Salesforce announced the launch of Agentforce 3, a major upgrade from its artificial intelligence (AI) platform, which provides companies with full visibility and control to scale AI agents independently.

Agentforce 3 is backed by a new Command Center for full observability, open standard support of the Context Protocol (MCP) Model for plug-and-play interoperability, and more than 100 ready-to-use industrial actions to accelerate business value.

With this update, Agentforce is expected to help companies expand successful things, improve those that are not yet optimal, and unlock the full potential of AI Agentics with clarity, control, and speed.

In recent months, we have listened to customer input and continued to innovate. The result is a large Agentforce 3 which brings higher intelligence, performance, and trust to every implementation," said Adam Evans, EVP & GM Salesforce AI.

He believes that Agentforce 3 will change the way humans and AI work together, significantly increase productivity, efficiency, and business transformation.

Agentforce Command Center

Command Center provides dashboards that allow teams to analyze any AI agent interactions, dig up details, understand usage trends, and view AI-based recommendations to improve Agentforce in a sustainable manner.

MCP Connectivity

Open standards such as the Context Protocol (MCP) Model not only open new opportunities in interoperability, but also present challenges in terms of governance, identity, and control.

Agentforce now supports the use of the Claude Sonnet model from Anthropic which is hosted through Amazon Bedrock. Thus, Agentforce is now more responsive to 50% lower latency since January 2025.


The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)