AI products use different pricing models. Here I examine where ChatGPT, GitHub Copilot, and Intercom Fin puts the risk.
ChatGPT charges for access through monthly subscriptions. Its subscription tiers give users access to models, tools, and higher usage limits for a recurring fee. Customers pay a fixed fee, whether they use it a lot or a little. The provider takes on the risk of rising usage and compute costs.
GitHub Copilot charges per person. Businesses pay for each developer using the product. The value proposition is increased individual productivity. GitHub gets predictable, recurring revenue, while the customer carries the risk that some developers use Copilot less than expected or gain little productivity from it.
Intercom Fin charges per outcome, specifically per resolved customer conversation, such as a resolved support conversation. This aligns the price with measurable business value. Intercom takes on more performance risk because unsuccessful attempts aren’t billed, while customers face potentially unpredictable costs as successful usage grows.
For our product, a Slack bot that lets teams answer questions about their projects and progress without constant check-in meetings, I would choose a subscription model based on team size.
We’re selling ongoing access to shared project knowledge, not individual AI usage or a guaranteed outcome. A predictable monthly fee makes budgeting easier and encourages teams to use the bot freely. This aligns with our goal of making collaboration easier.
