Leading AI at Work: Governance, Rollout and Practice / Unit 4 LAI 087 worked example, 60 to 90 seconds
Lending: Decision and Notice Cards
The idea
The lending playbook shows how cards and the matrix keep an AI underwriting model tied to a well-defined decision.
In practice
Step one: write the card that creates the initial Credit Decision, with its inputs: application data, bureau data, internal risk scores and the policy version. Step two: record every input used, assign adverse-action reasons when declining, and tie the decision to a specific policy version and model version. Step three: write the card that generates the adverse action notice from the decision, with timing and content that follow the applicable law. Step four: when an AI model is introduced, map its features to the decision's fields and make sure its outputs feed this defined Credit Decision. Step five: record in the matrix and the cards how each model change affects the decision.
Where people slip
The model version on the decision is the field teams forget. Without it, nobody can say later which version of the model shaped a declined application.
Your place on the trail
Every lesson rests on earlier ones and opens later ones. Tap any to jump.
Builds on
Something wrong in this lesson? Flag it · Sources for Leading AI at Work: Governance, Rollout and Practice