IMM Graduate School’s new AI-enhanced oral examination platform transcribed 2,835 spoken answers – 62.6 hours of student speech – in under 24 hours on 21 and 22 October 2026.
The platform, a first in South Africa, was introduced for the Semester 2 final assessment period, which started on 12 October 2026.
The IMM Graduate School AI-supported oral assessment application tackles a question facing higher education: when AI can produce a convincing written answer, how do institutions establish that a student understands the subject? IMM Graduate School’s solution is to bring students’ own spoken explanations into assessment, using AI to support the workflow while lecturers keep the academic decisions.
Lecturers review every response the AI flags and check up to 20% of those it approves.
“By shifting to structured oral defences, we ensure that students demonstrate real conceptual mastery and critical thinking in their own authentic voice, backed by the rigour of human academic moderation,” says Dr Cecelia Rosa, Dean of Learning, Teaching and Assessment at IMM Graduate School.
Webcam-based verification supports examination integrity, and potential irregularities are flagged for review.
Rosa adds, “AI enhances the speed and structure of our assessment pipeline, but authentic student articulation and professional academic judgement remain the true foundation of high-value higher education qualifications”.
The format mirrors the demands of marketing and supply chain roles, where professionals must defend recommendations, justify spending and weigh costs, risks and opportunities. The AI-supported oral assessment application was developed with technology partner Botlhale Village, which, like IMM Graduate School, is part of the UXi Private Education Group.
“A tool like this is only worth building if lecturers trust it, and they trust it when they can hear what the student actually said,” says Jan Goffing, CEO of Botlhale Village. “That is why the recording sits at the centre of the platform, and why we are glad to be building it with IMM Graduate School.”
Home turf
The platform runs on IMM Graduate School’s own infrastructure, with speech recognition, question writing and marking all handled on hardware the institution controls. No student data is sent to an outside AI company.
It is a standard Moodle activity, so students use their usual login and marks go straight into the gradebook. Questions can be pitched at NQF levels 5 to 8 using Bloom’s taxonomy and aimed at the module’s learning outcomes.
The platform works on phones and saves each answer as it is given, so a power failure or dropped connection does not wreck the exam. It marks understanding rather than accent or grammar.
Speak up
The AI-supported oral assessment application can draft personalised oral defence questions from a student’s quiz responses, typed answers or uploaded assignments. Students answer in a web browser, with each question shown on screen and read aloud, and can review their recording and transcript before submitting.
After submission, the system transcribes each response, suggests a preliminary score against the academic rubric and flags responses needing priority attention. On every flagged response, the lecturer listens to the original recording and makes the final decision.
If a transcript misreads a phrase because of accent or speech variation, the lecturer awards credit for what the student actually said, and not the platform’s interpretation thereof.
“Having every answer transcribed within a day means our lecturers can spend their time on what only they can do, which is judging the quality of a student’s thinking,” says Riana Prins, Dean of Undergraduate Studies at IMM Graduate School. “It is a strong start for a first run.”
Audio recordings and photo verifications are encrypted and accessible only to academic markers and moderators, and all webcam snapshots and audio recordings are permanently deleted once results are released and any appeals have been reviewed.
More to come
The oral assessment platform is one of three connected AI tools for teaching and assessment at IMM Graduate School, all working through Moodle and using the same locally hosted AI. The other two, an AI assessment and marking tool and an adaptive learning tool, are in development.
In all three, the AI does the time-consuming drafting and first marking, and people make every decision that affects a student.
The initiative combines IMM Graduate School’s longstanding academic standards with assessments relevant to marketing and supply chain careers, drawing on its partnership with Botlhale Village to adapt assessment methods as technology changes.