AI Insights on a Ticket
Open a ticket. The 'AI Insights' panel sits under the ticket description.
The AI service agent analyses the tickets it picks up and writes its conclusions into this panel. It tells you what it believes caused the ticket, whether it has seen anything like it before, and what it suggests doing about it β so a technician taking over an escalation starts with the analysis already done.
You can read the AI's root cause hypothesis, its search for similar past tickets and its suggested resolution, and find the same analysis preserved in the activity trail.

The panel
The panel is headed 'AI Insights' and tagged 'β‘ Powered by AI'. When the AI has analysed the ticket it shows three blocks.
| Block | Description |
| π Root Cause Hypothesis | The AI's explanation of what it believes caused the ticket, written in plain language. |
| π Similar Past Tickets | Matching tickets the AI has found. Reads 'Looking for similar ticketsβ¦' while the search runs. |
| β‘ Suggested Resolution | What the AI recommends doing, including any checks to make or people to confirm with. |
When there is no analysis
If the AI has not analysed the ticket the panel reads 'No AI insights yet for this ticket.'

Where the analysis is also recorded
The same root cause hypothesis and suggested resolution are written into the 'Activity' panel as an entry authored by 'IT AI Agent', with a timestamp. The analysis therefore stays on the ticket permanently, in order with everything else that happened to it.
Read the AI's analysis of a ticket
- Open the ticket from any ticket list β escalated tickets are the usual case. See Escalated to Technicians.
- Read the 'AI Insights' panel under the ticket description.
- Start with 'π Root Cause Hypothesis' to understand what the AI thinks happened.
- Check 'π Similar Past Tickets' for anything already solved.
- Follow or adapt 'β‘ Suggested Resolution'.
- Record what you actually did with a reply or an internal note. See Reply to a Ticket or Add an Internal Note.
Note: the insight text is generated from the ticket's own content, so it differs for every ticket. Treat the suggested resolution as a starting point and verify it before acting.
