Salesforce Docs Portal

About Generative AI Audit and Feedback Data

Use generative AI audit data (also known as audit trail) to track Einstein Trust Layer functionality such as data masking and toxicity detection that help ensure the safety and accuracy of the generated response.

Required Editions

Available in: Enterprise, Performance, and Unlimited Editions with an Einstein for Sales, Einstein for Platform, Einstein for Service, Einstein 1 Service, or Einstein GPT Service add-on. To purchase add-ons, contact your Salesforce account executive.

Types of Data Collected

Data 360 collects and stores two types of data for Einstein Generative AI: generative AI audit data and feedback data.

Type Source Data Collected Includes User Data
Generative AI audit data LLM Gateway
  • Prompt template ID and details
  • Request metadata and request ID
  • Org ID
  • Region and Language
  • Hydrated prompt text
  • Final PII-masked prompt text
  • Retrieved data
  • LLM response text
  • Final PII-unmasked response text
  • Response safety score
  • Response toxicity scores
  • LLM details and parameters (vendor model details and temperature)
 Yes
Feedback data Feedback API
  • Thumbs up and thumbs down reactions and reason text
  • Accept, regenerate, modify, and decline or ignore
  • Modified final response used
 No

Data Collection and Storage in Data 360

When you turn on Einstein Generative AI and Feedback Data Collection and Storage, data is stored in the default data space of your Data 360 instance. You can access this data using Data 360 reporting and through other standard platform capabilities.

You can delete your Salesforce org’s Einstein generative AI audit and feedback data at any time. Because this data is streamed to a data lake object (DLO) in Data 360, follow the steps in Remove Data Lake Objects from Data Spaces.

Data 360 refreshes the data streams once every hour.

See Also