Every AI Decision, on the Course Itself: One Tab per Course, One Context for Your Organization

Ask an academy administrator a simple question — “what is the AI allowed to do in the Year 10 biology course?” — and watch what happens. The tutor's behaviour is set on one screen. At-risk thresholds live on a report. AI marking strictness is somewhere in the quiz tool. Whatever the course was told about its learners is in a builder field nobody remembers filling in. Answering takes four screens and a guess, and the next course is set up slightly differently.

The instructor has the opposite problem. They are told the course has an AI tutor and AI marking, but not what either will actually do with their students. Will the tutor answer from my notes or from the internet? How strict is the marking? Is anyone being alerted when a learner slips? When the honest answer is “I'm not sure”, the safest move is to switch the AI off — and an AI that is switched off helps nobody.

Both problems have the same cause: AI decisions that belong to a course were scattered across the places where the AI happens to run. The fix is to put them back on the course.

The settings hunt

Scattered settings are not only slow. They make AI hard to trust, because nobody can see the whole picture for one course at once. A strict tutor paired with lenient marking might be exactly right for a writing course and exactly wrong for a compliance one — but you can only judge the combination if you can see it.

They also make AI hard to govern. An owner who wants to know whether every course is configured sensibly ends up auditing screen by screen, and the decisions that should be the organization's — its region, its currency, how much AI each learner gets — end up repeated, inconsistently, inside individual courses.

One tab per course: AI & Tutor

For courses built with the tool inside your academy, Course Settings now has an AI & Tutor tab: the tool's own course settings, embedded where the course lives. Everything the AI does in that course is decided there, in one place:

• AI Tutor grounding and guidance style. Whether the tutor answers strictly from the course's own materials or may also draw on general knowledge, and how it guides learners.
• At-risk alerts. The thresholds that decide who is flagged, and whether the course's teachers get a Daily or Weekly email summary of who was flagged — nothing is sent when no one is.
• AI marking. How strictly the AI marks short free-text answers in the tool's quizzes: Lenient, the default, or Strict for terminology-heavy subjects.
• Learning outcomes. What students should be able to do by the end of the course, mapped to its modules.
• Course context. Who the course is for, carried into what the AI generates for it.

The question “what is the AI allowed to do in this course?” now has a single answer: open the course's AI & Tutor tab. An instructor can read it in a minute; an administrator can compare two courses side by side without a spreadsheet. And because the settings belong to the course rather than the academy, one strict course does not make every course strict.

Learning outcomes, with the evidence

Learning outcomes are new on the same tab. Instructors write what students should be able to do by the end of the course — or draft them with AI from the course's own documents. Drafts are reviewed before anything is saved: AI drafts the outcomes; the educator owns them.

Each outcome is then mapped to the modules that teach it, and the new Outcome Progress report shows the evidence — outcome by outcome, from the content mapped to it. It is evidence rather than a certified mastery badge, which is exactly what makes it useful in a curriculum review: an outcome with nothing mapped to it, or with weak results in the modules that teach it, is the conversation to have.

The outcomes also reach the learner. The AI tutor frames its explanations against the course's outcomes, so help with this week's exercise is pitched at what the course is actually trying to achieve — without changing what a strict tutor may answer.

Tell the AI who a course is for

Generic content usually comes from generic instructions. The course builder now carries more of the course's own context into everything it generates. A free-text “who is this course for” travels into every lesson plan, deck, study guide, worksheet and question set the course generates, and the course's grade level now reaches the AI on every module build.

A course for adult returners and a course for first-year apprentices can cover the same topic and still need very different examples. Saying so once, on the course, is enough. Content you already generated is not changed.

One context for the whole organization

Some context is not about a course at all. If your learners are in Singapore, every course should use local currency and local examples — and nobody should have to type that into forty courses. Settings → Organization now has Organization AI context: one optional note, up to 500 characters, that everything the AI generates for every course in your organization takes into account.

• It is combined with the context each course sets for itself, not used instead of it — the organization says where and who, the course says what and for whom.
• It reaches question banks too, which belong to the organization rather than to a course.
• A change applies to the next generation; content already generated is not changed.
• Leave it blank and nothing changes.

AI you govern, in one place

Putting AI decisions on the course answers “what does the AI do here?”. The owner's question is the other half: “how much AI runs in our academy, and on whose terms?” The answer is the same principle applied one level up — decided in one place, by you.

• One pool you control. Every AI feature — course generation, the tutor, the assistant, AI marking — draws from one credit pool your organization controls, so AI never becomes a surprise or a free-for-all.
• A meter and a report. Your admin dashboard shows the meter, and the AI usage report shows where usage went.
• Per-student focus and limits. Settings → Features lets you keep the assistant on coursework and set per-student daily limits. You decide how much AI each learner gets; the platform makes sure what you decided is what happens.
• Learners never see billing. If a balance ever runs out, a learner is simply told the AI is temporarily unavailable, while your staff get the real message.

Your plan's AI terms are part of your agreement with us, set during onboarding — and when you need more headroom, that is a conversation, not a surprise.

FAQ