Analyst(s): Mitch Ashley
Publication Date: April 17, 2026
Salesforce is explicitly positioning Agent Fabric as a multi-vendor agent control plane, not a Salesforce-centric workflow extension. This is a platform positioning move: the company asserts it can anchor governance, orchestration, and observability across heterogeneous enterprise AI workflows. Guided determinism, Salesforce’s model of fixed handoff rules with bounded LLM reasoning, signals its architectural answer to the autonomy-versus-governance tension.
The multi-vendor claim is real at the discovery layer: Agent Fabric can scan and inventory agents from Bedrock, Foundry, and GoDaddy. Enforcement parity with Salesforce’s own Agentforce agents is unproven and remains the critical gap. Trusted Agent Identity, introducing mobile approval for high-risk non-human agent actions, is a meaningful early signal on the non-human identity governance problem, but not proof of maturity. MCP Bridge is positioned as an enterprise-grade on-ramp for MCP protocol onboarding. Fragmented availability persists: Agent Broker, the deterministic orchestration core, is in beta until June 2026. The strategic stakes: can Salesforce deliver operational enforcement and governance parity for third-party agents, or will the control plane remain Salesforce-centric in practice?
What is Covered in This Article:
- Salesforce Agent Fabric expansion: guided determinism, Trusted Agent Identity, MCP Bridge, and Agent Broker beta availability
- Guided determinism as Salesforce’s architectural answer to autonomy-versus-governance in agent orchestration
- Scrutiny of the multi-vendor control plane claim—discovery versus enforcement parity and what it takes for governance to be real
- Trusted Agent Identity as a directional signal for non-human identity governance, not yet evidence of maturity
The News: Salesforce advanced Agent Fabric with expanded MCP discovery, Trusted Agent Identity, and MCP Bridge for enterprise-grade protocol onboarding [1]. Agent Broker, the deterministic orchestration core, is in beta now and targets GA in June 2026. Agent Fabric can discover and inventory agents from Bedrock, Foundry, and GoDaddy, but enforcement parity with Agentforce is to be seen. Trusted Agent Identity introduces mobile approval for non-human agent actions, addressing a critical gap in agent identity governance. Availability remains fragmented, with full multi-vendor enforcement not yet demonstrated.
Salesforce Stakes Out Multi-Vendor Agent Control Plane—Determinism, Governance, Enforcement Remains the Test
Analyst Take: Salesforce’s Agent Fabric expansion is a platform positioning move, staking a claim as the multi-vendor agent control plane. The company aims to anchor governance, orchestration, and observability across enterprise AI workflows. The strategic tension is clear: can Salesforce deliver real enforcement and governance parity for non-Salesforce agents, or will its control plane remain essentially Salesforce focused?
Guided Determinism: Salesforce’s Operational Answer to the Autonomy-Governance Tension
Guided determinism, fixed handoff rules with bounded LLM reasoning, is Salesforce’s architectural answer to the autonomy-versus-governance challenge. This approach constrains agent behavior at runtime, moving governance from policy to enforcement. For engineering leaders, guided determinism offers a path to scale agentic automation without ceding control. The risk: if handoff rules are too rigid, agent utility and adaptability drop. If too loose, governance weakens. Salesforce’s approach is credible as a governance model, but operational reliability will depend on enforcement and stability across heterogeneous agent ecosystems. This is the core of the agent-governance-and-control challenge.
Agent Script, while less visible in the announcement, is the execution-level mechanism that operationalizes significant portions of this guided determinism. It codifies handoff rules based on Salesforce’s internal experience with agent development. This distinction between policy and enforcement is technically consequential, and its impact will be measured by how well it fills operational gaps as trust in AI grows and adoption broadens.
Multi-Vendor Control Plane: Easily Said, Hard to Achieve
Salesforce positions Agent Fabric as a truly multi-vendor control plane, claiming it can discover and govern agents from Bedrock, Foundry, and GoDaddy. Today, only Agentforce agents are fully governed with deterministic orchestration and rollback. For third-party agents, Salesforce can scan and inventory them, but enforcement of guided determinism, Agent Script, and Trusted Agent Identity remains unproven. Enterprises should demand operational evidence of enforcement, not just discovery, before treating Agent Fabric as a neutral control plane. Agent workflows across vendor boundaries are the ultimate providing ground, but don’t expect progress by leaps and bounds quite yet. The control plane land grab is underway, but demonstrated outcomes are the true gating factor. Until then, multi-vendor agent support remains aspirational across all vendors that have announced similar initiatives.
Trusted Agent Identity: Early Signal on Non-Human Identity Governance
Trusted Agent Identity introduces mobile approval for high-risk agent actions, directly addressing the non-human identity governance problem that most platforms sidestep. This is meaningful early evidence that Salesforce recognizes the need for agent identity controls beyond traditional user authentication. It is not proof of maturity, but it sets a directional marker for the market. Enterprises should watch how Trusted Agent Identity evolves—will it extend to third-party agents, and can it support enterprise-grade audit and escalation workflows? This is an early step toward agent governance and control, not a finished solution.
MCP Bridge and Agent Broker: Enterprise-Grade On-Ramp, with Fragmented Availability
MCP Bridge is positioned as Salesforce’s enterprise-grade on-ramp to adopting the MCP protocol, lowering friction for organizations as they onboard multi-agent workflows. The Agent Broker, anchoring deterministic orchestration, is in beta now and targets general availability in June 2026. This fragmented availability means enterprises cannot yet rely on Agent Fabric for full multi-vendor governance at scale. The operational gap between beta and GA is material: early adopters must plan for evolving APIs, incomplete enforcement, and shifting integration boundaries. Until Agent Broker is GA and enforcement parity is demonstrated, Salesforce’s multi-vendor control plane claim remains aspirational. The innovation and integration road ahead is long, and the market will judge on operational outcomes, not platform ambition.
What to Watch:
- Enforcement Parity: Will Salesforce deliver operational enforcement for third-party agents, or will governance remain Salesforce-centric through 2026?
- Guided Determinism and Agent Script in Practice: Can fixed handoff rules with bounded LLM reasoning, operationalized by Agent Script, scale across multi-vendor agent ecosystems without introducing new friction or governance blind spots?
- Trusted Agent Identity Evolution: Will mobile approval and non-human identity governance extend to third-party agents and support enterprise-grade auditability?
- Agent Broker General Availability: Does the June 2026 GA milestone deliver on deterministic orchestration and rollback for all supported agent types, or will enforcement remain fragmented?
Read the full press release on Salesforce’s website.
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.
Other Insights from Futurum:
Will Salesforce’s Agentic Contact Center Force a Rethink of CCaaS Sourcing?
Salesforce Overhauls Consulting Track, Recasting Partners as Outcome Architects
Can Salesforce’s Data 360 Pricing Overhaul Deliver Predictable ROI?
Author Information
Mitch Ashley is VP and Practice Lead of Software Lifecycle Engineering for The Futurum Group. Mitch has over 30+ years of experience as an entrepreneur, industry analyst, product development, and IT leader, with expertise in software engineering, cybersecurity, DevOps, DevSecOps, cloud, and AI. As an entrepreneur, CTO, CIO, and head of engineering, Mitch led the creation of award-winning cybersecurity products utilized in the private and public sectors, including the U.S. Department of Defense and all military branches. Mitch also led managed PKI services for broadband, Wi-Fi, IoT, energy management and 5G industries, product certification test labs, an online SaaS (93m transactions annually), and the development of video-on-demand and Internet cable services, and a national broadband network.
Mitch shares his experiences as an analyst, keynote and conference speaker, panelist, host, moderator, and expert interviewer discussing CIO/CTO leadership, product and software development, DevOps, DevSecOps, containerization, container orchestration, AI/ML/GenAI, platform engineering, SRE, and cybersecurity. He publishes his research on futurumgroup.com and TechstrongResearch.com/resources. He hosts multiple award-winning video and podcast series, including DevOps Unbound, CISO Talk, and Techstrong Gang.
