By Greg Allan, tradesman and creator of Data First Coaching
What office work is quietly costing a mechanic workshop its margin?
The paperwork that follows every logbook service, roadworthy and warranty job usually ends up with whoever has the least spare time in the workshop, and that is normally the owner. Job notes get written up at the end of a long day from memory instead of the scan tool readout, invoices wait for someone to catch up, and the real cost of a job on the hoist gets guessed rather than worked out. None of that shows up as one big problem because it is spread across the week in small pieces, but it adds up to real margin leaking out quietly, and a comeback that should never have happened because the job note was never right in the first place.
Why does the approval wait leave a hoist sitting idle?
A technician finds extra work during a service, brakes worn past a safe limit, a tyre that won't pass a roadworthy, a part that needs replacing before the car or ute goes back on the road, and the job stops until the customer says yes. The car sits on the hoist, the technician moves to whatever else is free, and the owner is meant to be chasing that customer between running diagnostics on the next job in. Most workshops lose real hours this way every week, not because the work itself is hard, but because the follow up has no one dedicated to it and keeps falling behind the next thing on the bay.
What's the difference between AI that answers and AI that works for a workshop?
There are two very different things being called AI right now, and knowing which one a workshop actually needs matters more than the brand name on it. One kind answers a question: it can tell you what a good job note should say or suggest wording for a message chasing approval. The other kind, an agent, goes and does it: writes the note, sends the message, books the follow up and moves the job along without you touching it. An AI agent is software that does a job the way you would, using your own prices, your own wording and your own rules, not a generic script written for every workshop at once.
Can an AI agent write up the job note and invoice line after a service?
Yes, and this is where it saves the most non-billable time in a mechanic workshop. An AI agent trained on the workshop's own job history, its past logbook services, repairs and warranty claims, can take a technician's rough on-site notes and turn them into a proper job note and the matching invoice line, picking pricing based on what similar jobs have actually charged before rather than a rule of thumb. That is the same admin a service advisor does after every car comes off the hoist in a mechanic workshop, except it happens as the job finishes instead of being chased up at the end of the day.
How does an agent know whether a job on the hoist actually made money?
It works from the workshop's own wages and overhead numbers rather than a feeling, which is why it tells the truth when memory does not. Picture a workshop owner who wanted one answer: what a billable hour on the hoist truly cost to run once wages, parts margin and every non-billable hour were accounted for. Spread across every job the usual way, the cost looked so high that nearly every job on the board seemed to barely break even. Strip the non-billable time out altogether and the number dropped so low it was clearly wrong too. The owner set an AI agent onto the workshop's live wages and overhead figures, labour rate, technician and apprentice time, rent, insurance, the lot, and had it work out the true cost of a billable hour with the non-billable and overhead cost sitting in its own bucket instead of spread across jobs. That is data first: the real figure comes from the workshop's own numbers, not an industry average, and the agent keeps it current as wages and overheads move.
What does an agent do with fleet accounts, warranty paperwork and bookings?
An AI agent can keep fleet accounts reconciled against each car's service history, chase warranty paperwork before it falls overdue, and manage hoist and bay bookings so an overrun doesn't collide with the next car booked in, all working from the workshop's own records rather than a generic template. Greg treats the AI like another staff member, not a magic box, which means it needs clear instructions and real context about how the workshop runs, who the fleet customers are, what counts as warranty work, how bookings get juggled when a job overruns, spelled out by the owner rather than left for it to guess. Overnight, the same kind of agent can work through a backlog of customer emails and run the workshop's own pre-written responses through a whole approval loop, so a reply is already waiting when the owner walks in rather than queued up behind diagnostics on the next car.
How does Data First Coaching approach this?
Data First Coaching looks at the workshop's own numbers first, through the Trade Business Intelligence, before any conversation about office work. For a mechanic workshop that usually starts with the real cost of a billable hour on the hoist and the gap between jobs raised and jobs invoiced, because that is often where an admin problem that looks like a time problem actually begins. Greg Allan built this after running his own trade business, and he puts it plainly: “Think of AI as well as I see these agents as another staff member.” The how, which agents to build and where they sit in a workshop's own workflow, is worked through inside the Academy, with members, because the Trade Business Intelligence is powered by every member who joins it.

