By Greg Allan, tradesman and creator of Data First Coaching
What's the real problem AI is supposed to fix for a gasfitting business?
The real problem is not that a gasfitter needs a chatbot to answer questions. It is that every job, a callout, a leak detection, a compliance certificate, a pressure test, gets created on-site on a phone, and the details either stay in the gasfitter's head or get rewritten later from memory, by which time the reading or the fitting used has already blurred.
A gasfitter servicing a hot water system or chasing a carbon monoxide safety check through an old flue is not sitting at a desk. The job happens standing beside the appliance, the meter or the regulator, often with the next callout already waiting. Whatever does not get written down properly in that moment is gone, or it gets reconstructed later, which is unbillable time spent remembering rather than fixing.
The same gap shows up with pricing. A regulator, a bayonet fitting or a length of gas line has its own price, and that price moves. A business running on a wholesaler price list from a season or two ago is quoting against numbers that are already wrong, and nobody notices until the job comes in under margin.
Why does on-site work keep getting lost before it's ever used?
On-site work gets lost because most systems capture it twice: once scribbled or remembered on the job, and once typed up properly back at the office or in the ute. Between those two moments, details drop out, a pressure test reading rounds off, a certificate note gets simplified, and the record that compliance depends on is already weaker than the work that was actually done.
For a gasfitter, that gap is not a small thing. A compliance certificate is only as good as the reading and the safety check behind it. If what gets entered later does not match what was found at the meter or the heater, the paperwork is doing nothing more than ticking a box, which is exactly the kind of risk a licensed trade cannot afford to carry quietly.
What's the difference between AI that just answers questions and AI that can actually do work for you?
Most AI that gasfitters have seen so far is a chatbot that answers a question when asked. The AI worth paying attention to is different: it is AI agents that can go and do a piece of office work for you, pulling from your own job records and pricing to get something finished, without you sitting at a desk typing it up yourself.
An AI agent is simply software that does a job the way you would do it, using your prices, your wording and your rules, rather than a generic version built for every trade at once. Some off the shelf tools still earn their place in a gasfitting business, a scheduling app that already works well does not need replacing just because an agent exists. What your own data shows is which parts of the office work are worth running on your own records instead of a tool built for every trade at once.
Picture a gasfitter who updated a single regulator price in the system, only to find every other price on the list suddenly flagged as wrong, because the AI doing the updating was reading an old wholesaler list it believed was current. At the same time, work typed up from the phone on-site kept failing to save properly, because that session on the phone was never actually connected to the business's central records. The job notes existed, briefly, and then they didn't.
What changes that picture is keeping all of the gasfitting business's data, jobs, pricing, compliance history, in one place as the single source of truth, with every AI agent session connected straight to that record. Work done from the phone under a bench or beside a meter actually gets saved, because there is only one place for it to land.
How can AI agents turn a pressure test reading into a certificate that's ready to issue?
AI agents can take the raw notes a gasfitter taps into a phone after a pressure test or a safety check and turn them into a compliance certificate that is already close to finished, in a ready for review state, so the gasfitter only needs to glance over it and correct anything, rather than write the certificate from scratch back at the office.
That matters for a gasfitter because compliance paperwork is where mistakes are most expensive. A wrong reading or a missed detail on a certificate is not just an admin error, it is a safety and licensing issue. When the agent does the first pass using the gasfitter's own wording and format, the person checking it is reviewing work, not creating it, which leaves more of the day for the next callout instead of the paperwork pile.
How can AI agents keep your pricing current instead of quoting off an old wholesaler list?
AI agents can build pricing items straight out of a wholesaler's current catalogue, so a regulator, a bayonet fitting or a length of gas line is priced from what the wholesaler charges now, not from a list someone keyed in by hand a season or two ago and never updated.
This is a data first way of pricing a job, the quote starts from what things actually cost today, not from memory of what they used to cost. For a gasfitting business quoting hot water installs, cooktop swaps or LPG conversions, where margin gets eaten quietly by stale prices on small fittings, that is the difference between a quote that protects the job's margin and one that only looks fine until the invoice is written.
How does Data First Coaching approach this?
Data First Coaching starts from the gasfitting business's own records, the jobs, the pricing, the compliance history, read through the Trade Business Intelligence, rather than from a conversation about what the owner thinks is happening. For a gasfitter, that usually means looking first at where on-site information gets lost between the job and the paperwork, and where pricing has quietly drifted away from what the wholesaler now charges. Greg Allan puts the aim plainly: AI “it can really get stuff done autonomously and we can build automations”, rather than leaving it to memory. The how of building that for a specific gasfitting business is worked through inside the Academy, with members, because the Trade Business Intelligence is powered by every member who joins it.

