If AI handles simple cases, how do you train juniors?
Track 01 · Managers and Business Leads · English edition
What you will learn to do
maintain an activity that develops a useful professional skill.
Fictional pedagogical case: a premium services company is preparing to automate its refunds.
The course
In customer service, beginners used to learn through simple refunds. They would locate the booking, check the terms and conditions, and justify a response. If an assistant handles these cases, new employees may primarily encounter exceptions.
The manager must examine what simple tasks used to teach. Their repetition did not all hold the same value. Certain steps built judgment; others consumed time without developing any particular skill.
The pedagogical objective may be to choose the applicable rule and identify missing information. The exercise must then force the beginner to make that decision. Rereading a response that has already been produced does not always reveal whether they would have found the rule on their own.
The company can compile anonymized or fictitious training cases. It prepares ordinary cases and known exceptions, then asks for an explanation of the decision before showing an automated proposal. The trainer observes the criteria used.
It is also necessary to verify transferability. A participant may succeed with a case very similar to the example without recognizing the same problem in a different presentation. A second case, therefore, changes the context while maintaining the targeted skill.
This organization constitutes a pedagogical proposal to be evaluated within the company. It does not allow one to assert that all automation degrades learning.
Your decision
Construct a five-minute exercise that verifies the ability to distinguish a general rule from an applicable exception.
Formulate an answer before opening the answer key.
Open the reasoned answer key
Present a request with a general policy and a dated exception. Ask the junior to choose the rule, justify its application, and signal any missing data. Then, show the assistant's proposal for comparison.
The success criterion is based on the justification and the identification of limitations. Mere agreement with the expected answer is not enough: the choice may be accidental.
Production to keep
skill to be retained and training activity.
Evaluation criterion
The activity requires a justified decision before showing an answer.
Sources and references
- UNESCO, Guidance for Generative AI in Education and Research, 2023
- Roediger and Karpicke, Test-enhanced learning, 2006
- McDaniel, Roediger and McDermott, Generalizing test-enhanced learning, 2007
The case and the answer key are VOCE pedagogical constructions; the references do not validate the effectiveness of this training.