Run the Data Suitability Recipe

Take a model's input data to a certified, point-in-time evidence record, from the Recipes page to the Complete step. For what each step does and why, see Define the Use Case and Capture the Evidence.

Steps

Step 1: Navigate to the Recipes page by clicking Recipes on the left side panel.

Step 2: On the Data Suitability card, click Get Started. The record list opens.

Step 3: Click Start a new record . The recipe opens on the Model step.

Step 4: Enter the Model Name, pick a Risk Tier, add a Description if you want, and click Next. The AI Use Case is created and the side panel footer shows the record it writes to.

Step 5: On Inputs, select the Input Datastores, then tick the Input Tables & Files the model reads. Click Next.

Step 6: On Standard, tick the checks in the Governed Check Set that define the standard, or leave it empty to include every active check. Turn on Check certification expiry on a schedule if the certification will carry a validity date, then click Next.

Step 7: On Baseline, click Capture Baseline . The card confirms Baseline captured. Click Next.

Step 8: On Evidence, click Submit for Review, then share the page with the approver. When the approver has clicked Certify, the badge reads Certified. See Certify an AI Use Case.

Step 9: Click Generate Attestation . The card Evidence record #N captured appears. Click Next.

Step 10: On Complete, click Download Evidence JSON and Download HTML Report to save the record, or Open the AI Use Case to review it in place.