How to Price AI: 3 Strategies That Actually Work
Maarten Laruelle The way you price AI matters more than you think. Most founders get it wrong because they treat all AI the same — one feature, one price. But AI isn’t one thing. The level of autonomy changes everything about how you should charge.
Not all AI pricing is the same
If your AI auto-generates reports with zero human input, that’s fundamentally different from an AI that helps a user write better emails. The autonomy gap changes the value — and the value should change the pricing model.
Three categories. Three different approaches.
1. Autonomous AI: price on outcomes
When the AI does the entire job — no human in the loop — you can price on what it delivers, not what it consumes.
Think: €0.75 per satisfied customer with an NPS above 7. Or a fee per successful resolution. The customer doesn’t care how many tokens you burned. They care about the result.
This is where value-based pricing really shines. The more autonomous the AI, the more you can charge for the outcome instead of the input.
2. Collaborative AI: subscription plus consumption
Most AI today sits here. The AI assists, but a human still drives. Think copilots, writing assistants, smart suggestions.
The model that works: a base subscription for access, plus per-action pricing for heavier usage. Something like €5–10 per output on top of the monthly fee.
This hybrid gives customers predictability (they know the base cost) while letting you capture value when they use the AI more intensively.
3. Efficiency AI: pure consumption pricing
When the AI is purely doing backend processing — tokenizing, classifying, transforming — it’s infrastructure. Price it like infrastructure.
Token-based pricing, aligned with your margins. Cost-plus, transparent, and scalable. Customers in this category are buying efficiency, not magic.
Which model applies to you?
It’s not always black and white. Your product might have autonomous features and collaborative ones in the same platform. That’s fine — you can layer models.
The core insight: the more autonomous your AI is, the more you can price on value instead of inputs. Your customers don’t care about tokens. They care about results.
Figuring out how to price the AI in your product? Let’s work through it together.