ChatGPT charges for access through flat subscription tiers, where users pay a fixed monthly fee regardless of how frequently they use the product, within usage limits. GitHub Copilot charges per person, meaning companies pay based on the number of developers using the tool. Intercom Fin charges for outcomes, where businesses pay for successfully resolved customer conversations.
In terms of who will carry the risk, for ChatGPT and GitHub Copilot, customers will pay for the risk when AI is run. Because they still pay even when the AI generates incorrect or unhelpful responses. However, the providers will carry risk when AI is expensive to run given that some users may use the product with usage within plan limits. For Fin, the risk is more on the provider because unsuccessful resolutions will not generate great amout of profits. However, as the reading points out, outcome-based pricing can be challenging because defining and measuring a successful outcome can be really hard.
We would prefer seat-based pricing for our team’s product. Our product aim to help cross-functional teams identify potential conflicts in, for example, Zoom meetings. Then they can discuss disagreements through shared annotations in the product, and track whether those conflicts have been resolved. Since our target customers are mostly businesses, charging per user would provide predictable costs and allow pricing to scale as more team members join. Also, the value of our product would come from collaboration among different team members, so charging based on seat would be the best fit for us.
