Salesforce Launches New Tool to Integrate Third-Party Agents With Customer Data
- Salesforce has expanded its AI agent ecosystem with the introduction of Agentforce 3, a major upgrade to its digital labor platform designed to improve the visibility, control, and...
- The update focuses on solving a primary blocker for enterprise adoption: the inability for teams to monitor agent activity and evolve those agents quickly enough to meet business...
- A central feature of the release is the Agentforce Command Center.
Salesforce has expanded its AI agent ecosystem with the introduction of Agentforce 3, a major upgrade to its digital labor platform designed to improve the visibility, control, and interoperability of AI agents within enterprise environments.
The update focuses on solving a primary blocker for enterprise adoption: the inability for teams to monitor agent activity and evolve those agents quickly enough to meet business needs.
Agentforce 3 and the Command Center
A central feature of the release is the Agentforce Command Center. This tool provides an observability solution that allows organizational leaders to manage, track, and scale how AI agent activity contributes to human productivity.
The platform is built on an updated Atlas architecture, which Salesforce states improves reasoning, performance, and trust. These technical updates are intended to provide lower latency, greater accuracy, and enhanced resiliency for enterprise readiness.
Agentforce 3 also introduces support for natively hosted large language models (LLMs), specifically mentioning Anthropic.
Interoperability and Open Standards
To facilitate seamless interoperability between different AI agents, Agentforce 3 includes built-in support for the Model Context Protocol (MCP), an open standard designed to standardize how AI models access data and tools.

The company has also expanded the AgentExchange, a marketplace where customers can access plug-and-play services. This exchange now includes over 30 partners, such as Google Cloud, AWS, IBM, Cisco, Box, Notion, PayPal, Stripe, Teradata, and WRITER.
Industry Application and Performance
Salesforce is applying these capabilities to specific sectors, including the launch of Agentforce for Communications on February 26, 2026. This industry-specific offering includes five prebuilt AI agents designed to help communications teams reclaim time.
The company reported several performance metrics from deployments since the initial October 2024 launch of Agentforce:
- 1-800Accountant autonomously resolved 70% of administrative chat engagements during critical tax weeks in 2025.
- Engine reduced its average customer case handle time by 15%.
- Grupo Globo saw a 22% increase in subscriber retention.
Other organizations utilizing the platform include PepsiCo and UChicago Medicine, benefiting from over 100 new pre-built industry actions and flexible pricing models.
Broader Ecosystem Context
The move toward agent interoperability is mirrored by other industry players. PwC has launched an AI Agent Operating System for enterprises that connects agents across different platforms and frameworks. This system integrates with various providers, including Salesforce, Microsoft Azure, Google Cloud, AWS, OpenAI, Anthropic, SAP, Workday, GitHub, and Oracle.
Salesforce has introduced other AI-related developer and governance tools, including the Agentforce Vibes coding tool and new AI data governance features.
