The problem
Franchisees buy the same things — shipping, procurement, IT — but they don't buy them equally well. One operator spots a wrong line on a freight invoice; the one next door pays it without a second look. One renegotiates an IT contract every year; another rolls over the same overpriced agreement for a decade. The spend leaks out a few percent at a time, across hundreds of bills, in categories nobody has the hours to police line by line.
The franchise we're working with wanted to close that gap. Not by hiring a procurement team for every site, and not by handing franchisees a dashboard they'll never open. They wanted the analysis done for them — bills read, anomalies flagged, concrete savings put in front of each operator — so the floor for the whole network rises to what their sharpest sites already do.
Constraints
This is a franchise, which means scale with variation. Hundreds of independent operators, each with their own suppliers, their own contracts, their own bills in their own formats. Whatever we build has to cope with messy, inconsistent inputs and still produce something an owner-operator trusts enough to act on. The output can't be a vague score — it has to be a specific, defensible "here's the line, here's what it should be, here's what to do".
What we're building
The AI assistant we built for spend intelligence. The premise is simple to say and hard to do: help every franchisee do more, automatically, with less. The assistant looks across a franchisee's spend categories — shipping and procurement, IT, and others — and runs a fully custom bill review across all of it. Not a fixed rules engine bolted to one supplier's format, but AI workflow intelligence that reads the actual bills, understands what's on them, and reasons about whether each charge makes sense.
From that review it produces two things: an analysis of where the money is going, and concrete suggestions on what can be optimised to improve that franchisee's profitability. The point isn't to surface data — it's to surface decisions. Cancel this, renegotiate that, you're being billed for a service you stopped using, this freight surcharge doesn't match the contracted rate.
A mobile app for both Android and iOS. The assistant only matters if franchisees actually reach for it, which means it has to live on the phone they already carry, whichever phone that is. We're building native for both platforms — not one and a grudging port of the other. That's a deliberate signal of how we work: Lumitec is platform-agnostic by design, building across Mac, Windows, Android and Apple, so the right tool decides the platform rather than the platform constraining the tool.
Why it's built this way
Bill review is exactly the kind of work that looks like it should already be automated and almost never is — because real bills don't conform. The same supplier issues two formats; a "shipping" cost hides inside a procurement invoice; an IT line is described three different ways across three sites. A rigid pipeline breaks on the first exception. We're using AI workflow intelligence here precisely because the job needs judgement applied at scale, not a lookup table. The model does the reading and reasoning that a sharp finance person would do if they had time to do it for every bill, every site, every month.
Building the assistant and the app together, rather than shipping a model and worrying about delivery later, is the other deliberate call. An analysis nobody sees changes nothing. Putting it on iOS and Android from the start means no franchisee is left out for being on the wrong device — and across a franchise network, "left out" quietly becomes "the savings never landed there".
We work in a two-week cadence on the client's own stack, so each fortnight there's something real to look at rather than a slide about something that might exist later.
Where it stands
Both pieces — the spend-intelligence assistant and the native Android and iOS app — are in active delivery. The value we're building toward is straightforward: every franchisee getting the same rigorous, automatic bill review that today only the most diligent operators manage on their own, and acting on it from their phone. We're not going to put a savings figure on that here, because the honest answer is that the real numbers come from the live system once it's reviewing real bills at scale.
What's next
As the assistant reviews more bills across more categories, the analysis gets sharper and the suggestions get more specific to each franchisee's actual suppliers and contracts. That's the shape of the work from here.
If you're running a network — franchise or otherwise — where the same costs are managed brilliantly in some places and not at all in others, that gap is usually worth more than people think. The best way to find out is our free, no-obligation AI Readiness Diagnostic: we look at where AI would actually move the numbers for you, with no commitment. If it's worth building, the first paid step is a proof of concept — see how we build production AI, or meet the team behind it.
“Lumitec understood the complexity quickly and turned it into practical systems our teams can actually use.”
— Darren, Operations Director