The challenge
A first-semester finance student wants to know whether a late hand-in means a lower grade or a re-sit, and the answer depends on which semester’s regulations she is under. A lecturer in business law answers that same question forty times a term, then answers a different version of it for the second-semester cohort whose rules changed. A study adviser in the counseling office fields everything nobody else caught, for programs she does not teach.
The academy had the answers written down already. They sat in course plans, semester descriptions, assignment briefs and study regulations, one set per program per intake, revised whenever a program was revised. A single assistant trained across all of it would have been worse than none, because it would have blended two semesters’ rules into one confident answer.
What an academy tells a student about a deadline, a re-sit or a grading rule is what that student will rely on. The rule that applies is the one written for her program and her intake, and an answer drawn from the wrong semester leaves her relying on a rule that was never hers.
The solution
One assistant per course, grounded in that course’s own material and nothing else, so the scope of the answer matches the scope of the rule. Where the same subject runs under different regulations, it runs more than one assistant: first semester and second semester are one subject and two rule sets.
Department assistants sit alongside the course ones. Student counseling and continuing education each run their own, scoped to a department’s guidance rather than to a single course. Students reach the assistant for the course they are enrolled on.
All 209 live assistants were built by academic staff rather than by a central integration project. The academy kept every system it had, nothing was migrated, and course plans and study regulations went in exactly as the academy already maintains them, so there is no second version of any document to keep current.
The capability that decided it
Scoping is a first-class thing on this platform. Each course assistant is grounded in that course’s own documents and no others, and that boundary is the entire product. Platforms built around a single institutional knowledge base could not express it, and the alternative was asking teaching staff to trust a blend of rule sets.
Building hundreds of narrow assistants is a normal shape here rather than an exception to be worked around, which is what let the estate grow a course at a time. It started where the volume was, on research methodology and the high-traffic shared modules every program sends students through. When a program runs again under revised regulations, the new cohort gets its own assistant rather than an edited old one.
The results
Read the 1,076 first. Across 1,084 days the estate went quiet on eight of them, through vacations, between semesters and across three intakes. A launch spike would have decayed long before the third intake arrived. Across the life of the account the estate has taken 95,028 questions, with the busiest single assistant carrying 15,897 of them on its own, and teaching staff have made 628 revisions to assistant instructions.
Separately, in the latest quarter, 97.1% of the answers to questions the academy’s own material covers landed well. That rate was measured across 1,543 in-context questions. Questions arriving on subjects no course’s documents cover are excluded from the figure rather than counted against it, and they are kept as a list of what students are asking for that the academy has not yet written down.
Just under half the assistants ever created are no longer live. For an estate scoped per intake that is the design working: a semester ends, its assistant retires with it, and the next cohort gets one built against the regulations that now apply.
Why it worked
Scope was treated as a feature rather than a limitation. The right number of assistants was the number of places where the same question has a different right answer, and the platform let the academy express exactly that number.
Teaching staff could build unaided. The estate reached this size because the people who own the material could stand up an assistant without waiting on a central team, and the same people kept tuning them afterward.
The material never had to change shape. Source documents went in as written, which meant the assistants stayed current with the way the academy already revises its own course plans and regulations.
| Measure | To date |
| Assistants in production | 209 of 405 created |
| Questions asked, raw count across the estate | 95,028 |
| Days with at least one question | 1,076 of 1,084 |
| Revisions to assistant instructions | 628 |
| Busiest single assistant, questions | 15,897 |
| Satisfied answer rate, latest quarter | 97.1% |
| In-context questions behind that rate, latest quarter | 1,543 |
Figures from CustomGPT.ai product usage records, scoped to the account that runs the estate. Shared anonymously at the customer’s request.