In this section (11)
- AI in ChaosDesk
- Fair use and usage
- AI and your data
- Draft replies with the copilot
- Translate conversations
- Intent, sentiment and language detection
- Turn AI features on or off
- Set up the AI agent
- Teach the AI agent with procedures and actions
- How ChaosDesk keeps AI answers safe
- Fill fields, predict CSAT and suggest macros with AI add-ons
Fill fields, predict CSAT and suggest macros with AI add-ons
Entity detection fills ticket fields from what customers write, predictions flag unhappy customers, and macro suggestions turn repeated replies into macros.
Available on Starter, Growth, Pro, Enterprise. Compare plans
The AI add-ons are three extra features that come with the copilot. Admins find them in Admin Center, Integrations, AI & translation, on the Add-ons tab.
| Add-on | What it does |
|---|---|
| Entity detection | Reads a new ticket and fills the fields you choose, such as an order number or a product |
| Predicted CSAT and escalation risk | Scores open tickets so you can spot unhappy customers early |
| Macro suggestions | Finds replies agents keep typing by hand and drafts a macro for each |
Entity detection
- Turn on Entity detection in the AI settings.
- On the Add-ons tab, choose Add field and pick a ticket field: text, number, date, pattern or drop-down. You can add up to 10.
- For each field, write What to look for, for example “the order number, like ORD-12345”.
- Choose Save.
When a ticket comes in, AI reads what the customer wrote and fills the fields that are still empty. Each value is checked like any other field value: a drop-down only takes one of its options, a date must be a date. A value that doesn’t fit is dropped.
The change shows in the ticket history as a system change, so your triggers can react to it. Entity detection runs in the background and does not use your agents’ copilot actions.
Predicted CSAT and escalation risk
When this is on, ChaosDesk scores every open ticket and writes two drop-down fields:
- AI: predicted CSAT: likely good, uncertain or likely bad;
- AI: escalation risk: low, medium or high.
The scores come from the ticket’s sentiment, words such as “manager”, “complaint” or “cancel”, the number of customer messages, how long the customer has waited, whether the ticket was reopened, and its priority. The CSAT prediction learns from your own good and bad ratings of the last 90 days.
Use the two fields in views, triggers and reports. Agents see both, with the reasons, in the copilot’s Auto assist tab.
Predictions make no AI model call, so they cost nothing extra.
Macro suggestions
- Turn on Macro suggestions in the AI settings.
- On the Add-ons tab, choose Find new ideas.
ChaosDesk looks at agents’ public replies of the last 60 days. When three or more similar replies on different tickets are not covered by an active macro, AI drafts one generic macro with placeholders.
For each suggestion you can Show reply to see the original replies, then Create macro or Dismiss. A created macro starts inactive: review it and switch it on in Admin Center, Business rules, Macros.
Not to be confused with the macro menu’s suggestions, which show existing macros used on similar tickets.
Your data
Entity detection and macro suggestions send ticket text to the AI provider with personal data masked first, as described in AI and your data. Their use shows in AI usage by feature.
Related
Was this article helpful?
Thanks for letting us know.
Sorry about that. Tell us what was missing and we will improve the article.