Next-gen pricing models can go from non-existent to must-have seemingly overnight. Just look at AI customer support for a preview of what might happen to you 👀
For the longest time pricing was rather, well, sleepy. There were a lot of seat-based subscriptions & Good-Better-Best packages. Plans ranged from about $20 to up to $150+ per seat per month.
Intercom shook things up when it introduced Fin, its AI agent that autonomously resolves customer conversations. Fin had a form of outcome-based pricing, charging just $0.99 per successful resolution.
This pricing was a win-win. It shifted risk from the customer to the vendor since the customer only pays when AI actually works. And it allowed Intercom to automatically make more $$ as the AI improved — turning R&D into a revenue-generating function.
It took a while for the next player to catch up. Zendesk announced an outcome-based pricing model about 18 months later.
Salesforce and Kustomer followed suit with per-conversation pricing — not fully outcome-based, but more aligned with the work delivered by AI. Help Scout changed pricing, too, announcing their pricing now includes unlimited seats (they charge based on « contacts helped »).
It has quickly become industry-standard to advertise an outcome-based model (see Sierra), even if customers might ultimately buy a work-based or consumption-based one (see Decagon).
The first company to shift takes on the biggest risk — and they have the most to gain if their bet is correct.
This all leaves me wondering: which category comes next?
How I’d make the decision:
1. Is there a clearly defined success metric in the market?
2. Do customers value that outcome?
3. Does your AI deliver predictable performance?
4. Do you expect that performance to continue to improve?
5. Do you have a clear path to attribute results to your agent (either directly, via A/B testing or with a POC)?
If so, you could be in the driver’s seat.