Skip to main content
Bring your team and maximize your impact at Dreamforce. Register three or more to unlock $999 passes.

Fix Configuration and Data Issues

Learning Objectives

After completing this unit, you’ll be able to:

  • Identify common causes of Agentforce configuration errors.
  • Resolve data issues affecting agent responses.
  • Configure Knowledge citations to appear in agent responses.

Resolve Configuration Errors

Configuration and data issues are two of the most common reasons an Agentforce agent underperforms, and both are usually fixable without touching a single action or prompt.

Symptoms include agents that won't execute configured workflows, Access Denied or Insufficient Privileges errors, or AI actions missing from the UI. Common causes are missing user or integration permissions, Flow Orchestration or Agentforce features that aren't enabled, or incorrect object-level access.

  1. Start with permissions. Confirm the agent user’s profile or permission set includes the minimum level of object permissions for each object that the agent interacts with via flows, Apex, or prompt templates. Always debug a flow by running it as the agent user to see permission issues that you possibly overlooked. For Apex, extend the security for any Apex class used by an agent to the agent user’s profile. See Configure Service Agent Access for the full range of access and controls you need to consider.
  1. Next, check Salesforce Setup to make sure the Agentforce features your workflow depends on are actually turned on.

A best practice is to put all your Agentforce and AI-related users into one centralized permission set group. It keeps access consistent and makes future audits much faster.

Troubleshoot Data Errors

Not every Agentforce problem is about prompting or routing. Sometimes the agent is set up correctly but the data feeding it, or the systems it talks to, are the actual issue.

Data errors show up as incomplete or irrelevant responses, inaccurate AI-generated insights, or automations failing because a required value is missing. The usual suspects are Data 360 sync delays or paused data streams, null or outdated field values, and inconsistent relationships between records.

  1. Start by checking the Data 360 Data Streams to confirm sources are active and syncing.
  2. Then review field mappings between your source systems and the Data 360 data model. Broken mappings can silently drop or change data.
  3. Finally, run SOQL queries or reports to spot missing, null, or stale records directly.

As a best practice, set up Data Quality Rules and alerts so these issues surface before they reach the agent.

Fix Missing Knowledge Citations

If an agent's response doesn't show a source citation, check these settings before assuming the content is unavailable.

  1. Open the Answer Questions with Knowledge action’s configuration and confirm citations are enabled.
  2. Test the action independently to confirm it returns a citation on its own.
  3. Confirm the response isn’t overwritten later in the flow.
  4. Verify the underlying Knowledge articles are published, indexed, and accessible to the agent.

You’re starting to get the hang of this. Next up are routing, action, and prompting issues.

Resources

Partagez vos commentaires sur Trailhead dans l'aide Salesforce.

Nous aimerions connaître votre expérience avec Trailhead. Vous pouvez désormais accéder au nouveau formulaire de commentaires à tout moment depuis le site d'aide Salesforce.

En savoir plus Continuer à partager vos commentaires