This article is auto-synced from its in-app version in Tai.
A flexible, general-purpose AI assistant — not tied to your operational data — for open-ended questions, drafting, and reasoning tasks.
When you'd use this
Use the TAI Assistant when you want a capable general assistant for tasks that don't need your TripBuilder data — brainstorming, drafting copy, summarising a document, explaining a concept, or doing quick research. It complements the TripBuilder Ops & BI Agent, which is grounded in your booking data: reach for the TAI Assistant when the question is open-ended rather than about your bookings.
It can also answer questions about TAI itself — what agents, skills, connectors, and knowledge are, and how to find your way around the Workbench — and coach you through writing or improving the system prompt for an agent you're building.
How to use it
In the Workbench sidebar, click Chats.
Select TAI Assistant from the agent picker.
Ask your question or describe the task. It can also generate charts and downloadable reports when that helps.
Tips and limits
It's general-purpose — it has no access to your bookings, connectors, or other data sources unless one is configured for it. The one source it has built in is TAI's own help center, which is what lets it answer questions about TAI.
Both of those are for where you work in TAI — Chats and the Terminal. Embed this agent in your own site with the Chat UI SDK, or call it from your own backend, and it drops them: no help-center answers, no prompt coaching, and a greeting that doesn't mention TAI. Your visitors get a general assistant, not a tour of our product.
It coaches you on your agent's prompt but can't edit it for you — copy its suggestions into the agent's System Prompt field yourself.
For a serious pass on a prompt, use Improve on the agent's own page instead — it rewrites the stored prompt rather than the excerpt you pasted into chat, and it's available to everyone who can author an agent. If you're a company administrator, an Advisor run goes further: it evaluates the agent's real conversations and proposes changes you can apply.
Related