Most AI pricing problems aren’t pricing problems
By the time you're debating tier structures and packaging, the decisions that constrain your pricing options have already been made.
I work at the intersection of product and pricing, which means I understand that the way you build early on quietly shapes what you can charge for, long before monetisation is on the agenda.
Three things come up repeatedly in product builds that most founders don't recognise as pricing decisions.
Whether you instrumented outcomes. If your product doesn't capture what actually changed for the customer: a decision made, a risk avoided, time saved, you can't price against it. You're left pricing inputs: tokens, calls, requests. Those are your costs, not your customer's value. The mismatch is structural, and it will show up in every commercial conversation you have.
Whether you built for value attribution. It's not enough for your product to deliver value, customers need to be able to see the connection between what they paid for and what changed. That's a product architecture decision: if you surface the before and after, and make the causal chain visible. Without it, outcome-based pricing has nothing to point at, and usage-based pricing feels like pricing for activity rather than impact. Attribution isn't a reporting feature. It's the commercial foundation of any pricing model that claims to reflect value.
Whether customers can govern their own usage. Most customers cannot reliably predict how they will use AI six months from now. If your product doesn't give them meaningful control over their consumption, spending becomes unpredictable and anxiety sets in. Anxious customers don't expand. They churn.
These constraints aren't permanent. Products evolve, and as they do, new pricing options open up, and models that were impossible to justify or operationalise at launch can become viable.
But that optionality only opens up if your monetisation layer can move at the same pace. A billing infrastructure that locks in today's model makes every subsequent pricing change an engineering project. That lag is costly: misaligned invoices, uncomfortable renewals, and churn that felt avoidable.
The founders who move fastest treat pricing iteration as a natural consequence of product iteration, and choose infrastructure that makes that possible.
If this is a problem you're sitting with, let's talk.