How it works
Each step is a fixed price agreed before it starts, and each one leaves you with something you keep.
01
£2,000–£5,000 · about a week
We shadow the people who do the work now and measure it: volume, cycle time, what a case costs, how often it goes wrong and why. You get that written up whether or not you go further, and it belongs to you.
This is where a pilot price comes from. Nobody can quote a build honestly without knowing your exception rate, and anyone who does is guessing.
02
£10,000–£40,000 · four to eight weeks
One workflow, built and put into use, with an accuracy and throughput standard written into the contract before the work starts. If it does not hit the standard, that is our problem to fix, not a change request.
Fixed price. Not time-and-materials, not day rates.
03
£1,000–£5,000 a month · continuing, cancellable
We operate it: exceptions worked, drift watched, the agreed measures reported monthly. Most of the value of this arrives in months three to twelve, when the process quietly stops being anyone's problem.
Optional. Plenty of clients take the build and run it themselves — you own the code either way.
Oversight
The workflow is run by a fleet of codified agents — that is the delivery mechanism, and it is why one operator can take on work that would otherwise need a team. It is also the part that needs the most honesty, so here is exactly where the boundary sits.
We do not claim the automation is fully autonomous, and we would not sell it to you if it were. The value is in the volume it handles without a person, and in the exceptions being surfaced rather than silently absorbed. Data handling and governance.
Why now
Not a forecast — a deadline.
Transparency obligations and enforcement powers commence on 2 August 2026. The high-risk obligations most people are bracing for have moved: Regulation (EU) 2026/1744 pushed Annex III systems — employment, credit, essential services — back to 2 December 2027, and product-embedded systems to 2 August 2028. Most vendor sites still quote the old date.
Underwriters increasingly price unsupported platforms and unpatchable dependencies as a risk of their own. Renewal questionnaires ask about end-of-life software directly, and the answer moves the premium. If your operational record lives in a system you cannot patch, that shows up at renewal before it shows up in an incident.
The market for engineers willing to work on jQuery, Angular.js, .NET WinForms or end-of-life PHP has been shrinking for years, and the people who know your particular system tend to be closest to retirement or the exit. The risk is rarely the technology. It is that the process now depends on one person's memory.
Agent Skills and the Model Context Protocol are published, open specifications with multi-vendor implementations, rather than one vendor's preview feature. That matters for one reason: work built on an open protocol can be handed over, audited, or moved. It is the difference between a supplier and a dependency.
In most firms of 50 to 500 people there is at least one process running on a shared mailbox, a spreadsheet and somebody's judgement: supplier onboarding, credit control, claims triage, compliance checks. It works, it is nobody's job to fix, and it quietly sets the ceiling on how fast the team can grow.
A week, a fixed fee, and a written measurement of what one workflow costs you now. It is the cheapest way to find out whether any of this is worth doing.
