Give the users of your app a chat that answers their questions from the app’s own data. The chat proposes changes, and they apply as soon as the user accepts them. AI credit use is visible per account and capped. That keeps the cost in view.
In use
The same pattern runs in production in one of our own applications: the agent answers questions in the chat panel, proposals are approved, and every account has its AI credit.
The Agentic Part
An agent in your app answers users’ questions from the data the app holds. When a user wants to change something, the agent creates a proposal, and the change applies as soon as the user accepts it. Every account has its own AI credit, and the credit check happens before any cost arises. The use of each account stays visible and capped.
Main Features
- Changes apply as soon as the user accepts them
- Answers from the own data of your app
- A chat panel directly in your app
- AI credit use is visible per account and capped
- The credit check happens before any cost arises
- The list of AI models is maintained by admins, and the access keys stay hidden
- The AI model is a simple option to choose
Use Cases
- A user of your app asks in the chat for a value from their data. The agent answers from the app's data.
- A user wants to change an entry and says so in the chat. The agent creates a proposal, and the change applies as soon as the user accepts it.
- A guest account asks the agent a few questions. The credit for that belongs to the guest account itself.
What we build for you
- The data of your app that the agent answers from
- Which changes the agent may propose
- The size of the AI credit per account
- The choice of AI model
- Colors and look of the chat in your app
