Skip to content
Monitor & improve

Tests and Compare

Save scenarios that check the agent still answers well, and compare prompts and models before switching.


Tests

  1. Go to Tests and click New test. Pick the agent.
  2. Write the customer's message, like “Can I see the doctor tomorrow?”, and optionally the conversation so far.
  3. Add checks: words the reply must mention or must never say, a webhook it must call (or that it must call none).
  4. Optionally add success criteria in your own words, like “Confirms the order number before acting”. The AI reads the reply and decides whether it meets each one, with its reason.
  5. Click Run. Run all of an agent's tests after changing its prompt, model or webhooks.

Compare

On the agent's Compare tab, run its saved tests on two or three prompts or AI models side by side, and see replies, pass rate, speed and cost before you switch. Costs suggests cheaper models worth comparing.

Runs use credits like the test chat.

Related guides