Agents and AI on Experience Sites

Experience CloudAgents and AI on Experience SitesPut an Agentforce Service Agent on an Experience Cloud site: messaging channel and Embedded Service deployment, guest vs verified conversations, self-service topics grounded in Knowledge, escalation through Omni-Channel, branding, security for agent actions, Data 360 personalization signals, and measuring deflection, with an end-to-end worked example.Advanced · about 90 minutesWhat you will learn

Describe how an Agentforce Service Agent reaches an Experience Cloud site through a Messaging channel and an Embedded Service deployment

Deploy and customize the chat client on Aura and LWR sites (snippet, hidden pre-chat fields, programmatic launch)

Design self-service topics and actions (case creation, order status, Knowledge Q&A) that are safe for unauthenticated and verified customers

Pass identity from a logged-in site user to the conversation and gate sensitive topics behind verification

Configure escalation from the agent to human service reps through Omni-Channel

Measure containment and deflection with MessagingSession data and site deflection signals

Contents

Architecture: how an agent reaches a site

Building the agent for self-service

Connecting the agent to the channel and the site

Guest vs authenticated conversations

Escalation to a human

Security for agents on sites

Search and personalization on sites

Measuring deflection

Worked example: Acme Support assistant, end to end

Best practices

Common pitfalls

Hands-on exercises

Key takeaways