ChatGPT, GitHub Copilot, and Intercom Fin use three different pricing models, each reflecting a different understanding of how AI creates value and who should bear the risk.
ChatGPT charges for access through monthly subscription tiers. Customers benefit from predictable costs and the freedom to experiment without worrying about every prompt. However, OpenAI bears the risk when heavy users consume more computing resources than their subscriptions cover. Meanwhile, users still pay even when the AI produces inaccurate or unhelpful responses.
GitHub Copilot charges per person, with companies purchasing licenses for individual developers. This makes budgeting and adoption relatively straightforward, especially for larger teams. However, the model assumes that each developer receives similar value, even though usage varies significantly. Companies risk paying for underused licenses, while GitHub manages computing costs through usage limits and allowances. Developers also remain responsible for reviewing AI-generated code.
Intercom Fin charges for outcomes, specifically customer conversations successfully resolved by AI. I find this model particularly interesting because customers pay for measurable results rather than simply having access to technology. Intercom takes on more financial risk when its AI fails to resolve issues. However, defining a successful resolution can be subjective, and businesses may dispute whether an interaction actually solved the customer’s problem.
For our team’s product, which helps small businesses organize tasks, prioritize responsibilities, and make decisions, I would choose a tiered subscription model. Unlike customer support resolutions, the value of a good decision is difficult to quantify. Charging per decision could discourage users from exploring alternatives, which is exactly what our product should encourage.
Kyle Poyar’s Your Pricing Is (Probably) Broken argues that every pricing model has limitations. For our product, I would combine an affordable base subscription with premium AI features and reasonable usage limits. This balances predictable costs for small businesses with sustainable growth for our team.
