.png)
R01 is the first ILR return validated against the 2026/27 specification, and it closes in early September. It covers a month of delivery run under rules your team has had for a matter of weeks.
Six checks catch most of what goes wrong. Three are about the new rates. Two are about the migration from last year's ILR. One is about the learners who are not on the new rules at all.
The reason to work through them now rather than in the submission window is arithmetic, not diligence. An error found at R01 costs the time it takes to fix a record. The same error found at DfE audit is clawback, and on training margins of 5 to 10% clawback is not a line item.
The three rate checks
1. New rates applied to every August start, not carried over. The changes to co-investment, continued focus on robust initial assessments and formulaic price reductions where content is omitted, defunding apprenticeship standards. The failure mode here is both ignorance of the change and a system that does implement these. Check the starts, not the settings.
2. Co-investment changes. The 26/27 funding rules have introduced a range of changes to co-investment for new starts from the 1st August 2026. For employers that do not pay the apprenticeship levy the government will now fully fund apprenticeship training for apprentices aged 16-24 (Up from 16-21 in previous years), with the government paying 95% of training costs for those 25 or older and the employer paying the remaining 5%. There are also changes for levy paying employers. The government will now fund all training costs for levy paying employers with insufficient levy left in their pot for apprentices aged 16-24, but for those aged 25 or older the employer contribution jumps from the previous 5% up to 25%. And finally the completion payment will no longer be withheld from training providers where 100% of co-investment has not been recorded but this does not mean that you should stop tracking and collecting this, particularly where this can equal up to 25% of the cost of training.
3. Learning support claims documented against the £150 monthly rate. Fixed rate, and evidence expectations to match it. A claim you cannot evidence on request is an error whether anyone asks in September or in the following spring.
The two migration checks
4. Validation errors from the 2025/26 migration, cleared before submission. Migrating from the 2025/26 ILR will throw validation errors wherever new fields or new rules apply. These are predictable and they are not a sign anything is wrong with your delivery. They are, however, much cheaper to clear before the submission window than inside it, when every error competes with a deadline.
5. Which version of the rules your processes were built against. Version 3 of the 2026/27 rules was published on 29 July, two days before they took effect, and it changed the employer co-investment table. If your R01 processes were designed against version 2, the co-investment section needs re-checking specifically.
The check most providers miss ...
6. Your transition learners. The learners who started before 1 August are still completing under the previous rules, including the ones on the 16 standards defunded for new starts from 1 September.
This is the check that catches organisations who did everything else right. The rates get updated correctly for new starts, and nobody goes back to confirm the old rules still apply cleanly to the learners who should still be on them. Two rule sets then run side by side inside the same return, each correct in isolation, with no single view showing which learner belongs to which.
That is the first question an auditor asks, and answering it is a data model problem rather than a paperwork problem. If the answer involves opening more than one system, R01 is the return where that shows.
Making the checks structural rather than seasonal
Every item above is a yes or no question. The difficulty is proving the answer learner by learner, and proving it before the return goes near a submission button.
MAYTAS validates ILR data in real time as it is entered, against the rule set that applies to each individual learner, so a misconfigured rate or an unevidenced claim surfaces while it is still a correction rather than a clawback. Tribal's dedicated Funding and Compliance team tracks every DfE rule change and builds it into the product, and the 140+ maintained compliance reports mean the evidence for a return is assembled rather than reconstructed.
That is the difference worth aiming for by next year's first return: R01 as a check you run, not a fortnight you survive.
If you want to see how that check happens automatically, the brochure is a quick read.
