COMPLIANCE · TEF
AI Assessment and the TEF: What UK Institutions Should Track
Not background policy noise. The revised framework changes how much a persistently weak assessment and feedback score actually costs you.
By Eduface · July 2026 · 8 min read
The Teaching Excellence Framework is in the middle of a genuine overhaul right now, not a minor update, and institutions that treat AI assessment tools purely as an operational convenience are missing how directly this connects to something that affects funding-relevant reputation. This is worth tracking closely over the next year, not filing away as background policy noise.
Quick answer
Under the proposed revision, student experience and student outcomes get separate ratings. A strong outcomes profile no longer quietly offsets a weak experience score, and assessment and feedback is the theme that most consistently drags that score down.
What is actually changing
The Office for Students has been consulting on a revised TEF model that departs from the framework most institutions are used to. Rather than a fixed four-year cycle producing a single Bronze, Silver, or Gold award, the OfS has proposed moving to a more continuous, cyclical model, with institutions assessed more frequently the lower their current rating, and with separate ratings published for student experience and student outcomes rather than one blended judgement. The OfS’s stated aim is to focus regulatory scrutiny where it is actually needed rather than treating every institution identically regardless of track record.
This matters for scope as much as mechanics. The TEF has already expanded in recent years to include further education colleges and private providers, not just traditional universities, which means a meaningfully wider set of institutions than the framework’s early years now has a direct stake in how it develops.
Why assessment and feedback keeps mattering here
Assessment and feedback has been a consistently weak-scoring theme feeding into TEF-relevant metrics for years, largely through its poor performance in the National Student Survey, which our companion piece on NSS-related feedback turnaround covers in more depth. The reason this connects to AI assessment specifically is straightforward: the revised TEF’s separation of student experience from student outcomes means institutions can no longer expect a strong outcomes story, good graduate employment, good progression rates, to compensate for a persistently weak experience score. If assessment and feedback quality is the thing consistently dragging that score down, it stops being something a strong outcomes profile can quietly offset.
What is still genuinely uncertain
It is worth being honest that this is a live consultation, not a settled framework. The OfS has indicated it planned to confirm its broad approach in spring 2026 and consult further on detailed methodology later in the year, which means the specific metrics and weightings that will actually determine future ratings are not fully fixed yet. Institutions should track this as an evolving picture rather than building a strategy around any single detail that could still change before the next assessment cycle.
What is worth tracking now, regardless of how the detail settles
Turnaround time
Likely to remain a tracked signal in some form, given how persistently it has shown up as a weak point across multiple TEF and NSS cycles already.
Feedback quality, not just speed
The specificity of feedback itself is likely to matter more under a framework explicitly trying to separate genuine experience quality from outcome-driven proxies.
Consistency across a programme
Whether every student gets a comparably good experience, or quality varies sharply by module or marker, fits a granular, cyclical model looking for sustained patterns.
What this means practically for institutions using AI-assisted marking
If AI-assisted marking is helping you turn assessments around faster and more consistently, that is not just an operational efficiency story to keep internal. It is directly relevant evidence for exactly the part of TEF-relevant data collection that institutions have historically struggled to evidence well. Worth building now, ahead of the next assessment cycle: a genuine before-and-after picture of turnaround times, and, where you can measure it, whether feedback specificity and quality improved alongside speed rather than being traded off against it. A faster process that produces thinner feedback is not the win it might look like on a dashboard; the two need to be tracked together.
Frequently asked questions
Is the TEF changing significantly right now?
Yes. The Office for Students has proposed moving away from the fixed four-year cycle toward a more continuous, cyclical model with separate ratings for student experience and student outcomes, though full methodology details were still being consulted on as of mid-2026.
Does the TEF apply to private providers, or only traditional universities?
It has expanded to include further education colleges and private providers in recent years, not just traditional universities, which broadens who has a direct stake in it.
Why does assessment and feedback specifically matter for TEF, not just NSS?
Because assessment and feedback feeds into the student experience metrics that inform TEF, and a revised framework that separates experience from outcomes means a weak experience score is less likely to be offset by a strong outcomes profile than it may have been in the past.
What should institutions using AI-assisted marking start tracking now?
A genuine before-and-after comparison of turnaround times, alongside evidence that feedback quality and specificity held up or improved rather than being sacrificed for speed.
Build the turnaround evidence now
Eduface returns criterion-based feedback fast, and gives you the before-and-after data to show for it. Book a demo or start free.