Frameworks, core principles and top case studies for SaaS pricing, learnt and refined over 28+ years of SaaS-monetization experience.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
.png)
Once the product does the work, price it for the work it does, measured as closely as you can. When the result is clear and the customer can confirm it, charge for that result. Fin charges $0.99 each time it resolves a question, where "resolved" means the customer says it is solved or simply stops asking. HubSpot charges $0.50 for each support chat its AI resolves, and $1 for each good sales lead it finds. These are the cleanest models.
The result is a yes or no, the buyer can check it, and the value is easy to see. HubSpot put the principle on a banner when it made the switch in April 2026: AI should be measured in outcomes, not output. It could afford to say that because its support agent was already resolving 65% of conversations across more than 8,000 customers. When the AI does not deliver, the customer does not pay, which quietly moves the risk from the buyer to the vendor.
Not every product gives you a clean result like that. When it doesn't, get one step closer to value than seats and stop there: charge for actions the agent takes, tasks it finishes, or jobs it runs. Either way the rule holds. The closer the price sits to the value, the longer it lasts.
The setup underneath is usually a mix, at least at first. A base fee covers access to the platform and a set amount of AI work. Anything past that, the customer pays for as they use it. Intercom kept its seats on the Essential, Advanced, and Expert plans as that base fee, and let the AI's results become where the money came from. The seat stopped being the engine. It became the cover charge.
There is a cost side that seats never had. A seat costs you almost nothing to serve. A result is different: every outcome an AI delivers carries a real cost behind it, the compute and the model calls and the work the agent does. So the price you pick has to sit above what each outcome costs you, with room to spare. The gap between what an outcome costs you and what you charge for it matters as much as the headline number, because with AI that cost is real and it can move. Watch it as closely as you watch the price. Intercom's finance chief calls this the "it's not zero" rule: the cost of an AI outcome is never nothing, so you forecast and price as if it is real, because it is.
How you package the product matters as much as the number, because the same feature feels different depending on how you sell it. HubSpot built Breeze straight into its Professional and Enterprise plans instead of selling it on the side. Rangan's view was simple: there should be one product, built around AI, not a feature bolted on.
And when you charge for a result, that result is a promise, so you have to define it before a rep ever quotes it. Intercom couldn't charge per resolution until "resolved" was clear enough that customers would accept the charge. Both sides have to agree on what they are counting, or the model has nothing to stand on.
Credits are a step on the way there, not the finish line. HubSpot's credit system broke the link between price and seat count, which it needed to do. But credits are hard to read. A customer can't feel what "100 credits" is worth the way they can feel "$0.50 per resolved chat." Credits buy you time while you find the real number. They don't replace it. HubSpot still runs credits underneath, where a resolved conversation costs 50 of them, but it now quotes the outcome, $0.50 a resolution, because that is the number a customer can actually feel.
Join companies like Zoom, DocuSign, and Twilio using our systematic pricing approach to increase revenue by 12-40% year-over-year.