liveBook · Manning

Redress path

As explained in AI Governance

A redress path is a process through which an affected person can request review and reversal of an AI-driven action.

A redress path lets someone affected by an agentic action request human review and reversal. It ensures no one is trapped in an automated decision without a way to reach a person. section 6.7.1

Make the path reviewable and testable:

  • Record the agent’s goal, plan, tool calls, inputs, handoffs, approvals, final action, and errors. section 6.7.1
  • Store records somewhere tamper-resistant and outside the agent’s reach; test the redress path end to end. section 6.7.1

Human escalation

Salesforce Agentforce can escalate complex cases to a human representative through Omni-Channel, handing over the full conversation history. section 6.7.1

How AI Governance covers it

How deep it goes
Connects review and reversal to recordkeeping and end-to-end testing without specifying a complete case-handling procedure.
Examples
Uses Salesforce Agentforce escalating complex cases through Omni-Channel with conversation history.
What you can do after
Operate a workflow that routes affected-person requests to a human for review and supports reversing agent actions.
How it connects
Links redress paths with human oversight, audit logging, and graceful failure.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of AI Governance. The section links point to the parts of the book it draws on.