Turning Student Feedback into ESG / ENQA Accreditation Evidence

Quality assurance directors face a recurring problem at accreditation time: you have years of student survey data, but turning it into convincing evidence that the institution listens and acts is slow, manual, and often thin. This guide maps the ESG 2015 — Standards and Guidelines for Quality Assurance in the European Higher Education Area (Part 1, internal quality assurance) to the specific evidence student feedback should produce, and shows where AI-moderated evaluation makes that evidence stronger.

It is written for QA directors, deans of teaching and learning, institutional-research leads, and self-assessment authors preparing for ENQA-coordinated reviews, NVAO accreditation, or programme revalidation in the European Higher Education Area.

What reviewers actually look for

The ESG were adopted by EHEA ministers in 2015, prepared by ENQA with the E4 group (EUA, EURASHE, ESU) and EQAR. They are the reference framework for both internal and external quality assurance across the EHEA. Crucially, the ESG are principles-based, not prescriptive — they tell you what must be assured, not which tool to use. That means panels judge the quality of your evidence and your improvement loop, not the brand of your software.

For student feedback specifically, reviewers want to see three things:

1. Systematic collection — feedback gathered consistently, not ad hoc.
2. Genuine analysis — that you understand what students are telling you, including qualitatively.
3. Closing the loop — documented action taken in response, communicated back to students.

That third point is where most institutions are weakest, and where evidence quality most often determines outcomes.

ESG Part 1 standards mapped to feedback evidence

The table below maps the relevant ESG 2015 Part 1 standards to the concrete evidence a modern evaluation programme should produce.

ESG 2015 standard	What it expects	Evidence student feedback should provide
1.1 Policy for quality assurance	A public QA policy with stakeholders involved	Documented student-voice mechanism embedded in QA policy
1.3 Student-centred learning, teaching and assessment	Respect for and response to learners; feedback used	Evidence students are heard and that teaching adapts to them
1.4 Student admission, progression, recognition, certification	Consistent, fair processes across the lifecycle	Student experience evidence across stages, not just end-of-term
1.7 Information management	Reliable data collected, analysed, and used for management	Analysed feedback feeding programme decisions, with an audit trail
1.9 On-going monitoring and periodic review of programmes	Regular monitoring leading to continuous improvement	Longitudinal cohort evidence + documented changes over review cycles

(Standards 1.2, 1.5, 1.6, 1.8 and 1.10 also reference stakeholder input but are less directly feedback-driven.) Standard 1.9 is the heart of the matter: it requires that monitoring leads to continuous improvement, which means you must be able to show the loop from feedback to action to outcome.

The evidence gap with traditional surveys

Most institutions run Likert-scale Student Evaluation of Teaching (SET) surveys. These satisfy "we collect feedback," but they struggle against the analysis and closing-the-loop expectations:

• Numbers without meaning. A 3.8/5 score tells a panel little about why or what you did about it.
• Underused free-text. Open comments contain the actionable insight, but manual coding is slow and inconsistent, so qualitative evidence is often summarised superficially or omitted.
• No documented loop. Surveys rarely link a specific piece of feedback to a specific change, which is exactly the chain reviewers want to follow.
• Late signal. End-of-term-only collection means issues are documented after the affected cohort has already left the module.

These are well-documented constraints of the fixed-survey paradigm, not failures of any single vendor.

How AI-moderated evaluation produces stronger evidence

Koji approaches evaluation as moderated interviews at scale. Instead of a fixed questionnaire, each student has a short conversational session in which an AI moderator probes adaptively for specifics, and every response is automatically analysed into themes. This changes the evidence you can put in front of a panel.

Mapped to the standards above:

• For 1.3 (student-centred): Adaptive probing captures the student experience in depth, and built-in action tracking documents how teaching responded — direct evidence that learners are heard and answered.
• For 1.7 (information management): Automatic thematic analysis turns raw responses into structured, themed findings with representative quotes and prevalence, giving you a reliable, auditable analysis layer rather than a backlog of unread comments.
• For 1.9 (ongoing monitoring & review): Standardized, repeatable interviews across terms produce longitudinal cohort evidence, and the closing-the-loop tracker shows the change history a periodic review panel expects.
• For 1.1 / 1.4 (policy & lifecycle): Formative, mid-module collection lets you embed the student voice across the learning lifecycle, not just at the end, and document it in your QA policy.

Because moderation is standardized, the qualitative evidence is consistent across cohorts and programmes — reducing the moderator variance that weakens human-run focus groups while keeping the richness a numeric survey cannot reach. Data is handled under EU/GDPR with EU hosting, which matters for European institutions assembling an evidence base.

A practical evidence checklist

Before your next self-assessment, confirm you can produce:

• A policy statement naming your student-feedback mechanism (ESG 1.1).
• Themed qualitative findings per programme, with representative student quotes (ESG 1.3, 1.7).
• An action log linking specific feedback themes to specific changes and dates (ESG 1.9).
• Longitudinal evidence showing how themes and actions evolved across review cycles (ESG 1.9).
• A "you said, we did" communication closing the loop back to students (ESG 1.3).

If you can produce the first two but not the last three, that is the gap most likely to weaken a review — and the gap automatic theming plus action tracking is designed to close.

Honesty: where traditional tools still fit

If your accreditation evidence chain is already built around a validated numeric instrument used for sector benchmarking, you should keep collecting that quantitative trend data — panels value it, and consistency over time is itself evidence. A pragmatic model is survey numbers for benchmarking, AI-moderated interviews for the qualitative depth and the documented loop. Tools such as EvaSys remain strong for standardized and paper-based survey administration; the gap they leave is the analysed, closing-the-loop qualitative evidence that ESG 1.9 rewards. See our Koji vs EvaSys comparison for the detailed trade-offs.

Beyond Europe

The same evidence logic applies to AACSB, EQUIS, and national frameworks like the Netherlands'' NVAO: all reward demonstrated use of stakeholder feedback in a continuous-improvement cycle. The mapping differs in wording, but "collect, analyse, act, document" is universal. Koji''s evidence outputs — themed findings, action tracking, longitudinal reporting — are framework-agnostic.

The same AI interview engine also powers the main Koji platform (koji.so) for customer and user research, so institutions running both teaching evaluation and wider stakeholder research can standardise on one method.

Related Resources

• Koji vs EvaSys: A Fair Comparison for Course Evaluation (2026)
• Gender Bias in Student Evaluations of Teaching: Evidence and Mitigation
• Explore Koji for Education

Next step

If your self-assessment is approaching and the qualitative, closing-the-loop evidence is your weak point, that is exactly what AI-moderated evaluation is built to strengthen. See how Koji for Education works.