Question bank management software for universities and examination bodies

A question bank is worth no more than the metadata attached to it. Every item should carry the subject and topic it belongs to, its cognitive level on Bloom's taxonomy, its intended difficulty, the learning outcome or curriculum code it measures, and the language it was written in. Without that, retrieval is guesswork and blueprinting an exam is impossible.

With structured tags in place, assembling a paper becomes a query rather than a search: forty items from cardiology, sixty per cent at application level or above, nothing used in the last two sittings. The bank answers; the coordinator reviews.

Most institutions already have a review policy written down. What they lack is a system that enforces it. iTest routes each item through the roles your regulations require — the academic who authors it, the peer or committee that reviews it, the head of department or examinations board that approves it — and records who did what, and when.

An item reaches a live exam only once it carries approved status, so a question drafted the night before cannot slip into a paper. Rejected and revised items keep their full history, which matters when an appeal arrives months later and someone has to show exactly how a question was cleared.

The exam ends and the data begins. Item analysis turns candidate responses into a verdict on each question: how many answered correctly, whether the item separated stronger candidates from weaker ones, and whether any distractor was doing no work at all.

Feeding those results back into the bank is what stops it decaying. Items that misbehave are flagged for revision or retirement instead of being asked again next term, and the difficulty tag stops being the author's estimate and becomes an observed value.

Writing items is the bottleneck. AI Questions generates draft questions from material you supply — lecture notes, textbook chapters, clinical guidelines — across 15 question types and 18 languages, so an academic starts from a candidate item to edit rather than a blank page.

Generated items are drafts, not published questions: they enter the same author–reviewer–approver workflow as anything written by hand. The platform has generated over 10,000 questions since March 2026, and is in use at Ajman University. It also marks what it produces — objective items automatically, alongside essay marking and OMR bubble sheets.

A question written once should serve for years. Usage history per item lets coordinators exclude anything used in a recent sitting, rotate items between cohorts, and hold high-value questions in reserve — which is precisely what makes a bank an institutional asset rather than a file.

The same bank feeds delivery. iTest handles 120,000+ concurrent candidates and more than 2 million exams a year, runs Egypt's national medical licensing examination (EMLE), sat by 9,397 doctors, and integrates with Moodle, Blackboard, Canvas and Google Classroom. It deploys in the cloud, on-premise or hybrid, and is built against GDPR and FERPA.