What is Data 360 for Agentforce in Agentforce architecture?
In Agentforce architecture, connect harmonized customer data to Agentforce for more useful grounded responses. This guide explains the design decisions, controls, and implementation checks needed to apply the pattern in production.
Customer context often lives across CRM, commerce, service, web, and external systems. Data 360 can harmonize that evidence so an agent receives a more coherent view instead of reasoning over disconnected identifiers and stale copies.
A unified profile increases usefulness and risk at the same time. Identity resolution, calculated insights, permissions, consent, and purpose limitation must be part of the agent design.
Use unified context deliberately
- Select only the attributes needed for the agent's current job rather than exposing the entire unified profile.
- Understand identity-resolution confidence and the consequence of joining the wrong records.
- Apply consent, classification, and access policy before the data reaches grounding or an action.
- Distinguish current transactional facts from derived scores and explainable calculated insights.
Connect data to an agent use case
- Map the job to required source fields, unified attributes, and freshness requirements.
- Test profiles with missing, conflicting, recently changed, and incorrectly matched data.
- Limit retrieval and action inputs to the minimum context required.
- Trace which data informed consequential recommendations and provide human review where appropriate.
Resources
This Tucario article is based on the following Salesforce learning and product documentation.