Scenarios
Six problems we have already built for
Different trusts describe their pain differently, but it usually lands in one of six shapes. Each has a defined pilot, a fixed price and a known duration — because we have scoped the work step by step rather than estimating it.
At a glance
Every pilot fits inside a direct-award window
All six are priced under the threshold at which most NHS organisations can award directly, and all six complete inside four to eight weeks.
| Scenario | Pilot | Duration | Readiness |
|---|---|---|---|
| Closing consultancy gaps The review happened. Nothing changed. | £12,000 | 4–6 weeks | Ready today |
| Paper to structured data The OCR mountain | £25,000 | 6–8 weeks | Ready today acute |
| Manual process replacement A person retyping between two screens | £18,000 | 5–6 weeks | Ready today |
| Regulator evidence Reconstructed the week before an inspection | £20,000 | 5–7 weeks | Ready today |
| EPR migration data quality Go-live happened. The data did not all arrive. | £15,000 | 4 weeks | Ready today |
| Estate map and silent failure Nobody knows what stopped working | £15,000 | 4–6 weeks | Ready today |
Prices differ because the work differs. A read-only reconciliation is a lighter build than clinical-event extraction from scanned pages, and we price it accordingly rather than averaging everything into one number.
Scenario one
Closing consultancy gaps
What it looks like
A review was commissioned. The findings were accepted. The recommendations went into a programme plan competing with everything else for the same handful of people, and eighteen months later the same problem is on the board report.
The advice was not wrong. There was simply nobody free to act on it, every day, forever.
What we do
Take one accepted recommendation and make it operational software. The reconciliation that only happened when somebody remembered now happens daily, with a log.
Ready today
£12,000 4–6 weeks. Read-only — no writes to any system.- Lightest possible governance — nothing is changed, only compared
- Fastest of the six pilots for that reason
- Scales to a department at £45,000, trust-wide beyond that
Scenario two
Paper to structured data — the OCR mountain
Scanning is not the problem. Most trusts have already paid somebody to scan. The problem is that scanning produces pages, and a clinician needs facts.
The historic archive
Decades of case notes in storage or half-scanned into a document system. Ten to fifty million pages at a medium trust. Digitised, but not readable as clinical data.
The daily inflow
Paper referrals, dictation, consent forms, handwritten charts, faxed results. Five to fifteen thousand documents a month, arriving whether or not the backlog is cleared.
The data prison
Clinical information locked in legacy systems with no interface. Readable on screen, not retrievable by any other means. We read the screen.
Where the gap actually is
A document management supplier delivers pages digitised and indexed. What does not reliably happen is extraction — turning those pages into the structured clinical records an electronic patient record can actually use.
So a clinician opening a returning patient sees only what was captured under the current workflow. Everything before that is a scanned image nobody can search.
That gap is the work. We sit alongside your scanning supplier rather than replacing them.
Ready today acute trusts
£25,000 6–8 weeks. A 500-document sample, drawn from content you already hold.- No new scanning needed at pilot stage
- Pilot does not write to production systems
- Seven acceptance criteria, agreed before we start
Mental health trusts are not ready yet. Safeguarding-signal extraction is in development, not built. If you are a mental health trust we will tell you that in the first conversation rather than the third.
Scenario three
Manual process replacement
What it looks like
A member of staff reads something on one screen and types it into another. Every day, hundreds of times. Referrals transcribed from PDF into the referral system. Results copied between systems that do not talk. Letters assembled by hand.
Where this bites hardest is waiting times: a referral that takes two to five days to reach a consultant list spent most of that time in a transcription queue, not in clinical triage.
Why we usually start here
Of the six scenarios this is the only one where every single step is already built. Nothing is a small build, nothing is roadmap. It is the lowest-risk way to find out whether we are any good.
Every step ready today
£18,000 5–6 weeks from engagement letter to user acceptance.- One workflow, end to end
- The original document stays attached for clinical reference
- Additional workflows take one to two weeks each afterwards
Scenario four
Regulator evidence
What it looks like
Evidence for an inspection or a follow-up gets assembled retrospectively — a team pulling records together in the weeks before a deadline, reconstructing what happened from whatever was logged at the time.
It is expensive, it is stressful, and it proves what could be found rather than what actually happened.
What we do
The operational work produces the evidence as a by-product. Because the bot is doing the task, it logs every instance as it happens — so the evidence pack is a report rather than a project.
This one pairs well with a remediation programme: the same engagement satisfies the operational need and the assurance need.
Ready today
£20,000 5–7 weeks. One evidence domain.- Continuous rather than periodic
- Each pack format is configured to the regulator it serves
- Departmental scope adds further domains at £65,000
Scenario five
EPR migration data quality
What it looks like
Go-live happened. The programme is winding down. And then the queries start — historical context missing at the point of care, patient information split across the old systems and the new one, information governance requests that cannot be answered from the new record alone.
This is normal. Migrations of any size generate six to eighteen months of data quality remediation afterwards, and it is almost never in the vendor's scope.
What we do
Compare the legacy state against the new record field by field, classify each gap — missing, partial, conflicted, stale — and write structured remediation back. Every write is read back and verified before it counts.
We work alongside your EPR supplier, not against them. The gap we close is one their contract does not cover.
Ready today
£15,000 4 weeks. Reconciliation and reporting only — no write-back at pilot.- Sample cohort stratified across your main specialties
- Write-back begins at departmental scope, one specialty at a time
- Nothing is written to clinical decision data. Ever.
Scenario six
Estate map and silent failure detection
What it looks like
Nobody has a current picture of which systems talk to which, what flows between them, or who owns each connection. So when an interface quietly stops, it is found by a clinician noticing something missing — often weeks later.
A queue backing up, a scheduled job that stopped running, an integration returning success codes for messages nobody is reading. None of it raises an alarm, because nothing is watching for absence.
What we do
Build the map from your existing records and live discovery, then watch every connection for liveness, latency and failure rate. We establish what normal looks like, and alert when it stops.
An instrument that has gone quiet is a fault, not an all-clear.
Ready today
£15,000 4–6 weeks. One division or clinical area.- Usually better as a second engagement than a first
- Makes every other scenario easier to deliver afterwards
- Full-division scope at £55,000
If more than one fits
The second engagement is always faster than the first
By then the relationship exists, the governance route is known, and the technical interfaces are already in place. Trusts that run two scenarios pay less for the second than they would for a standalone engagement, and it lands sooner.
We would still rather do one properly first.
Not sure which one you have?
Describe the problem in your own words on a thirty-minute call. If it maps to one of these we will tell you which, what it costs and how long it takes. If it does not, we will say that instead.