Publishing a Salesforce Agentforce Agent to Experience Cloud gives customers 24x7 conversational self-service. This guide covers the business case, the Salesforce configuration it needs, and what to watch after go-live.
Introduction
Customers want fast, accurate answers, and an Agentforce Agent on an Experience Cloud site is one of the quickest ways to give them one. Budget about 60 to 70 minutes to get the agent live.
Why publish Agentforce on Experience Cloud?
Embedding Agentforce in Experience Cloud puts AI support in front of customers inside a web experience they already know. Typical benefits:
- Reduced load on customer service teams
- Faster response times and improved CSAT
- 24×7 availability for information-based queries
- Lower cost per interaction compared to live agents
Business use case
For customer-facing companies, airlines, e-commerce and utilities among them, a lot of support volume is repetitive and information-driven: cancellation policy, seat changes, entitlement queries. An Agentforce Agent lets customers self-serve those answers in chat, which frees your agents for higher-value work.
Big quick win
If you already have a set of FAQs, policies or structured documents, the quick win is:
- Uploading content to the Agentforce Data Library
- Linking that data library to your Agentforce agent
- Publishing the agent to your Experience Cloud site
High-level steps
The full deployment touches a number of Salesforce configurations. At a high level you will:
- Create a new Service Agent and verify permission sets
- Enable Messaging and Omni-Channel routing, and create queues, presence statuses and routing flows
- Create a Messaging Channel and embed it via an Experience Cloud site
- Add trusted URLs, configure CORS and test the agent end to end
Step-by-step checklist
- Create a new Service Agent
- Verify Permission Set
- Enable Messaging
- Enable Routing Configuration
- Create Queue
- Create Presence Status
- Create Presence Status Permission Set
- Click Presence Configuration
- Create a Messaging Channel
- Create Omni-Channel Routing Flow
- Update the Messaging Channel with the Routing Flow
- Enable Digital Experiences
- Create an Embedded Service Deployment
- Add Trusted URL
- Create an Experience Cloud site
- Add CORS (Cross-Origin Resource Sharing)
- Test Agentforce Agent in Experience Cloud
Best practices
- Start with a small, high-value dataset (FAQs, policies) to train the agent.
- Monitor usage and transfer-to-agent rates so you can refine prompts and data sources.
- Keep the source documents in the Data Library clear and structured; responses get more predictable.
- Configure routing so complex or sensitive queries escalate to human agents.
Next steps & resources
For the detailed walkthrough with screenshots, see the full step-by-step guide and the Agentforce Workshop:
Conclusion
Publishing an Agentforce Agent on Experience Cloud adds self-service to a customer portal without major backend changes, and the setup cycle is short enough to make it a sensible first Generative AI project for admins and architects. Admins get a faster deployment, developers still own routing and escalation for the complex flows, and the service team sees a measurable drop in load along with faster outcomes for customers.
Leave a Comment