Apprenticeship funding audit
How a funding audit works, what auditors sample and ask for, what triggers a review, and what happens when evidence cannot be produced.
What an audit is testing
An audit tests whether the money claimed was properly claimed. It works from your funding claim backwards: a sample of apprentices is selected, and for each one the auditor asks to see the evidence that supports every element of the claim.
It is a documentary exercise. What you remember, what you intended and what actually happened are not the point. What is on file on the day is the point.
What gets sampled
Typically a cross-section: different standards, different start dates, different employers, and anything that looks unusual in the data. High-value apprenticeships, apprentices with recorded breaks, prior learning reductions and subcontracted delivery all attract attention because that is where errors concentrate.
A failure in the sample tends to widen the sample rather than end the exercise.
What triggers a closer look
Rapid growth in starts. A high proportion of apprentices on a single standard. Off-the-job hours that cluster suspiciously near the minimum. Prices consistently at the top of the band. Subcontracting above the reportable threshold. Data that does not reconcile between your records and the individualised learner record.
What happens when evidence is missing
Funding is recovered for the affected apprentices. If the failure looks systemic rather than isolated, recovery can be extrapolated across the population, which is how a handful of missing files becomes a six-figure clawback.
The practical defence is not argument at the time. It is having collected the evidence at the point it existed.
Common questions
How much notice do you get of a funding audit?
Usually several weeks, with a request for a data return first. That period is for producing evidence you already have, not for creating it.
Can findings be appealed?
Findings can be challenged where you can produce evidence that was overlooked, or show the interpretation was wrong. Challenging on the basis that the training genuinely happened, without records, rarely succeeds.
What is extrapolation?
Applying an error rate found in a sample across the whole population. It is why a small sample failure can produce a large recovery figure.
Track this yourself
Skills Radar follows every change to the funding rules, the standards register and T-Levels, with what changed, what it changed from, and what follows. Free to read.