The explanation function can provide information on the reason for the incorrect suggestion.
- Was the customer number previously assisted incorrectly?
- Is the customer number not or not correctly associated in master data?
- If the correct number cannot be assisted, it may not be present in the master data at all and must be maintained accordingly.
- In the case the wrong number has been "learned", please reach out to us, so we can possibly make corrections and re-learn incorrectly trained information, i.e. re-train the AI.
Important: To retrain the AI, we need specific information:
This should be generally applicable logic and not one-off case specific (e.g. always the same for this particular customer, not an exception to the rule).
Template to contact: email@example.com
What is suggested incorrectly?
What should be suggested instead?