Your Best Technician Is Your Biggest Single Point of Failure

Every MSP has one. Usually two.

The senior technician who fixes the thing nobody else can fix. The one whose name gets said out loud when a ticket goes sideways. The one you’d genuinely struggle to replace.

And in almost every MSP I’ve looked at, that person is also the single largest hole in the organization’s knowledge, because the way they work leaves no trace.

An MSP told me exactly this recently, without prompting. Their two most senior technicians routinely resolve escalated issues off-system — a phone call, a screen share, a walk over to someone’s desk. The problem gets solved. The ticket gets closed. Nothing is written down about how.

They weren’t describing a discipline problem. They were describing their best people being efficient in the way that has always been rewarded.

Why it happens (and why yelling about it won’t fix it)

The standard response is a documentation mandate. It doesn’t work, and it doesn’t work for structural reasons.

Escalation is a conversation, not a workflow. A tier-1 tech pings a senior in Teams. The senior says “check the DNS suffix on the VPN profile.” Ninety seconds. Problem solved. There is no natural moment in that exchange where anyone opens a ticket to write it down — the ticket belongs to the tier-1 tech, and the knowledge belongs to the senior.

The person who knows isn’t the person who documents. The tier-1 tech writes the resolution note, and writes what he did, not what he learned. “Escalated to Nathan, resolved.” The actual diagnostic reasoning never enters the record.

Nothing rewards writing it down. Utilization targets reward closing tickets. Nothing on any MSP scorecard I’ve seen rewards making the next person’s ticket faster.

Seniors are the most time-poor people in the building. They’re the escalation path for everyone. Asking them to add a documentation step to every interrupt is asking the busiest person to absorb the most new work.

So it’s not a compliance failure. It’s an incentive design failure, and it’s been baked into how MSPs run service desks for twenty years.

What it actually costs you

The cost used to be tolerable. It isn’t anymore.

The same escalation, repeatedly. Because nothing was captured, the next tier-1 tech escalates the same issue in three weeks. Your senior solves it again. Multiply by every recurring issue in your client base.

Onboarding takes a year. New technicians learn by osmosis because there’s nothing to read. Your ramp time is a direct function of how much of your operational knowledge is undocumented.

Escalation loops. When a ticket bounces between tiers, nobody can reconstruct what’s been tried. I’ve watched a technician spend forty minutes reading a ticket with over a hundred entries just to find out where things stood.

Key person risk on your balance sheet. If half your operational capability lives in two people’s heads, that’s a real discount at exit, and any competent acquirer will find it in diligence.

And now: your AI can’t learn. This is the new one. Every automation, every resolution suggestion, every similar-ticket match is built on the record of how problems were solved before. Off-system resolutions are invisible to all of it. You can buy the best service desk AI on the market and it will be permanently blind to your hardest 5% of tickets — the exact ones where help would be worth most.

Four fixes that don’t require anyone to become a technical writer

Because that’s the constraint. Whatever you implement has to cost your senior technician under sixty seconds, or it won’t survive contact with a busy Tuesday.

1. Make the escalating tech capture the escalation

The tier-1 technician who escalates writes a short note at the point of escalation: what they tried, what the symptoms are, who they’re escalating to and why.

This sounds like it solves the wrong half of the problem — and partly it does, since the resolution still comes from the senior. But it does two useful things: it creates a structured moment in the workflow where escalation becomes a recorded event rather than a Teams message, and it gives any AI layer enough context to connect the ticket to similar past cases.

One MSP we work with landed on exactly this as their practical starting point. It’s a workaround, not a complete answer. It’s also the one that actually got adopted.

2. Voice, not typing

Let the senior speak the fix into a phone for thirty seconds and have it transcribed onto the ticket. The barrier to documentation isn’t willingness — it’s the cost of switching from thinking mode to writing mode. Speech removes most of it.

Rough transcription with the right technical detail beats polished prose that never gets written.

3. A resolution note standard, enforced at close

One page. Symptom, root cause, fix, steps. In the resolution field, customer-readable.

At one partner, a named senior person’s resolution-note framework was cited to me as a genuine competitive advantage for that organization — because it meant their historical ticket data was actually usable. That’s not a documentation policy. That’s an asset.

4. Make the AI do the drafting, and the human do the correcting

The realistic division of labour: let the system draft the knowledge article from the ticket, and have the senior technician spend sixty seconds fixing what’s wrong.

One design requirement matters more than any other here: the draft has to be grounded in the ticket record itself — the exact command, the exact path, the exact order — not a competent-sounding summary of the troubleshooting approach. A paraphrase of the fix is not documentation. Sixty seconds of correction from the person who knows is sustainable; twenty minutes of authorship from the busiest person in the building is not.

This is exactly the gap Mizo’s Documentation Agent closes: it captures every resolution into IT Glue, Hudu, or SharePoint automatically, drafted from what actually happened on the ticket — and the QA Agent flags the “escalated, resolved” notes at close, while the technician still remembers the specifics.

The reframe worth sitting with

We’ve spent years telling ourselves that the senior technician who can fix anything is our biggest asset.

They are — and simultaneously they’re a concentration risk you’d never accept anywhere else in the business. You wouldn’t run a single client at 40% of revenue without worrying about it. You wouldn’t keep one copy of a backup.

Undocumented expertise is the same exposure wearing a friendlier face.

The organizations pulling ahead right now aren’t the ones with the best technicians. They’re the ones whose best technicians’ knowledge is available to everyone else, including their automation. That’s a different kind of asset — one that compounds instead of walking out the door.

Start here

Pull last month’s escalated tickets. Read the resolution notes on the ones that went to tier 3.

If most of them say some version of “escalated, resolved,” you already know what your senior technicians’ knowledge is worth to anyone but themselves. It’s worth exactly nothing to the next technician, and exactly nothing to any tool you buy. (Mizo’s QA Agent runs exactly this audit for you, on every ticket, at close — so you never have to do it retroactively again.)

That’s fixable. It just isn’t fixable with a memo.

From a former MSP operator who was, for several years, his own single point of failure.