Choosing an AI model
Pro servers can pick which AI model powers their answers, from fast and inexpensive to top-shelf. The Free plan uses the default model.
Where to choose
Open AI, then Configuration and pick from the model list. The change applies to new answers immediately.
How to think about it
- The default model is tuned for support work: fast, accurate on grounded questions, and cheap enough that included usage goes a long way.
- Premium models reason better on genuinely hard questions but cost more per answer. Because usage is metered, a pricier model simply consumes your included usage faster and shows a higher per-answer cost.
A practical approach
Start with the default. If you see the AI struggle on nuanced questions that your knowledge does cover, try a stronger model for a week and compare: the dashboard shows each conversation's exact cost and your deflection rate, which is exactly the trade you are weighing.
There is no wrong choice to fear: the spending cap protects you either way.