Decathlon: knowing a tent starts when you have to put it up

A tent in a catalogue is usually standing perfectly. In real life it has poles, wind and someone who would already like dinner. Sports product training is therefore a useful common-sense test. Can we explain equipment in the conditions it was bought for, or are we simply delivering an elegant recital of the product page?

Decathlon publishes sports advice material. That is a real context for discussing learning. What follows is my independent proposal for AI-supported consultant practice, not an account of Decathlon's internal system. I would connect digital rehearsal with mandatory hands-on work; otherwise we produce a remarkably informed expert on packaging.

Teach through a situation

For example, an employee receives a customer task: a first short camping trip, exercising at home or travelling with limited luggage. They ask useful questions and select an appropriate product class. AI plays the customer and changes conditions. The objective is to explain needs, specifications and compromises, not guess a single approved item.

The conversation then needs to become a practical task based on verified materials: demonstrate adjustment, assembly or care. A coach observes performance and the system records the confirmed skill. Technical limits come from current documentation for the exact model. Similar names do not justify transferring specifications between products, however persuasively a model can discuss them.

Recurring customer questions and return reasons can supply fresh exercises once personal data is removed. Frequent confusion about sizing or adjustment should change a learning scenario. Training then stays connected to the shop floor. There is no need to invent another motivational theme when existing difficulties are already recorded quite clearly in customer contacts.

Can the employee explain it with their hands?

The three measures would be suitability of the chosen product for the brief, correct independent demonstrations and returns linked to inaccurate usage advice. The last needs a meaningful observation period. Looking only at a quiz immediately after training makes it easy to confuse short-term memory with a skill that will survive until the next season.

A small team working on one category, such as straightforward camping equipment, could test the approach. Before and after preparation, staff complete an unfamiliar task using an actual product. A comparable group uses ordinary learning materials. Coaches assess against agreed criteria rather than their personal impression of how confidently someone delivers the explanation.

AI must not diagnose conditions, prescribe exercise or promise safety in extreme activities. Questions beyond a sales associate's competence require an appropriate specialist. Practice records should not become public rankings. The result I want is an employee who explains possibilities calmly and acknowledges limits honestly. A tent needs to stay upright; it does not need to validate an algorithm's confidence.