Showcasing Innovation: Salesforce Headless 360
Salesforce Headless 360
Salesforce Headless 360 enables organizations to expose Salesforce capabilities through APIs, MCP tools, and command-line interfaces, allowing both people and AI-driven agents to interact with business processes seamlessly across any engagement channel.
Key Innovations in Salesforce Headless 360
Comprehensive Platform Access for AI Agents
A new set of MCP tools and development capabilities empowers coding agents with direct access to Salesforce resources, enabling them to automate tasks, execute workflows, and extend platform functionality with minimal friction.
Unified Experience Across Every Channel
An advanced experience layer delivers consistent, native interactions across a wide range of touchpoints, including Slack, voice interfaces, WhatsApp, and other digital channels, ensuring a cohesive user experience regardless of where engagement occurs.
Agent Governance and Operational Control Organizations gain powerful controls to manage agent behavior throughout the entire lifecycle. These capabilities support configuration, testing, monitoring, and optimization before deployment as well as ongoing management in production environments.
A Shared Platform for Humans and Agents
Salesforce serves as the common foundation where employees and AI agents collaborate effectively. Whether users engage through Customer 360 applications, Slack, Headless 360, or other interfaces, agents can leverage existing applications, workflows, business rules, and data models without requiring organizations to rebuild them.
At the core, both human users and AI agents depend on the same trusted assets—enterprise data, business processes, and governance controls. While interaction channels and user experiences may vary, the underlying Salesforce platform remains the consistent engine that powers and orchestrates all operations.
Here is an Example of how to invoke an Agent from another source
We will consider two salesforce orgs for this example, considering one org as the master org and the other as the CRM org, we can considering more than one org as to be the CRM Org
Create two or more org considering one org as the Master org and the rest of the Orgs as the CRM orgs which would connect to the Master Org
Turn the required settings of Data Cloud in the Master org and assign the required PS to the user, then connect Another Salesforce Org from Connect an Org setting under Data Cloud setup
You should be able to visit the Data Cloud home as an application
The connected Orgs should be visible under the Salesforce CRM and should be Active once you add a Org from the Data Cloud Setup
Next, we will create a Agent force data library to add essential context to your AI Agent’s responses.
Attached is a example of the library which you can use an dwait for the batch to move the status of “Ready”.Once the status is ready please check the chunks of the library has been create, you can verify under AI model tab.
Next, we will create an Agent in the Master Org itself which will use a static Agent force Data Library resource to provide the results for the Agent.
Turn on Agent force on the Master Org from Agent force Agents
Create a Topic/Subagent with the below details for Book recommendations to be generation from the library
Please test the agent in the conversation preview.
Now we can set up the connection to connect to a third party app, for now we will connect to another salesforce org to invoke this Agent.
Under connections, navigate to Connections and Add to select a External App, you can create a External Client App from here and select it here
Once the connected app has been created with the below Oauth scope
From here we can test the Agent API from Postman using the Agent API collection from SalesforceSetup the connection parameters as below -
Next we can Test the Send Message – SynchronousBelow is the body
And we should get the below recommendation
We can invoke this connection from any Node JS application or any other Salesforce ORG.
Conclusion
Every Developer org of Salesforce with Salesforce Hosted MCP Servers at no cost, and any MCP-compatible AI agent. The above example is a brief use case of how we can connect a Agent API with another Salesforce Org and in the next article we will go though a stand alone third party app invoking the Agent through the Agent API.
Comments