Thinking Messages: Real-time status updates while your AI agent works

The moment between a user’s question and an agent’s answer is usually a second or two. Three animated dots fill the space, but they don't say anything. Whether the agent is writing a plain response or running three API calls in the background, what users see looks identical. As reasoning models take longer to work through complex problems, that silence grows, and the dots become an even weaker signal.

Users don’t know if the agent understood the question, is retrieving their account, or got stuck in a loop. That uncertainty creates friction that has nothing to do with how good your agent actually is.

Thinking Messages replaces the generic typing indicator with custom, per-tool status messages, so users always know what's happening while they wait.

How it works

• Per-tool message configuration: Write a specific status for each tool call in your agent’s actionbook. Each tool gets its own message, written in your voice.
• Real-time display: The message appears the moment a tool is invoked and replaces the typing indicator for the duration of that call. When the tool completes, the normal response flow resumes.

Thinking Messages is part of Trust OS, turning the space between question and answer from silence into a signal.