The 80-page problem

The real bottleneck in accreditation assessment

Accreditation is infrastructure. That’s not a metaphor – it’s the World Bank’s conclusion.

Its flagship World Development Report, published in December 2025, is the first comprehensive global analysis of what quality standards really do for economies. It names accreditation as one of four pillars of quality infrastructure globally, alongside metrology, conformity assessment, and standardisation. Countries with stronger accreditation systems grow faster and participate more effectively in international trade and talent mobility. ¹

UNESCO’s 2026 roadmap for transforming higher education reaches the same conclusion. With over 22,000 accredited institutions worldwide, it calls for systems that are “more relevant and impactful.” At that scale, accreditation cannot function as a bespoke, expert-intensive process applied institution by institution. It must work like infrastructure. ²

Most people working inside accreditation bodies already sense this. The question is whether there’s a significant gap between what accreditation is being asked to do and how it currently gets done.

The gap

The World Bank states it can cost up to $425,000 USD per institution in low- and middle-income countries to complete a full accreditation procedure. At that price, the reach of international accreditation is structurally capped –The real bottleneck in accreditation assessment

Accreditation is infrastructure. That’s not a metaphor – it’s the World Bank’s conclusion.

Its flagship World Development Report, published in December 2025, is the first comprehensive global analysis of what quality standards really do for economies. It names accreditation as one of four pillars of quality infrastructure globally, alongside metrology, conformity assessment, and standardisation. Countries with stronger accreditation systems grow faster and participate more effectively in international trade and talent mobility. ¹

UNESCO’s 2026 roadmap for transforming higher education reaches the same conclusion. With over 22,000 accredited institutions worldwide, it calls for systems that are “more relevant and impactful.” At that scale, accreditation cannot function as a bespoke, expert-intensive process applied institution by institution. It must work like infrastructure. ²

Most people working inside accreditation bodies already sense this. The question is whether there’s a significant gap between what accreditation is being asked to do and how it currently gets done.

The gap

The World Bank states it can cost up to $425,000 USD per institution in low- and middle-income countries to complete a full accreditation procedure. At that price, the reach of international accreditation is structurally capped – not by lack of demand, but by the economics of how assessment works. ³

The bottleneck isn’t the judgment. It’s everything surrounding it.

Before a site visit, every panel member reads the self-evaluation report individually. Eighty pages. Sometimes a hundred and twenty. Appendices, curriculum maps, staff CVs. Each person builds their own picture of the institution, with their own annotations, in their own document. There’s no shared layer. No structure connecting what one assessor flagged to what another is seeing three chapters later.

Then comes the site visit – two or three days of interviews and observation that produce the most important evaluative data in the whole process. Mostly captured in personal notes that were never designed to feed into anything systematic.

Good panels produce inconsistent reports not because the people aren’t capable, but because the infrastructure around the judgment was never built to support what the judgment requires.

What AI actually changes

The claims you hear about AI in accreditation are mostly wrong in one direction: “AI will read faster, so we need fewer panel members.” That’s not the interesting proposition.

The judgment that drives accreditation outcomes – does this programme meet the standard, does this institution have the governance capacity to sustain quality – cannot be automated. The moment you remove the human from that decision, you’ve changed what accreditation is. Nobody running a serious accreditation body wants that.

ENQA, the European body that regulates accreditation agencies across the Bologna Process countries, published its guidelines on responsible AI use in November 2025, the outcome of consultation with its 60 member agencies. The central principle is unambiguous: AI is intended to support and not replace human judgment. Final decisions remain the responsibility of human experts, in line with the peer review principle that underpins the European Standards and Guidelines. ⁴

That’s not a caution – it’s a description of where value lies.

A more useful question is: what is the panel spending time on that isn’t judgment?

Quite a lot. Navigating documents without a shared structure. Synthesising observations without a common scaffold. Manually cross-referencing standard numbers against the specific framework version in front of them. Drafting a written artefact largely from memory when the evaluative work has already been done, verbally, in the room.

This is where AI-assisted tools are genuinely useful – not replacing the panel but removing the manual work that slows them down and obscures the evidence. A system that maps an institution’s evidence against the relevant standards before the panel arrives. A site visit agenda scaffolded from the institution’s own documentation, with evidence linked to criteria. The panel still reads, still visits, still judges, still signs off. What changes is the administrative load sitting around the judgment, not inside it.

The consistency question

If you sit on accreditation panels, or manage them: is inter-panel consistency as reliable as the process assumes?

Two programmes, same cycle, same standards, panels drawn from the same pool. In principle the resulting reports should be comparable in rigour and language. In practice they often diverge – not because assessors are inconsistent, but because the process gives each panel a different starting point and different tools to work with.

This matters more than it used to. UNESCO’s call for accreditation to be “more relevant and impactful” carries an implicit accountability expectation. Institutions appeal decisions. Bodies are scrutinised. The written record of how a decision was reached is now a central part of the accountability framework. ²

Infrastructure that supports consistent, evidence-linked assessment doesn’t just make panels more efficient. It makes decisions more defensible and makes scale possible.

What it looks like in practice

At Ripltec we tested our approach against a prior programme review – working from the same self-evaluation documentation provided to the assessment team. Our report was compared against the final expert output produced after the site visit and supervised drafting.

According to the client, the evidence linkage in our version was stronger and more granular. Nothing about the panel’s judgment was replicated or replaced. What changed was the infrastructure around it.

Where this is going

The World Bank’s diagnosis is clear: accreditation needs to scale in reach, in consistency, and in its capacity to serve institutions where the economics of current methods simply don’t work. That’s as much a technology problem as a policy problem.

The bodies that work this out first will have a meaningful advantage in the markets the World Development Report is pointing to.

Notes

1. World Bank, World Development Report 2025: Standards for Development (December 2025). https://www.worldbank.org/en/publication/wdr2025

2. UNESCO, Transforming Higher Education: Global Collaboration on Visioning and Action (March 2026). https://www.unesco.org/en/articles/transforming-higher-education-global-roadmap-future

3. WDR 2025, op. cit.

4. ENQA, Guidelines for the Responsible Use of AI in External QA (November 2025). https://www.enqa.eu/wp-content/uploads/ENQA-AI-guidelines-final-2025.pdfill have a meaningful advantage in the markets the World Development Report is pointing to.