A practical ranking of support ticket categories by how safely and profitably AI can handle them.
Key takeaways
Password resets, status and outage checks, basic how-to questions covered in existing docs, license or seat count queries, and simple configuration walk-throughs consistently top the list across MSPs and software support teams. These categories share two traits: high volume and low ambiguity, meaning the correct response rarely varies once you know the category.
Ranking by these two traits, rather than by which category seems most technically interesting to automate, keeps the rollout focused on where the time savings are actually largest. A niche edge-case ticket type might be fun to build a workflow for, but it won't move your queue metrics the way password resets will.
Billing disputes and anything touching account termination or cancellation should stay fully human, even once your AI triage system is mature. Both categories carry outsized relationship and revenue risk if handled wrong, and both frequently involve nuance — a client's specific history, an unwritten exception previously granted — that a triage system won't reliably capture.
This isn't a permanent restriction so much as a standing default. As your knowledge base and review processes mature, some firms do introduce limited AI drafting support even here, but always with a senior team member reviewing before anything reaches the client, never as an autonomous response.
Pull the last three to six months of tickets and sort by volume and by how consistent the correct resolution has been historically. Categories that are both high-volume and highly consistent are your starting point; categories that are low-volume or highly variable are candidates to leave for later or permanently.
This exercise usually surfaces one or two categories specific to your client base that don't appear on any generic list — a particular software integration your clients frequently ask about, for instance — and those are often worth prioritizing precisely because they're specific to you.
Your ticket mix changes as your client base and product offerings change, so the ranking that made sense at launch won't necessarily hold a year later. Building a quarterly review of ticket categories into your operating rhythm keeps the automation list current rather than frozen at whatever it looked like during the initial rollout.
This is also a natural point to retire automation that's no longer earning its keep — if a category that used to be high-volume has shrunk, the review effort spent maintaining that workflow might be better redirected elsewhere.
A 30-minute call is enough to tell you whether AI pays for itself here.
