How we work
Four stages. Each one has to earn the next.
No retainer you cannot leave, no platform you have to adopt first, and no build priced before anyone has looked at your systems.
-
00
Diagnose
We map what you run and measure what the manual work costs.
from £7,500 fixed
-
01
Prove
We prove the fix on your own systems, not on a demo.
from £5,000 per system
-
02
Build
We build the software that removes the work for good.
scoped from the proof
-
03
Run
We keep it running, in your environment, as your contracts change.
on application
Fields populated, by brand
Darker is more complete. Every brand is a full count of its own active sites, so nothing here is sampled.
| All 1,410 active sites | Fin code | Phone | Region | Area mgr | Reg FM | Snr mnt | Mnt mgr | Hours |
|---|---|---|---|---|---|---|---|---|
| Brand Aall 58 | 58 | 56 | 57 | 58 | 58 | 57 | 57 | 54 |
| Brand Ball 44 | 44 | 41 | 43 | 44 | 44 | 43 | 43 | 39 |
| Brand Call 38 | 38 | 35 | 37 | 38 | 38 | 37 | 37 | 35 |
| Brand Dall 27 | 27 | 26 | 27 | 27 | 27 | 27 | 27 | 27 |
| Brand Eall 17 | 17 | 14 | 17 | 17 | 17 | 17 | 17 | 1 |
| Brand Fall 16 | 16 | 15 | 16 | 16 | 16 | 14 | 14 | 2 |
| Vacantall 38 | 38 | 29 | 29 | 38 | 36 | 29 | 29 | 28 |
| Distributionall 4 | 4 | 2 | 4 | 4 | 4 | 3 | 3 | 0 |
| Brand Gall 571 | 571 | 522 | 539 | 571 | 569 | 537 | 538 | 504 |
| Brand Hall 594 | 593 | 545 | 547 | 593 | 592 | 547 | 547 | 510 |
Fin code, Phone, Region, Area mgr, Reg FM, Snr mnt, Mnt mgr, Hours. Two of those columns read as populated on 1,312 sites and name an actual person on one.
Anonymised from a ProofWorks review, 2026. The counts are real. Client, brand and contractor names are removed.
00Diagnose
from £7,500
fixed price, not an estimate
We count it before anyone proposes anything.
Read only. Nothing created, edited or saved, and every request logged, so we can prove it rather than ask you to take it on trust.
You get the census: what you hold, how complete it is, where the structure is wrong, what is open, what is overdue, who is carrying it and what the manual work costs you in a year. Counted across the whole estate, not sampled.
If the answer is that you do not have a problem worth paying us to fix, that is what the document will say, and the stage ends there.
Terri working a queue and showing its reasoning on every job. Demonstration data, October 2026.
01Prove
from £5,000
per system, on yours not ours
Proof means your system, your data, your work orders.
We build against a copy of your live environment and run real volume through it. 97 of 100 test calls became dispatchable jobs on one of these.
You see the result in the system you already use, not in a slide. If it does not work, you have spent £5,000 finding that out instead of a year.
The build running: completed jobs written into the client’s system and read back to prove they landed. Demonstration data, November 2026.
02Build
scoped
from the proof, not from a guess
Priced after the evidence, not before it.
The build is scoped from what the proof actually showed, so the number is based on measured work rather than a discovery workshop.
It runs against the systems you already have. API where there is one, controlled automation where there isn't. You are not migrating onto a platform, and nothing here depends on you staying with us.
03Run
POA
price on application
We keep it running, in your environment.
Monitoring, fixes and changes as your contracts change. It runs where your systems run, on your infrastructure, under your own controls.
Cyber Essentials certified and ICO registered, which matters the first time your client's IT team asks who is touching their system.
Straight answers
The four questions we actually get asked.
Will it work with the systems we already have?
Yes, and Diagnose is where that stops being a promise. You get the answer in writing, for a fixed fee, before anyone asks you to commit to a build. We work with what you run: API where there is one, controlled automation where there isn’t.
The method does not depend on us having seen your system before. ProofMap reads one from the outside, field by field, and everything else is built against what it finds. Five read that way so far.
Nothing we do asks you to move onto something else, and nothing asks your client to either. That is the point of the whole approach, not a concession in it.
What happens when something needs a person?
It goes to a person. Anything safety critical is routed out by rule rather than by judgement, and every decision is recorded with its reasoning, so you can see what was decided and why instead of taking it on trust.
The claim is not that nobody is involved. It is that nobody is retyping.
Who can see our data?
We start read only, and we can prove it rather than assert it. Every request is logged with its method, and anything that would write is stopped before it leaves the browser. That provable restraint is why we get let into live systems.
What we build then runs inside your own environment, under your own controls. Cyber Essentials certified, ICO registered ZC167703.
Who supports it after go-live?
The person who built it. Not a ticket queue, not a first-line script, and not a different name on every reply. You deal directly with whoever wrote the code, and they already know your contract.
And none of it sits on a platform of ours. It is deployed into your environment, alongside the systems you already run, so there is nothing to migrate onto and nothing to migrate off. Your data and your configuration stay yours throughout.
Everyone sells you a deck. We prove the fix.
Next
Bring us one process that's getting in the way.
Thirty minutes. We'll tell you whether we can take it out, and roughly what it would cost. If we can't, we'll say so.