AI ticket triage in 2026: what nine IT service desks document, stage by stage

Triage is the part of the ticket lifecycle between logging and resolving: deciding what a ticket is, how urgent it is, and who should work it. In our Lifecycle Automation Census it covers two stages, 02 Categorised and 03 Routed. AI triage means the tool sets the category, priority and destination itself, rather than an agent reading each ticket or an admin maintaining a long list of keyword rules.

Two patterns stand out. Freshservice documents AI routing but not AI categorisation on the pages we read. Jira Service Management and ServiceNow are marked Partial on both stages: Atlassian lists Rovo Agents and queues without an auto-categorisation claim [Atlassian pricing], and ServiceNow describes Now Assist 'AI-guided recommendations' and AI agents that 'streamline incident management and request fulfillment' rather than automatic classification [ServiceNow ITSM] [ServiceNow AI agents]. Partial means the capability may exist but the pages we read do not describe it as automatic.

Triage is often metered or gated by plan, which matters more than whether it exists.

• monday service: AI Ticket Triage covers 1 ticket board on Standard, 3 on Pro and unlimited boards on Enterprise [monday pricing].
• Zendesk: Intelligent Triage is listed on Suite Enterprise + Copilot, which is quoted rather than priced [Zendesk Employee Service pricing].
• Freshservice: Freddy AI Copilot is a $29 per agent per month add-on [Freshservice pricing].
• Jira Service Management: Standard includes 250 Rovo credits [Atlassian pricing].
• Halo: Halo AI is included in the single per-agent price [Halo pricing].

When you compare quotes, check how many boards, queues, agents or credits the AI triage on your plan covers, and what happens when a month's volume runs past it.

Four announcements this year point the same way: less rule-writing by admins, and more model choice.

• 27 January: Ivanti introduced persona-based agentic AI for Neurons for ITSM, with a customer preview from Q1 2026 [Ivanti press release].
• 24 July: ServiceDesk Plus Cloud added Resolution Assist, which generates a resolution from similar resolved requests [ManageEngine release notes].
• 9 September: SysAid launched an AI Automations Builder. Admins describe an automation in their own words and the tool sets triggers, conditions and actions for assignment, prioritisation, categorisation and escalation [SysAid blog].
• 14 September: ServiceDesk Plus Cloud added Anthropic's Claude as an AI provider for Ask Zia and its GenAI features [ManageEngine release notes].

SysAid's builder changes who writes triage rules, not whether rules exist. The rules still need an owner and a review, so ask who maintains AI-built automations after go-live (question 9 on our buyer's checklist). All dated items are on our news page.

Vendors quote headline figures, and they rarely describe the same thing.

One figure refers to routine queries across service contexts, the other to incidents before a ticket exists. Neither says what share of your tickets will be categorised and routed correctly. Measure that yourself: our two-week trial plan has a stage-by-stage test.

1. Which fields does AI set on a new ticket: category, priority, assignee, or all three?
2. Can we see the reason for each AI routing decision, and correct it?
3. Which AI model or provider processes our tickets, and can we choose?
4. Is AI triage included on our plan, or metered by board, agent or credit?
5. What happens to a ticket the AI cannot classify?

Then read the census notes on each review, for example monday service, Halo and ServiceNow ITSM.