Salesforce Decode
Salesforcedecode
Back to questions
AgentforceIntermediatehuman-handoffservice-cloudmessagingcontext

Implement human handoff without losing conversational context

Real World Scenario

Customers complain that when Agentforce escalates to a live agent, they must repeat their issue. The contact center uses Service Cloud with Messaging for In-App and Web.

Expected Answer

• Persist full conversation transcript and structured summary as fields on MessagingSession or Case • Generate a handoff brief: intent, entities extracted, actions attempted, and customer sentiment • Use Omni-Channel routing to prioritize sessions with failed containment or high urgency scores • Pass customer verification status so live agents do not re-authenticate unnecessarily • Configure agent desktop with embedded summary component and links to retrieved knowledge • Define SLAs for pickup time when bot hands off—alert supervisors on breaches • Close the loop: feed human resolution notes back to improve agent topics and knowledge gaps

Follow-Up Questions & Answers

Click to expand — each follow-up includes a direct, interview-ready answer

Security first: Persist full conversation transcript and structured summary as fields on MessagingSession or Case. Enforce CRUD/FLS explicitly — `with sharing` alone is not enough. Generate a handoff brief: intent, entities extracted, actions attempted, and customer sentiment. Handoff is a product moment, not a technical afterthought. Balance speed of delivery with maintainability.

Architect Perspective

Handoff is a product moment, not a technical afterthought. Structured summaries beat raw chat dumps—agents scan intent and attempted actions in seconds. Feed resolution data back into your knowledge and topic gap analysis pipeline.