IT Ticketing System: How a Ticket Should Move, and Where It Stalls

A user emails at 10 a.m. on a Tuesday: the shared printer in accounting won't scan to email. Your tech replies asking for the printer's model and whether anything changed on the copier last week. Then the tech moves on. The ticket stays "In Progress." Nine days later the client's office manager calls asking why nobody has fixed the scanner, and your tech says, truthfully, "We were waiting on you."

Nobody did anything wrong, exactly. The ticketing system recorded the question. What it didn't record was whose move it was. That gap, more than any feature on a vendor's comparison grid, decides whether an IT ticketing system works.

What is an IT ticketing system?

An IT ticketing system is software that turns every support request into a tracked record, a ticket, with a requester, a description, a priority, an owner, a status and a history of everything done on it. It exists so that no request depends on someone's memory: anyone on the team can open the ticket and see what was asked, what was tried, and what happens next.

Most of the pages that rank for this term are written by vendors of internal IT service management tools, such as ServiceNow, Freshservice or Jira Service Management. Those tools serve one organization's employees. If you run a managed service provider (MSP), your IT ticketing system is almost always the ticketing module of your professional services automation (PSA) tool: ConnectWise PSA, Autotask, HaloPSA or Syncro. The difference is the contract. An MSP ticket belongs to a client company, carries that client's agreement and becomes a billing record. We covered that difference, and why the phone is the channel that leaks, in MSP ticketing systems: the phone call test most reviews skip.

This post is about something that applies to both kinds of shop: how a ticket should move from open to closed, and the places where it stops moving.

How a ticket moves: the lifecycle in seven steps

Every ticketing system, from a free help desk to a full PSA, runs the same lifecycle. The names change. The steps don't.

  1. Intake. The request arrives by email, portal, chat, phone, or an alert from your remote monitoring and management (RMM) tool, and becomes a ticket tied to a company and a contact.
  2. Classification. Someone sets the type, subtype or category, so the ticket lands on the right board or queue and shows up correctly in reports.
  3. Prioritization. The ticket gets a priority, usually from how many people are affected and how urgent the work is for them.
  4. Assignment. A person or team owns it.
  5. Work and communication. The tech troubleshoots, logs notes and time, and talks to the requester.
  6. Waiting. The ticket pauses for the customer, a vendor, a part or a scheduled visit.
  7. Resolution and closure. The fix is recorded, the requester is told, and the ticket closes.

Vendor guides spend most of their words on steps 3 through 5: SLA timers, assignment rules, dashboards. In practice those are the steps that work best, because a tech is actively looking at the ticket. The tickets that go wrong usually go wrong at step 1, step 6 or step 7, where nobody is looking.

Where tickets stall, and why

At intake: the ticket that is too thin to work

An email that says "scanner broken" becomes a ticket titled "scanner broken." Before anyone can work it, someone has to go back and ask which scanner, which user, what error, since when. If that question goes out late, the clock has been running on a ticket nobody could start.

The phone is worse. A call creates no ticket at all unless the tech stops and writes one. Some calls get a ticket typed from memory later; some never get one. If you want to see how big that gap is in your shop, compare a week of call logs against a week of tickets. The MSP ticketing systems post walks through that test.

While waiting: the status that doesn't say whose move it is

This is the scanner story. A status list like "New, In Progress, Closed" can't tell you whether a ticket is stuck on your team or on the client. So the ticket looks active, sits untouched, and nobody follows up, because everyone assumes someone else will.

The fix is boring and it works: make waiting a separate status, and make it specific. "Waiting on Client," "Waiting on Vendor," "Scheduled" and "Waiting on Parts" each tell the next person what has to happen before work resumes, and who has to do it.

Your PSA also uses these statuses to run the SLA clock, so getting them right changes your reported numbers, not just your queue. Autotask is explicit about it. It tracks four SLA events, First Response, Resolution Plan, Waiting Customer and Resolved, and a status mapped to the Waiting Customer event "pauses the clock while you are awaiting action or information from a customer" (Autotask: introduction to service level management). The default mapping hides a trap, though: any status you add that isn't mapped explicitly counts as First Response, not as a pause (Autotask: mapping ticket statuses to SLA events). If you create "Waiting on Vendor" and never map it, Autotask keeps the clock running while you wait on the vendor. HaloPSA lets a status put a ticket's SLA on or off hold, set under Configuration > Tickets > Statuses (HaloPSA guide: statuses). ConnectWise PSA and Syncro each have their own status settings too. Whichever PSA you run, open the status list and check what every waiting status does to the clock.

There's a second trap: a paused clock is comfortable. Once a ticket is in "Waiting on Client," your SLA report stops caring about it, and so does everyone else. The client, meanwhile, still has a broken scanner. A waiting status needs a follow-up rule attached: who chases the client, how often, and what happens when they never answer. Without that rule, "Waiting on Client" is where tickets go to age.

At close: the note that says "fixed"

The last place a ticket fails is the resolution note. "Resolved, closed" tells the next tech nothing when the same user calls back in a month with the same problem. On a PSA, the note also often travels. Autotask's help on adding a time entry says summary notes "may be customer-facing and appear on invoices, notification emails, and in the Client Portal." We wrote up what a usable one looks like in what a good ticket note looks like.

How to design statuses that keep tickets moving

If you only change one thing in your ticketing system this quarter, change the status list. A few rules that hold up across PSAs:

  • One status per owner of the next move. Your team, the client, a vendor, a calendar. If two statuses mean the same next move, merge them. If one status hides two different next moves, split it.
  • Name waiting statuses for who you're waiting on. "Pending" means nothing. "Waiting on Client" means someone at the client has to reply.
  • Tie every waiting status to the SLA on purpose. Decide for each one whether the clock should pause, then check that your PSA actually does that. Waiting on the client usually pauses it. Whether waiting on a vendor should is a contract question, so put the answer in your agreements instead of leaving it to whatever the default is.
  • Give every waiting status an exit. A number of follow-ups, a last-notice message, and a status the ticket moves to when nobody replies, whether that's closed or back in front of a tech.
  • Keep "In Progress" honest. It should mean someone is working the ticket now or is about to. If tickets sit there for days, the status list is hiding a waiting state you haven't named.
  • Review stale tickets by status, not by age. A weekly look at everything in each waiting status for more than a few business days finds the scanner tickets before the client calls about them.

What to look for when choosing an IT ticketing system

If you're picking a system or reconsidering yours, ask about the stalls above, not only the feature list.

  • Intake from every channel your clients use. Email and portals are table stakes. Ask how a phone call becomes a ticket and who has to do it. For RMM alerts, ask how you keep them from flooding the board; we covered that tradeoff in should RMM alerts create tickets?
  • Configurable statuses per board or queue, each with its own SLA behavior. Different boards wait on different things. A projects board and a help desk board shouldn't share one status list.
  • Customer communication from inside the ticket. Replies, reminders and resolution notices should go out from the ticket and be recorded on it, so the history is complete.
  • Notes and time on the same record. For an MSP this is what turns a ticket into a billable, reportable record. It's the main reason MSPs run a PSA rather than an internal IT help desk; see our PSA software guide for how ConnectWise PSA, Autotask, HaloPSA and Syncro compare.
  • Reporting you'll actually run. At minimum: open tickets by status and age, first response time, time to resolution, and reopened tickets.

Where AI fits in a ticketing system

AI is most useful at the three stall points, the steps where no human is looking. It doesn't need to replace the ticketing system to do that. It needs to read and write the tickets you already have.

That's how Vight works. It runs on top of ConnectWise PSA, Autotask, HaloPSA and Syncro rather than replacing them, and it targets those same three places:

  • Intake. Vight pulls each recorded call from your phone system, matches it to the client and ticket, and posts a summary of the issue, the troubleshooting and the resolution as a ticket note, with a time entry. Its ticket triage matches each new ticket to a scenario you set up, emails the requester for the details they left out, posts an internal note with what's answered and what still needs confirming, and moves the ticket to a ready-to-work status you choose. Ticket agents can also rewrite vague titles, set the type and subtype, and merge duplicates.
  • Waiting. A client follow-up agent starts when a ticket moves into one of the waiting statuses you pick, such as "Waiting on Client." It sends one reminder each weekday, written from the ticket, and stops as soon as the customer replies or the ticket leaves the waiting status. If nobody answers after the last reminder, it closes the ticket, or escalates it to a technician when closing would leave the client with a real problem such as an outage or a security issue.
  • Close. Because call summaries and work-session notes are written from what actually happened, the resolution on the ticket says what was changed, not just "fixed."

None of this works without a status list that means something. An agent that chases "Waiting on Client" tickets is only as good as your team's habit of putting tickets there.

Common questions

What's the difference between an IT ticketing system and a help desk? A help desk is the team, or the function, that takes support requests. The ticketing system is the software that records them. Most help desk software is a ticketing system plus extras like a knowledge base, a portal and reporting.

What's the difference between an IT ticketing system and a PSA? A PSA includes a ticketing system, and adds what a service business needs around it: client agreements, time tracking, billing, projects and per-client reporting. Internal IT teams rarely need those extras. MSPs almost always do.

Should the SLA clock stop when a ticket is waiting on the customer? Usually yes, since the delay isn't yours. But pause it through a status that's mapped on purpose, and attach a follow-up rule so paused tickets still get chased.

How many statuses should a ticket board have? As few as it takes for each status to name one next move and one owner. For most service desk boards that's a new status, one or two working statuses, three or four specific waiting statuses, and a closed status.

If you'd like to see how calls, triage and client follow-ups look on your own PSA tickets, book a demo and we'll walk through it on ConnectWise PSA, Autotask, HaloPSA or Syncro.

See Vight document every call and ticket in your PSA