Assessment integrity software for universities: what should institutions evaluate?
A university-focused guide to evaluating assessment integrity software across exams, long assignments, AI-use governance, evidence review and human decision-making.
Assessment integrity software for universities now has to support more than one assessment format. Timed exams, take-home work, long assignments and assessments that permit some use of AI create different operational and academic-integrity needs.
For secure exams, institutions should evaluate readiness checks, controlled assessment delivery, session continuity, recovery from genuine technical interruption and the evidence available to authorised staff after the attempt. A lockdown browser may be one component, but it should not be the entire evaluation.
For long assignments, the focus shifts from locking down a device to understanding how work developed. Writing-development history, saved sessions, AI-use declarations, similarity signals and reference checks can provide context when a lecturer needs to review authorship or process.
The software should also support policy differences between assessments. Some tasks may restrict AI. Others may allow bounded assistance or require an AI-use declaration. Those rules should be visible to students and manageable by authorised academic staff.
Most importantly, evidence should not become an automatic misconduct decision. Signals can be incomplete or ambiguous. Academic staff need enough context to interpret them fairly, follow institutional procedures and, where appropriate, ask the student to clarify how the work was produced.
When comparing assessment integrity software, universities should therefore assess delivery controls, assignment workflows, AI governance, evidence quality, access controls, retention, student notice, review processes and operational resilience together.