Agree the baseline
Record current evidence-preparation time, sample quality, reviewer roles and the definition of a relevant related event.
SAVINGFOUNDING CUSTOMER OFFERLimited to the first five customers
HOW THE PILOT WORKS
Historical evidence first.
Integration only after proof.
The flagship pilot tests one CAPA repeat-failure workflow on approved historical records over 4–6 weeks. It measures retrieval and preparation quality while every relevance, investigation and closure decision stays with named customer reviewers.
SCOPEOne priority CAPA workflowDATAApproved historical exportTIME4–6 weeksSTARTING RANGE£6,500–£9,500
FIVE CONTROLLED STAGES
From baseline
to a fixed decision.
Each stage has an owner, evidence requirement and review gate. Progress does not imply approval to integrate or scale.
Prepare approved history
The customer data owner provides the agreed export. RepeatProof maps only the fields and records inside the signed boundary.
Configure and verify
Normalisation, retrieval and calculation rules are tested against labelled examples before the controlled evaluation starts.
Run named reviews
Quality reviewers accept, mark maybe or reject surfaced history and inspect evidence, confidence and limitations.
Decide with evidence
The outcome pack compares the agreed measures and records proceed, change-scope or stop - with no automatic production commitment.
SUCCESS AGREED BEFORE SIGNATURE
Measure preparation quality.
Do not promise scrap reduction.
Scrap or recurrence reduction may take longer than a 4–6 week historical pilot. The primary measures test whether the workflow finds relevant evidence and makes review work better.
Median preparation time
Comparable manual and assisted evidence-pack preparation using the agreed start and stop points.
Retrieval precision
Accepted related events divided by displayed related events, using the named reviewer decision.
Evidence completeness
Required evidence present at first review divided by the agreed evidence checklist.
Accept, maybe, reject
User decisions on prepared outputs, including reasons and any recurring correction pattern.
WHAT THE CUSTOMER PROVIDES
Small, named inputs.
No hidden effort.
The final proposal records the expected hours, dates and decision owners. These are planning assumptions until both parties approve the pilot plan.
Three named contacts
- Quality lead and final reviewer
- Data owner for the export
- Security or privacy contact
Minimum comparison fields
- Event reference, type and dates
- Description or failure-mode context
- Product, process, machine or supplier where available
- Action links and available cost fields
Planned review points
- Kickoff and baseline agreement
- Data-quality and mapping review
- Two structured output reviews
- Final outcome and go/no-go decision
Decision boundary
RepeatProof retrieves, groups, calculates and prepares. It does not decide related-event relevance, root cause, containment, corrective action, disposition, effectiveness or CAPA closure. The customer's QMS and authorised people remain authoritative.
Check the evidence
before booking delivery.
Start with four details: work email, company, event volume and primary problem.