution is a negotiation over definitions, not over a price. Nail the definitions, cap the rate, rescue your stranded credits, and sign short. And if the proposal is already on your desk, we will benchmark it before you sign.
Twenty five percent of what we save, and nothing otherwise.