Activities are not outcomes
Business schools are reviewed every few years by their accreditors, AACSB, EQUIS and AMBA, and the question that decides the review is whether the school can show that its programmes produced the learning they promise. The applied work in those programmes, the capstone, the consulting project for a company, the case competition, is where most schools expect that proof to come from. You can show a peer review team a hundred such activities and still not answer the one question they came to ask. A capstone ran. A consulting project was delivered. A competition was won. All true, all documented, and none of it is evidence that a competency was achieved, improved and measured again. Every accreditation regime for business schools asks for that second thing, and AACSB’s own numbers say it is where schools most often fall short.
What every accreditor asks for
Strip the vocabulary away and AACSB, EQUIS and AMBA converge on four objects. Stated learning outcomes for each programme. Assessment that is documented and can itself be defended. Proof, drawn from what learners produced, that the outcomes were reached. And a closed loop: an improvement made because of what the evidence showed, then measured again.
AACSB’s Standard 5 asks for assurance-of-learning processes with direct and indirect measures whose results lead to curricular and process improvements. Its definition of a direct measure is evidence from learner work, such as examinations, assignments and projects, based on direct observation of individual performance. The documentation it suggests is a table per programme with a column for each competency, its performance target, how and where and when it was assessed, the results, and the improvement made, with dates. Those are the standards’ own words.
EQUIS asks whether the school can show that the assessment regime itself is evaluated rigorously, and how it responds to gaps between intended learning outcomes and actual results, in its 2026 criteria. AMBA expects learning outcomes to be mapped and measured against assessment, and an accredited MBA to give evidence of impact for its graduates, including transfer of learning to the workplace, in its MBA criteria. Its 2025 update adds a clear policy on generative AI, with disclosure of AI use in assessed work.
Where schools stumble
AACSB publishes what its review teams find. In the 2026 State of Accreditation report, covering visits from July 2025 to June 2026, schools in initial accreditation most frequently identified Standard 5, assurance of learning, as the standard needing improvement, and had done so for five years running. Standards 3 and 5 together were cited for 54% of the schools that received a recommendation to come back before the next full review. Almost every continuing review, 96.8%, had at least one standard flagged for attention.
The societal impact findings say the same thing in different words. Reviewers most often asked schools to define specific focus areas and to “separate activities from measurable outcomes”. Some schools presented engagement as if it were impact. And the report notes that AI proficiency remains less commonly measured as a core competency, which matters now that most programmes claim to teach it.
None of this is new. The 2023 report already called assurance of learning an area of difficulty, with 48% of schools wanting progress in their processes. What has changed is that the richest piece of learner work a programme produces has become harder to trust, for reasons the first piece in this series covers.
The evidence is in a drawer
Here is the odd part. The direct measure accreditors value most, learner work at programme level that shows a competency in use, is exactly what a capstone, a thesis or an in-company project is. Every programme already produces it.
It just does not survive as evidence. The project is graded as one document, usually by one reader, against criteria that live in the examiner’s head or a rubric used once. The grade is kept. The work, the drafts, the returns and the reasoning behind the signature are not, or not in any form a second person can read later. Nothing is measured again next cycle against the same criteria. So when the review comes, someone assembles the table by hand from spreadsheets and sample papers, and the longitudinal columns, results over a cycle and the improvement that followed, are the ones that stay thin. That is the part reviewers keep asking about.
What evidence of learning must have
If you are deciding how a programme’s projects should run, these are the properties that turn the project from an activity into evidence.
- Criteria written before the work, per competency, not per course, and the same for every participant in the cohort.
- Drawn from the work itself, at the step where the competency shows, not from an exit survey about it.
- Kept per participant and per cohort, across cycles, so that an improvement can be re-measured against the same bar.
- A person’s judgement on record: who signed, what was returned and why.
- Exportable into the shape a reviewer reads, without a semester of spreadsheet work.
A programme that has these does not prepare for accreditation. It is already prepared, every day, as a by-product of running its projects.
Where Ulern stands
We build Ulern, the applied-learning platform for business schools and executive education, so read this section knowing that.
In Ulern the supervisor sets a project’s steps and what each step must contain, once for the whole programme. Every draft is checked against those criteria before a person sees it, every return and every signature is recorded, and the finished deliverable exports with that record. Reports show each participant and each cohort by step, and what the class got wrong. What is not there yet is a one-click export into an accreditor’s table; the record and the cohort reports exist, the table does not. We have also not yet been through an accreditation visit alongside a school, so the first programmes will teach us what a review team wants to see and in what order.
This is the second of seven pieces in Assign and hope, a series on what goes wrong while a programme’s project runs, and what a remedy must have. The others: Did they write it, or did the AI? · Supervision does not scale · The 90 days after the programme · What corporate clients now ask of schools · The thesis in the age of AI · What to ask an AI tool before December 2027