evidoria

← Back to browse

Good practice Imported

Integrevise AI Viva Pilot at Lindenwood University — Preliminary, Mixed Evidence on Verifying Student Understanding

United States of America · Saint Charles · See the United States of America profile

Evidence: Descriptive / self-reported Top 89% 27/100 · Ask Evidence Copilot about this practice

A Fall 2025–Spring 2026 pilot at Lindenwood University paired AI-adaptive oral 'vivas' with written work to verify understanding. Of 52 Spring-2026 vivas, only 7 students gave formal feedback: 14.3% said it deepened their thinking, 57.1% disagreed — an honest, preliminary result.

52 vivas
Vivas completed (Spring 2026)
7 respondents
Survey respondents (out of 52 vivas) (Spring 2026)
14.3 %
Respondents agreeing the viva deepened their thinking (Spring 2026 survey)
57.1 %
Respondents disagreeing or strongly disagreeing (Spring 2026 survey)
20 sessions
Sessions with no feedback recorded (Spring 2026)
Integrevise AI Viva Pilot at Lindenwood University — Preliminary, Mixed Evidence on Verifying Student Understanding

Details

Maturity
Pilot
Promoter
Lindenwood University
Period
Fall 2025 – Spring 2026
Keywords
higher education, edtech, academic integrity, generative AI, assessment

Context

In late 2025, Integrevise, a startup based in Newport, Wales, began piloting an AI-supported 'viva' tool designed to check whether students truly understand the coursework they submit, rather than simply detecting whether text was AI-generated. Students pair a written assignment with a short, adaptive spoken conversation in which an AI system asks them to explain their reasoning and defend their work aloud. Lindenwood University, a private liberal arts institution in St. Charles, Missouri, ran the tool across Fall 2025 and Spring 2026 as two separate implementation cycles.

Activities

In the Spring 2026 cycle, 52 vivas were completed. Only 7 students completed a formal survey afterward: 14.3% agreed the oral exercise deepened their thinking, while 57.1% disagreed or strongly disagreed. Session-level feedback data was incomplete — 22 sessions logged positive reactions, 10 logged negative reactions, and 20 had no feedback recorded at all.

Results

The peer-reviewed pilot report, co-authored by a Lindenwood faculty member and two Integrevise staff, is explicit that the results are preliminary: it cites technical issues (voice lag, application freezes) and some student resistance to the AI-mediated format, and states the findings support further investigation rather than any claim of proven efficacy or readiness to scale. The vendor's own marketing separately claims far more favourable figures (80% reporting stronger confidence, 75% saying the approach could reduce disengagement) that do not match the peer-reviewed survey data — a gap worth flagging for any institution considering adoption.

Conclusions

This is preliminary, mixed evidence from a small, single-institution pilot with an incomplete survey response and an author conflict of interest (two of three authors are vendor staff); it supports further investigation, not a claim of proven efficacy.

Implementation

Indicative cost
Low (< €50k)
Time to results
Short (< 1 year)
Staffing & skills
Integrevise (vendor, Newport, Wales), Lindenwood University faculty (pilot lead / co-author)

Conditions for success

  • pairing a written assignment with a short adaptive spoken viva to surface comprehension and authorship confidence
  • running across two separate cycles (Fall 2025, Spring 2026) to accumulate implementation experience

Common failure modes

  • technical issues reported: voice lag and application freezes during vivas
  • some student resistance to the AI-mediated format
  • very low, incomplete survey response (7 of 52 vivas; 20 of the logged sessions had no feedback at all)
  • a majority of survey respondents (57.1%) disagreed the exercise deepened their thinking
  • undisclosed-in-abstract conflict of interest: two of three authors of the evaluating study are Integrevise staff
  • vendor marketing claims (80% stronger confidence, 75% reduced disengagement) do not match the peer-reviewed survey figures

Where it fits

Governance type
private university piloting a third-party vendor product
Scale
single institution, two small pilot cycles
Income level
high-income country (United States); vendor still pre-seed funded as of May 2026

Commonly funded by

Own resources / municipal budget

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful