AI Sandbox: Test Smart Data Capture To Improve Accuracy

 Product Area: Smart Data Capture

 Impact of Change: New Functionality

 Roles Impacted: Legal Operations teams managing Smart Data Capture

 Description: We've introduced AI Sandbox, a testing environment that lets you validate Smart Data Capture accuracy before running it on your live contracts.

 Problem Addressed:

Legal teams were frustrated with SDC accuracy issues that only surfaced after processing their entire contracts. You'd write metadata descriptions, run SDC on hundreds of contracts, and discover results were not acceptable - requiring manual cleanup. Even worse, you didn't know which descriptions were problematic or how to fix them. There was no way to test descriptions beforehand or get guidance on improvements, creating anxiety about deploying SDC.

 Solution Overview:

AI Sandbox provides a risk-free testing environment where you can perfect your Smart Data Capture setup before going live.

 What's New:

Test on Sample Contracts: Select representative contracts from your repository and run extractions without affecting live data

See AI Reasoning: View the exact contract sections and logic the AI used for each extraction decision

Get Improvement Suggestions: Mark results as correct/incorrect and receive AI-generated recommendations to refine your descriptions

Track Accuracy: Monitor success rates and iterate until you achieve target accuracy (75%+ recommended)

 How to Access:

On legacy metadata workspaces: Navigate to Manage on the navigation panel > Contract Metadata > Click on the Contract Type > Click the three dots next to any metadata field, and select "Test Smart Data Capture".

On global metadata workspaces: Navigate to Manage > Click on Metadata Manager > Click the three dots next to any metadata field name, and select "Test Smart Data Capture".

This will be available to all customers with Smart Data Capture enabled.