In other lessons, we provided the basic flow of how to create AI rules, train them in the Coach Inbox, and make them live.
In this article we want to go much deeper into HOW to train them. If you manage this process well, you should be able to get to highly functional rules in about 15 learners, and then verify them and make them live within 30. But training AI is a new skill for most of us in L&D, so we want to make this concept super practical and walk through some examples of best practices.
First - let’s talk about what we actually should be reviewing.
- Core Rules
- Unexpected Alternatives
- Transcription Alternatives
- Specific Coaching
So with this in mind, we've provided some sample submissions below - these are real submissions, only the companies and learner names have been changed to protect the innocent.
- NEED DEMO AND DETAIL HERE
Our top piece of guidance for this process is DON’T just send AI moments in training back to learners. Having submissions in your inbox is a precious commodity because it gives you more to work with to vet that your rules are going to be able to handle the diversity of submissions learners will throw at it once live.