The ticket, onboarding and reporting admin costing UK MSPs their billable week

Every MSP owner knows the shape of the problem even if nobody has put a number on it. The contracts and the projects are what you bill for. The ticket queue, the onboarding runs, the documentation and the client reports are what actually fill the week - and none of it shows up on an invoice.

That gap is about to get wider, not narrower. The government's Cyber Security and Resilience Bill is set to bring medium and large managed service providers ("relevant MSPs") into UK cyber regulation directly for the first time - incident reporting, supply-chain oversight, attached to MSPs specifically, not just their clients. It's a draft Bill, published 12 November 2025, and not yet law. At the same time, Openreach's PSTN and ISDN switch-off, now set for 31 January 2027, is landing on service desks as a client-by-client migration project: discovery calls, documentation, testing and sign-off, repeated across every client still on a traditional line. Neither of those is admin you can wave away, and both add tracking, evidence and reporting on top of a week that is already full.

The admin that sits around every contract

Engineers get paid to fix things and build things. In between, someone still has to triage whatever lands in the queue, onboard new client staff and offboard the ones who leave, keep documentation and runbooks current, chase quotes for hardware and licences, and pull together the SLA and client reports that keep the relationship healthy. Most of it is invisible to the client. All of it takes real hours, and those are the hours that go missing from the timesheet at the end of the week.

Stack a regulatory change that names MSPs specifically on top of a telecoms migration that touches every client account, and the admin pile does not sit still. It grows one ticket, one migration and one onboarding form at a time, until reporting season arrives and someone has to explain where the month went.

Where the hours actually go

We don't put one headline figure on MSP admin, because the mix decides it - a break-fix-heavy shop loses time differently to a fully-managed, project-led practice. What we see consistently across the UK service desks we assess is the same handful of drains, and in a business of this size they typically run something like this:

• Ticket triage and routing - reading each new ticket, judging priority, categorising it and getting it to the right engineer. In a small-to-mid service desk this often runs to somewhere in the range of 5-9 hours a week across the team, almost all of it low-value reading and re-typing rather than fixing anything.
• Client onboarding and offboarding - discovery, documenting the environment, setting up accounts and access, and the reverse when someone leaves. For a business taking on a steady trickle of new seats and starters, this can absorb 3-6 hours a week, and it is exactly the kind of work that gets rushed and then comes back as a follow-up ticket.
• Documentation and runbooks - the client documentation, network diagrams and procedures everyone relies on and nobody has time to keep current. Left unmanaged, this is less a fixed weekly cost and more a debt that quietly compounds until an audit, a new starter or an incident exposes how out of date it is.
• SLA and client reporting - the monthly pull-together of tickets closed, SLA performance and the value narrative, repeated for every managed client. For a desk running a dozen or more managed contracts, this regularly swallows a full day or more around month-end alone.
• Quoting, procurement and alert triage - hardware and licence quotes, supplier back-and-forth, order tracking, and sifting genuine monitoring alerts from the noise. Individually small, but each one is a context-switch away from chargeable work, and across a week they add up faster than most desks track.

There's a sixth drain that doesn't fit neatly into a weekly figure: Cyber Essentials and Cyber Essentials Plus recertification, which many MSPs hold themselves and increasingly have to evidence on behalf of clients bidding for public-sector work. It's an annual cycle rather than a weekly one, but the evidence-gathering behind it - patching records, access reviews, configuration screenshots - lands in exactly the same pile as everything else above, usually just before a deadline.

None of those figures are a benchmark to hit. They're a starting point for a conversation about your business specifically - which is exactly what our AI workflow assessment for IT MSPs is built to map, using your actual ticket volumes and client count rather than a sector average.

What's safe to hand to AI, and what stays with an engineer

The caution here is well founded. You hold the keys to your clients' systems, so a tool that leaks data or acts without a human checking it first is not a risk worth taking. That's the first thing we look at in the assessment, before anything else.

This matters more for MSPs than most sectors, because the data passing through a support ticket is very often the client's own customer data, not just yours. We only point at business-account AI plans that keep inputs private and are not used for training, and where UK or EU data residency is available, we flag it as a checkbox worth ticking rather than an afterthought.

In practice, the safe ground tends to be the drafting layer rather than the decision layer. AI can read an incoming ticket and draft a summary, suggest a category and priority, and put together a holding reply, with an engineer approving before anything is sent. It can draft client documentation and runbooks from your existing ticket history and configuration data, so they stop drifting the moment they're written rather than replacing the person who owns the accuracy. It can draft the commentary in a monthly SLA report from your PSA and RMM data - whether that's Autotask, ConnectWise, Halo or NinjaOne - while the numbers themselves still come straight out of the system of record.

The decision itself doesn't move. Nothing gets escalated, actioned on a client system or sent externally without an engineer's sign-off, and the assessment flags the specific data-protection and client-consent points worth checking before any of this touches a live environment. The tools stay in your stack, the responsibility stays with your team, and AI takes the reading and the first draft off the desk rather than the judgement call.

One task, before and after

Take ticket triage on its own, since it's usually the biggest single drain and the easiest to see the shape of. Before: a new ticket lands, an engineer stops what they're doing to read it, works out roughly what it is, guesses a priority, picks a category from a list that half the team ignores anyway, and either replies immediately or leaves it in the queue for someone else to triage again later. Multiply that by every ticket in a busy week and the interruption cost is often bigger than the triage time itself.

After: the ticket lands, AI drafts a one-line summary, suggests a category and a priority based on the wording and the client's contract, and drafts a short holding reply acknowledging the issue. An engineer glances at the three lines, agrees or overrides in seconds, and the reply goes out. The reading and the first draft are gone, and the send button stays exactly where it was, with an engineer.

The assessment runs that same exercise across onboarding, documentation and reporting, then ranks the drains by what's actually worth fixing first for your business - not a generic priority list borrowed from another desk's stack.

Find your own number

If you want a rough steer before anything else, our admin cost calculator will give you a general sense of what non-billable admin is costing across a business your size. It's a starting estimate, not a diagnosis - useful for a gut check, not for planning a rollout, and it won't know your PSA, your client mix or your contract types the way an assessment will.

The real number comes from the 60-minute assessment. We sit down with your business, map where the non-billable hours actually go across the service desk, engineers and account management, and hand back a report built around your tools and your client mix: named tools, step-by-step recipes, and a prioritised plan, not a template. If it doesn't surface at least 5 hours a week worth reclaiming, the fee comes straight back.

If deal and pipeline admin costs you more hours than the service desk does, our piece on the CRM and workflow gaps that cost UK service businesses is worth a read alongside this one - a lot of the same drafting-versus-deciding logic applies whether the queue is tickets or opportunities.

Not ready for the full assessment yet? Start with our free two-minute quiz for a first read on where your business's admin is heaviest, then book the assessment to see where your hours are actually going.