Salesforce saddles up MuleSoft Agent Fabric for orchestration & governance
Agents need control.
More specifically, agentic software services that bring forward new strains of AI automation and acceleration into modern IT stacks need orchestration and governance to oversee their (machine) identity access and privileges, their state of iteration (for updates and maintenance), their ability to interconnect (via protocols such as MCP, A2A as well as through Application Programming Interfaces and more) and to assess their wider status and ability to function within the modern IT stack.
As such, Salesforce has announced MuleSoft Agent Fabric to discover, orchestrate, govern and observe any AI agent, regardless of where it’s built or operating.
Agent sprawl is real
As companies accelerate their adoption of AI, they’re facing an explosion of agents across teams, platforms and vendors. While this surge holds massive potential, it also introduces a new form of fragmentation (often known as) where disconnected workflows, redundant automations and compliance blind spots emerge.
Unmanaged agents risk making it harder to govern data, enforce security and deliver consistent user experiences at scale.
“That’s why Salesforce has announced MuleSoft Agent Fabric, a new solution that transforms unmanaged AI agents into a secure and intelligent network. MuleSoft Agent Fabric provides a single place to register, orchestrate, govern and observe every agent, regardless of where it was built. Much like an air traffic controller ensures planes from different airlines take off, land and share airspace safely, MuleSoft Agent Fabric acts as the backbone that connects and coordinates an enterprise’s disparate digital workforce, turning chaos into cohesion and transforming fragmented agents into a trusted, high-performing network,” said the company, in a technical statement.
Imagine a global retailer with one agent tracking inventory, another updating prices and a third detecting fraud. MuleSoft Agent Fabric helps to ensure these agents don’t operate in silos but rather work together. So when stock runs low, pricing adjusts automatically and fraud checks happen in real time, all while governance keeps sensitive data secure.
“MuleSoft Agent Fabric delivers the missing foundation… governing, orchestrating and monitoring agents across ecosystems, so they can work together securely, intelligently and at scale. This empowers organisations to confidently accelerate their journey to becoming Agentic Enterprises, where humans and AI agents work side by side, with the clarity and control needed to facilitate security and compliance,” notes the company.
Saleforce says that MuleSoft is well-positioned to tackle this challenge.
MuleSoft Agent Fabric helps customers orchestrate and govern a diverse ecosystem of agents with four capabilities that work together to power the Agentic Enterprise:
- Discover Any Agent or Tool: MuleSoft Agent Registry is the central catalogue where every AI agent or tool – including MCP and A2A servers – can be registered and made discoverable by developers or other agents. This makes every AI asset easy to find, reuse and compose into workflows, helping enterprises avoid duplication and accelerate delivery.
- Orchestrate Agents Across Ecosystems: MuleSoft Agent Broker is an intelligent routing service that organises AI agents and tools into business-focused domains and dynamically routes tasks across them. Powered by your LLM of choice and connected through A2A and MCP, the Broker enables even the most complex, multi-step processes to execute seamlessly across diverse agents and systems.
- Enforce Trust and Security at Every Agent Interaction: MuleSoft Agent Governance provides enterprise-grade guardrails that apply security, compliance and policy controls to every agent interaction. This allows organisations to scale AI adoption safely, helping to ensure every action is consistent, secure and aligned with enterprise and regulatory requirements.
- Gain End-to-End Visibility into Agent Decisions: MuleSoft Agent Visualizer gives IT teams a dynamic map of their agent ecosystem, showing how agents connect, interact and perform. By turning black-box AI into transparent, observable systems, it provides the insights needed to optimise performance, prevent failures and build trust in the agentic workforce.
Why all this matters
AI agent adoption is projected to surge by 327% in the next two years, with magical analyst house Gartner suggesting that 40% of enterprise applications are expected to include agents within a year. Each embedded agent promises to streamline tasks within its own application, but they are blind to the broader enterprise ie. they are unable to orchestrate actions across domains and lacking the necessary guardrails for secure, cross-system collaboration.
“The reality is, most enterprises live in a multi-vendor world and that won’t change with AI. Just as companies rely on countless applications across different ecosystems, they’ll soon have AI agents from every SaaS provider and every major LLM. The strategic challenge isn’t building a single agent, but enabling all of them to work together. MuleSoft Agent Fabric gives every organisation the ability to govern and orchestrate all their agents in a cohesive and trusted way,” said Andrew Comstock, SVP and GM of MuleSoft, Salesforce.
With Salesforce’s Agentforce, users get an out of out-of-the-box platform for building and deploying autonomous AI agents. MuleSoft Agent Fabric also extends Agentforce to orchestrate with other third-party agents that don’t interact through Agentforce.

IT teams can use MuleSoft Agent Broker to structure agents and MCP servers into business-focused domains. The Broker can navigate across these domains and dynamically route tasks to the best-fit agents and tools to quickly and accurately resolve incoming prompts.

Architects can use MuleSoft Agent Visualizer to see a visual map of their agent network. Visualizer provides real-time insights into agent interactions, decision flows, and dependencies, helping teams quickly identify bottlenecks, troubleshoot issues and optimise performance across their multi-agent ecosystem.