Reinforcement where the error happens

In a traditional course, the mistake is discovered at the final exam, when there is nothing left to do about it. The learner either passes or repeats the whole thing.

Mi Campus was built around a different assumption: the moment an error appears is the best moment to correct it, and the correction should be specific to the error, not a repeat of the entire module. This is what reinforcement means here, and it is the reason two people in the same program rarely complete the same set of capsules.

Error patterns and severity

Every activation challenge produces a result, and every result that does not meet the passing criteria carries information about why. A wrong dose because of a unit conversion mistake is not the same error as a wrong dose because the prescribed concentration was misread. A checklist failed because a step was skipped is different from a checklist failed because the steps were done out of order.

Mi Campus records these as error patterns. When the same pattern shows up across attempts or across learners, it becomes visible in the Reinforcement Hub, where your team can see which errors are occurring, how often, and in which capsules.

Each error pattern carries a severity: low, medium, high or critical. Severity is not about how many people made the mistake; it is about how much the mistake matters. In health training, confusing two similar drug names is critical even if it happened once. Formatting a date incorrectly on a record might be low. Severity lets an instructor look at a cohort and know immediately which errors need attention this week and which can wait.

The prescribed capsule

An error pattern on its own is a diagnosis. The treatment is a reinforcement capsule.

Reinforcement capsules are written by the institution in advance, one for each error worth correcting. They have the same anatomy as any capsule: a short lesson that addresses that specific gap from a different angle than the original, a challenge that tests whether the gap is closed, and the evidence that it was. They do not appear in the learner’s normal sequence. They stay out of view until a learner’s results match the pattern they were written for. Then the capsule is prescribed to that learner and appears as “needs reinforcement” in their path.

The effect is that the program adapts to the person without anyone rewriting it. A learner who never makes the unit conversion error never sees the reinforcement capsule about it. A learner who makes it sees that capsule, and only that one, before continuing. The core sequence stays the same for everyone; the reinforcement layer is different for each person.

This is also why “nobody does the same course” is a description, not a slogan. The set of capsules a learner actually completes is the core sequence plus whatever reinforcement their own errors called for.

The competency map

Passing capsules is a means, not an end. What an institution cares about is whether a person holds the competencies the program was meant to build. The competency map shows exactly that.

For each learner, every competency in the program is in one of three states: mastered, in progress or pending. A competency is mastered when the capsules that evidence it have been completed under the passing criteria. It is in progress when some evidence exists but the criteria are not yet met, or a reinforcement capsule is open. It is pending when the learner has not reached it.

The map lets an instructor answer questions that a gradebook cannot. Which competencies is this cohort weakest in? Which learner is ready to move to the next path? Which competency do most people still have in progress after the core sequence, meaning the original capsule probably needs rewriting? Alongside the KPIs of capsules completed, activations achieved, activation rate and time spent, the map is where the outcome of a program becomes legible.

What the AI does, and what it does not

Institutions can enable the Sommatic AI Command Center inside Mi Campus. It is an assistant that helps people find what they need and act inside the platform, so that a learner or an instructor spends less time navigating and more time on the work.

It is worth being precise about what it does not do. It does not grade challenges. Whether a challenge is passed is decided by the rubric and passing criteria your team defined. It does not decide which reinforcement capsule a learner needs; that follows from the error pattern the institution mapped to that capsule. And it does not replace the instructor. The instructor writes the lessons, the challenges, the rubrics and the reinforcement capsules, and reads the Reinforcement Hub to decide what to change. The AI assists inside the platform. The judgment stays with people.

If you already run a program, start with one question: what is the single error your learners make most often on the job? Write a reinforcement capsule for that error, map it to the pattern, and run the next cohort. You can do this on the free plan, and the Reinforcement Hub will show you whether the error rate moved.