Support  ·  Your other tools

AI help

Turn on the AI buttons for techs, with your own Claude or Azure OpenAI key.

AI help adds buttons to tickets that summarize them, draft a reply, suggest a category, and turn a solved ticket into a knowledge base draft. It also lets a knowledge base search that finds little try other ways of saying it. It’s off until you turn it on, and nothing runs until someone clicks a button.

Turn it on

1. Tick Turn on AI help.
2. Pick the AI service: Claude (Anthropic) or Azure OpenAI.
3. Paste a Key from your own account with that service, so the cost goes to you. It’s stored encrypted.
4. Fill in the Model. For Azure, that’s your deployment name, and you also need the Azure endpoint.
5. Click Save, then Test it under Check it. The test sends one tiny question with no ticket data.

Only techs see the buttons, on tickets they work. A draft reply lands in the reply box for the tech to read and change. AI help never sends anything to a requester on its own.

Choose what it can see

• Send internal notes too is off by default, so only what the requester can see goes out. Summaries are better with notes, but notes can hold things you’d rather keep in house.
• Which departments: untick Every department to pick. A department that handles sensitive requests (HR, say) can stay off.

The What gets sent box spells out exactly what each button sends. Email addresses, files, Staff only and All techs articles, and other tickets are never sent. Every use goes to the audit log.

Check your organization’s own rules on AI before turning it on.

Turn it off

Untick Turn on AI help and save. Remove the key deletes the saved key and turns AI help off.