The AI model, chosen in the conversation
In a conversation you open with a worker, a menu at the bottom right chooses the model that answers, from those this worker is allowed to use. The choice holds for the whole conversation.
- A team member can only choose from the models the workspace owner has named for this worker.
- Under each answer, a discreet line names what answered and the model the answer reports, or “model not reported”.
- The model you chose is never silently replaced: if it can no longer answer, the message is refused and the reason is shown.