Salesforce is expanding access to its Agentforce ecosystem by allowing external AI assistants to interact with Salesforce agents and business tools through Model Context Protocol, or MCP, while deepening the integration of AI agents within Slack.
The move is aimed at making Salesforce's AI capabilities accessible beyond its own applications. Through Salesforce-hosted MCP servers, organisations can expose Agentforce agents and Prompt Builder templates as tools that external AI assistants can invoke as part of a workflow. Salesforce's developer documentation lists assistants such as Slackbot, ChatGPT and Claude as examples of systems that can delegate specialised tasks to Agentforce.
MCP is an open standard designed to connect AI models with external tools, data sources and business systems. Salesforce introduced MCP support for Agentforce earlier this year, allowing companies to register third-party MCP servers and give agents access to approved tools while applying governance policies around their usage.
The expanded interoperability means companies can retain specialised agents and workflows built within Salesforce while making them available to AI assistants operating elsewhere. Agentforce agents created using Salesforce's newer Agent Script Builder can be exposed as MCP tools, allowing external assistants to call them without recreating their business logic.
Salesforce is also extending this approach through Slack. Organisations can configure Salesforce MCP servers in Slack, enabling employees to access information and take actions in Salesforce through conversations with Slackbot. Developers can additionally build custom MCP servers that connect Slackbot with tools including Apex Actions, Lightning Flows, APIs, Prompt Builder templates and Agentforce agents.
The integration is particularly relevant for sales workflows. Agentforce Sales is now available through Slackbot, where employees can use AI agents to manage pipelines, prioritise leads, progress deals and identify next steps without moving between applications. Actions performed through Slack can be synchronised back to Salesforce records.
Salesforce has been steadily expanding Agentforce as it looks to position AI agents across enterprise workflows. In August, the company began enabling the Agentforce platform by default for organisations with Agentforce access. The platform is being provisioned without an additional platform charge, although separate licensing and usage requirements can apply to individual Agentforce capabilities.
The latest expansion reflects a wider shift towards interoperability in enterprise AI, where agents built on different platforms are increasingly expected to access shared tools and work across existing applications. For Salesforce, MCP and Slack provide routes for Agentforce to operate beyond the traditional CRM interface while keeping enterprise data access and agent permissions within governed environments.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.