AI won’t fix a broken process

Every vendor in field service software is selling you an AI layer this year. Agents that write your job notes, assistants that answer questions about your data, tools that brief your techs before they arrive.

Some of it is genuinely useful, and we’re building in that direction too. But after 30 years of watching Australian service businesses buy software, we’d say this: none of it helps if your techs still can’t see today’s run and your invoices still go out nine days late.

We’ve been through a few of these waves. Handhelds were going to fix field service. Then the cloud was. Then mobile apps. Each one genuinely helped the businesses that had their process sorted, and each one disappointed the businesses hoping the technology would supply the discipline they didn’t have. There’s no reason to think this wave is different.

What has changed is the pace of the claims. A feature that would once have been sold as a time-saver is now sold as a member of your team. Read the fine print on most of it and you find something genuinely handy — automatic job summaries, better search, a nicer way to ask a question of your own data — wrapped in language that implies it will run the business for you.

Automation is a multiplier, and multipliers work both ways

Point a clever model at a job record nobody filled in properly and you get a confident summary of nothing. Automate a quoting process that’s wrong and you’ll produce wrong quotes faster than you ever could by hand.

The businesses getting real value out of automation are the ones whose data was already clean — because their process was already sound. That’s the uncomfortable bit. The prerequisite isn’t a licence. It’s a workflow your team actually follows.

It’s the same reason a report is only as good as the data behind it. If half your jobs are missing their parts, no amount of clever summarising will tell you your real margin. It will tell you the wrong number faster, and with more confidence than a spreadsheet ever had.

The unglamorous things that actually move the numbers

In our experience there are four, and none of them will ever be a launch announcement.

One thread from quote to invoice, so nothing is re-keyed. A schedule everyone can see. Job costing that includes the parts, so margin is a fact rather than a feeling. And a preventative maintenance program that runs itself, because that’s the revenue you can forecast twelve months out.

Get those four right and the business feels different within a quarter.

If you only do one, do the first. Almost every problem an owner brings us — margin they can’t explain, invoices going out late, techs ringing the office all day — traces back to the job record being re-typed somewhere between the site and the ledger.

The second is the schedule everyone can see, because it’s the fastest to fix and the whole office feels it inside a week. The third is job costing that includes parts, which is where most businesses discover their real margin is several points off what they assumed. The fourth — preventative maintenance running itself — is the one that changes how the business is valued, because forecastable revenue is worth more than reactive revenue to a buyer, a bank, or you.

Where automation genuinely earns its place

We’re not luddites about it. Automation is excellent at the repetitive, rule-based work: raising the PM jobs when they’re due, pulling the parts onto the invoice, flagging the job that’s been sitting at “completed, not invoiced” for a fortnight.

That’s the boring bit, and the boring bit is where the money is. It’s also safe to automate, because the process behind it is deterministic — you’re not asking a model to guess what happened on site.

The distinction we hold to is simple: automate the things where being wrong is obvious and cheap, and keep a human on the things where being wrong is expensive and invisible. Raising a scheduled PM job is the first kind. Deciding what to charge for a contested three hours is the second.

That’s also the honest reason we don’t lead our marketing with AI. We’d rather show you a scheduling board your dispatcher can run on day one than a demo of something clever that depends on data you don’t have yet. When the data is there, the clever parts get much more useful — and we’ll build them.

Ask any vendor these three questions

Including us. What does week one look like, with dates? Who answers the phone when it breaks, and where are they? And what won’t your software do for a business like mine?

That third one is the tell. Everyone can demo the good bits. Thirty years in, we know where we fit — HVAC&R and electrical contractors who need configuration around their own process — and we know where we don’t. We’d rather say so before you sign than six months into an implementation.

Australian-founded, and it matters more than it sounds

We don’t raise the flag for the sake of it. It matters for practical reasons: our support hours are your working hours, we know what a Reece or Actrol account looks like, we’ve built for Australian compliance and after-hours rates, and when something breaks at 7am you speak to someone who can change it — not someone logging a ticket for a team who start work as you’re knocking off.

Thirty years in the same market also means we’ve watched businesses grow from eight techs to forty. We built for that transition specifically, which is why we don’t chase the sole-trader end of the market and don’t pretend to.

If you’re a one-van operation, there are simpler and cheaper tools than us and we’ll happily name them. If you’re running 10 to 50 techs across HVAC&R or electrical, with maintenance contracts, project work and real asset obligations, that’s the shape of business we’ve spent three decades building for.

Ask us the hard question. Call 0457 228 884 or book a demo — Australian-founded, Australian-led, and you’ll speak to someone who knows the industry, not a script.