Hi Community đź‘‹
Quick thought: customer sentiment is already available in Jira Service Management (Positive / Neutral / Negative). It’s useful, but right now it mostly stays as a “nice signal” on the ticket.
What I’d love Atlassian to do next is make Sentiment actually usable for triage and prioritization in a native way without having to build fragile workarounds.
What I mean by “native”
Not “AI changes priority randomly”.
More like A priority suggestion banner:
“Sentiment is Negative (high confidence). Consider raising priority / moving to an escalation queue.”
Optional admin-controlled actions (opt-in):
“If sentiment stays Negative for 2 customer replies → set priority to High + notify the right queue/on-call.”
Make it easy to use in queues/automation with a clear audit trail and predictable behavior.
Why I think this matters
A lot of escalations aren’t “impact/urgency” issues at first, they’re frustration issues. Two tickets can look identical on paper, but one customer is fine and the other is clearly escalating. Sentiment is often the earliest signal.
Guardrails I’d want (so it doesn’t become chaos)
Confidence threshold (don’t act on weak signals)
Only analyze customer-side content (portal/email replies), not internal notes
Transparency (show what triggered the sentiment)
Opt-in per project/request type
No silent changes unless explicitly configured
Curious what you all think
Would you use sentiment-based priority suggestions?
Would you allow automation to act on it, or keep it suggestion-only?
Biggest concerns: false positives, language support, compliance/privacy?
Interested to hear how others handle “emotional urgency” today manual priority bump, special queues, escalation rules, etc.
Arkadiusz Wroblewski
1 comment