As of 13 August 2026, AI can only partly triage IT helpdesk tickets.

Most people should hand this to a purpose-built tool.

Can you do it?

5 minutesto a draft.

30 minutesto something you’d act on.

Cost, all in£0

Skill neededchat-fluent

Who has to check ita colleague

What the alternative costsThe supplied tool information gives no price for a comparable helpdesk service.

If this goes wrong: an urgent security or access problem is placed in a low-priority queue and the resulting downtime or data exposure is yours to deal with.

What to actually do

Use a tool built for this
The route this page recommends
One pick for this task
Botpress
Botpress is a purpose-built platform for building support agents that can apply your routing rules, although you still need to configure its data access, escalation rules and human approval process.

Do it yourself

Second choice

A chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.

How to actually do it

Open your helpdesk export or ticket list and remove passwords, API keys and personal data that is not needed for triage.
Write down your priority rules, including what counts as urgent, which systems are critical, your support hours and the queues or people who receive each type of ticket.
Gather current information about known outages, maintenance windows, important devices and services, and any existing escalation contacts.
Paste the prompt into a chatbot, then replace each bracketed section with your rules, service information and ticket text.
Compare every proposed priority and queue with your written rules, and inspect every P1 or P2 ticket for evidence of outage, compromise, data loss or critical-user impact.
Send only the checked routing decisions to your helpdesk or colleague, and keep the original ticket IDs and the AI's flagged uncertainties attached.

Prompt

You are helping triage IT helpdesk tickets for a small UK business. Do not take actions, contact users, change systems or close tickets. Use only the ticket text, the business rules and the current service information provided below. If information is missing, say exactly what is missing and lower your confidence rather than guessing.

For each ticket, return a table with these columns: ticket ID, short summary, category, priority as P1 urgent, P2 high, P3 normal or P4 low, proposed queue or owner, immediate safe next step, information to request, confidence as high, medium or low, and reason. Treat suspected malware, account compromise, data loss, widespread outage, loss of access for a critical user, and risks to personal data as candidates for P1 or P2, but apply the supplied business rules and explain the reason. Do not diagnose malware, invent technical facts, promise a resolution time or recommend risky changes. Flag any ticket that needs human IT review before routing. Keep the original ticket ID attached to every row.

Business rules and queues:
[PASTE YOUR PRIORITY RULES, OPENING HOURS, ESCALATION CONTACTS AND QUEUE NAMES]

Current service information:
[PASTE YOUR SYSTEMS, DEVICES, KNOWN OUTAGES, MAINTENANCE WINDOWS AND CRITICAL SERVICES]

Tickets:
[PASTE ANONYMised TICKET EXPORT OR TICKET TEXT, REMOVING PASSWORDS, API KEYS, PERSONAL DATA NOT NEEDED FOR TRIAGE]

After the table, list the five tickets most in need of human review and state the missing fact that would change each decision.

Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

Hand it to a person
The distant third
A person who owns the outcome does this end to end, worth it when the failure is dear.

What it gets wrong

AI cannot know that a minor-looking ticket affects a key customer, a payroll run or a critical business process unless you provide that context.
AI cannot reliably distinguish a genuine security incident from an ordinary fault using sparse ticket text alone.
AI cannot see whether a suggested fix would conflict with a live outage, maintenance window or undocumented dependency.
AI cannot take responsibility for a missed escalation, delayed response or exposure of business data.

What caps this at PARTLY: judgement under ambiguity, stakes of error and context depth.

How we scored this

Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.

Axis	Score (0–2)
Output	2
Inputs	1
Verification	1
Liability	1
Effort delta	2
Total	7 / 10

The methodology and its thresholds are published in full.

FAQ

Can ChatGPT sort my IT support tickets?: Yes, it can sort ticket text into categories, suggest priorities, summarise the issue and propose a queue. Use it as a draft triage layer, not as an unattended router, because it cannot see every business dependency or assess an unclear security incident reliably.
Can AI decide which IT tickets are urgent?: It can apply clear urgency rules to the facts written in a ticket. It cannot reliably infer hidden business impact, so a person should approve urgent, security-related and ambiguous cases before they are routed.
Is it safe to put helpdesk tickets into AI?: Only provide the minimum information needed and remove passwords, API keys and unnecessary personal data. Check the chatbot or helpdesk product's data-handling terms before uploading confidential business or customer information.
Can AI automatically assign helpdesk tickets to the right person?: It can suggest an owner or queue when you provide current routing rules and keep the available queues up to date. Automatic assignment is unsafe for unclear, high-impact or security-related tickets unless a colleague reviews the decision.

Nearby answers

All Helpdesk & devices rulings

Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.

The newsletter

AI news, new answers and product picks, straight to your inbox.