University of Reading Prepared for University of Reading Integrevise

Integrevise for University of Reading

Move beyond polished output to evidence of what each student understands.

Reading's blind test showed that markers did not identify 94% of 33 submitted AI-generated answers as AI. Integrevise adds a short adaptive oral discussion after submission so students can explain, justify and defend their work, producing structured evidence for staff and personalised feedback for the student.

Explore Integrevise See the value

Value at every level

Move from polished output to evidence of learning.

Integrevise adds a practical evidence layer while preserving Reading's positive approach to AI, academic judgement and the assessments faculty already use.

For Reading

Evidence beyond detector scores.

Support Reading's positive educational approach to AI with evidence of understanding rather than a claim about how the text was produced.

For Faculty

Questions grounded in each submission.

Use the existing brief, rubric and student work to create a structured explanation record.

For Student

Make reasoning visible.

Give students a fair chance to show their own understanding and receive feedback on where reasoning needs work.

Peer reviewed2026University setting

Evidence before adoption

A practical approach, tested in higher education.

Integrevise's adaptive oral approach has been evaluated in a university setting and published in the peer-reviewed journal Trends in Higher Education.

Read the peer-reviewed study

Directly aligned with Reading

The principles are in place. Integrevise makes them operational.

Reading combines a three-category approach to AI in assessment with direct research showing that polished output is weak evidence of individual understanding: markers did not identify 94% of 33 AI-generated answers in its blind test. Integrevise makes the response constructive by creating direct evidence of reasoning while instructors retain every academic decision.

A three-category assessment framework A real-world blind test Making GenAI use visible
Reading prioritisesTransparent AI use, visible reasoning and human judgement
Integrevise addsEvidence of individual understanding

A simple next step

Compare the evidence Reading would require.

A short conversation is enough to compare priorities, explore fit and decide whether the approach deserves a closer look.

Book a 20-minute conversation No preparation, module selection or assessment redesign needed.