HubSpot is shifting its Breeze AI pricing from per-use to per-outcome. Starting April 14, the Customer Agent costs $0.50 per resolved conversation (down from $1 per conversation). The Prospecting Agent moves to $1 per lead recommended for outreach (previously a flat monthly charge per enrolled contact).
CCO Jon Dick framed it simply: "You pay when it works, full stop."
The numbers so far: Breeze Customer Agent resolves 65% of conversations across 8,000+ customers and cuts resolution time 39%. Prospecting Agent activations are up 57% quarter-over-quarter, with a 10% lift in close rates across 10,000+ customers.
The Real Story: HubSpot Is Late
Sierra launched outcome-based pricing in early 2024. Co-founders Bret Taylor and Clay Bavor built the model from day one: pre-negotiated rate per autonomous resolution, free if escalated to humans. Sierra hit $100M ARR in 21 months, then $150M+ ARR by February 2026. $10 billion valuation. Customers include ADT, SiriusXM, Rivian, and Cigna, seeing 50 to 90% of service interactions fully automated.
Taylor's take: outcome-based pricing gives startups a structural advantage. Legacy CX providers depend on seat-based revenue. Better AI means fewer seats, which undermines their own business. Sierra has no conflicting incentives.
Intercom's Fin agent ran the same play. Launched at $0.99 per resolution, it scaled from $1M to $100M+ ARR. Fin now resolves 2 million customer issues per week across 8,000 companies. Resolution rates climbed from 27% at launch to 66 to 67% today. Intercom backed it with a $1M performance guarantee: hit 65% resolution or they pay you.
Salesforce took a different route with Agentforce: three pricing models at once. Per-conversation ($2), per-login session ($1.50 to $2), and flat monthly ($500 to $1,000 per agent). The market is still figuring out which model sticks.
What This Means for Sales Teams
If you are using HubSpot for prospecting or support, you now pay only when AI delivers a qualified lead or closes a ticket. No more paying for dead-end conversations or junk leads. This shifts the risk from buyer to vendor.
For CRMs generally, this is the direction. Sierra, Intercom, and now HubSpot are proving the model works at scale. If your sales stack is still charging per seat or per message sent, ask why. Comp structures for sales teams using AI tools will also shift: if the tool only bills on outcomes, quotas should reflect that efficiency gain.
HubSpot has 215,000 customers globally, 10,000+ in ANZ, and $2.8 billion ARR as of Q4 2024. The company employs around 150 to 200 staff in Sydney and Melbourne. No recent ANZ-specific hiring or expansion tied to this pricing change.
Worth noting: outcome-based pricing works when resolution rates are high and consistent. HubSpot's 65% is solid. Sierra's customers hit 50 to 90%. Intercom averages 66 to 67%. If those numbers drop, the model breaks. The question is not whether outcome pricing is better. It is whether the AI actually works reliably enough to bet revenue on it.
For now, the data says yes. Whether that holds in 12 months is the real test.