A new architecture that gives AI a shared understanding of the customer and the business — and enables it to act with trust
The Agentic Enterprise is changing how work gets done—and redefining the role people play as AI becomes increasingly embedded across business functions.
As AI agents take on increasingly complex tasks, they are moving beyond simple automation to understand what is happening, determine what should happen next, take action across multiple systems, and work alongside people and other agents.
That evolution creates a new challenge for enterprises: How can businesses give AI agents the context, intelligence, tools, and controls they need to work reliably, securely, and at scale?
Salesforce is addressing that challenge with its Enterprise AI Harness, a trusted foundation designed to bring together the capabilities agents need to understand the business, reason and plan, take action, and operate within enterprise controls—without requiring companies to build and manage those capabilities separately for every agent or AI experience.
Salesforce’s Enterprise AI Harness brings together six capabilities spanning context, agency, action, governance, security, and models, delivered through a common, composable architecture built on the customer relationships, processes, and controls that already run the business.
Alongside these capabilities, Salesforce is introducing a new AI Control Plane, giving businesses a centralized way to see, manage, and control agents and AI as they expand across the enterprise.
Customers can use all six capabilities together as one system or adopt only the capabilities they need. They can also combine Salesforce technologies with their existing technology environments, including third-party models, agents, and systems.
Why an Enterprise AI Harness Matters
Consider a seemingly straightforward customer question:
“Can we fulfill this order today?”
The answer rarely exists in a single system.
CRM contains information about the customer and the relationship. ERP manages inventory and fulfillment. Contracts contain commitments and entitlements.
Analytics provides business definitions, while policies determine what can be promised. Previous interactions provide important memory, and business processes determine how the work should actually be completed.
The challenge, therefore, is not simply connecting these systems.
An agent must understand all that information in the context of a specific customer and a specific moment, determine the appropriate action, execute it securely, and remain within the company’s policies and controls.
AI reasoning can be open-ended, but enterprise execution often cannot be. The Enterprise AI Harness is designed to connect flexible AI reasoning with the business rules, policies, and controls required for predictable execution.
It also makes these capabilities reusable across the enterprise. Context can be shared across agents and models, while actions and workflows can be securely invoked wherever they are needed. Governance and security can likewise be applied consistently as AI moves across different parts of the business.
As people and agents increasingly work together, every interaction and action can generate new signals, outcomes, and memory. Those insights can enrich the context available for subsequent work, creating a compounding intelligence loop over time.
While models will continue to improve and become more widely available, an enterprise’s proprietary context remains uniquely its own—including its customers, knowledge, semantics, memory, processes, relationships, and history.
As the underlying intelligence evolves, that proprietary context can continue to compound, becoming a durable source of differentiation.
Salesforce’s Enterprise AI Harness
The Enterprise AI Harness represents an architecture Salesforce says enterprises need for the AI era—and one the company has been building toward for decades.
The trusted customer data, metadata, semantics, business logic, workflows, permissions, security, and governance that companies already depend on to run their businesses can now serve as the foundation for how AI understands, reasons, and acts.
Salesforce is bringing together core technologies from Data 360, Informatica, MuleSoft, Agent Fabric, Tableau, Agentforce, Salesforce Guardian, and the Salesforce Platform through a common, composable architecture and unified experience for its Enterprise AI Harness.
For existing Salesforce customers, the technologies they already use provide the foundation for this expansion, with an upgrade path designed to unlock additional capabilities as they become available.
The Six Trusted Capabilities
Trust is designed into each of the six capabilities from the outset. Together, they provide AI with what it needs to understand the business, determine what to do, take action, and operate within enterprise controls.
For the question, “Can we fulfill this order today?”, each capability plays a distinct role.
Trusted Context
Trusted Context brings together customer context with data, metadata, semantics, knowledge, real-time signals, memory, and an understanding of how work gets done across the enterprise.
Grounded in governed enterprise data, it gives AI a shared understanding of the customer and the surrounding business context.
For an order, this means understanding the customer, their contract and entitlements, available inventory, and what “available” actually means for that particular customer.
Trusted Agency
Trusted Agency provides agents with the reasoning, planning, state, memory, collaboration, and orchestration required to pursue complex business outcomes.
It combines flexible AI reasoning with deterministic controls where certainty is required.
In the order example, it can help determine whether the order can be fulfilled and what should happen next while following business rules that require certainty.
Trusted Action
Trusted Action securely connects AI to applications, APIs, workflows, tools, and business processes, allowing agents to move from understanding what needs to happen to actually doing it.
An agent can use these capabilities to reserve inventory, update an order, trigger fulfillment, or involve a person when necessary.
The outcomes of those actions can then flow back into Trusted Context, helping enrich the information available for future interactions.
Trusted Governance
Trusted Governance helps enterprises govern the data, metadata, policies, and processes that AI relies on, with capabilities for lineage, quality, guardrails, and controls.
It helps ensure that an answer is based on trusted information and that agents follow the policies governing how an order or other business process should be handled.
Trusted Security
Trusted Security applies identity, permissions, privacy, data protection, and runtime security to what AI can access and what agents are permitted to do.
This helps ensure that an agent can only access customer information and perform actions for which it has the appropriate authorization.
Trusted Models
Trusted Models gives enterprises flexibility to securely connect the appropriate model for each job, with intelligent model routing based on factors such as accuracy, performance, cost, and business requirements.
This allows work to be directed to the right intelligence for a particular task while giving businesses the flexibility to change models as AI technology continues to evolve.
“The Agentic Enterprise won’t be defined by which model a company chooses. Models will keep changing, and intelligence will increasingly be available everywhere. What will differentiate an enterprise is the trusted, proprietary context it brings to that intelligence — starting with the customer — and its ability to turn that context into action securely,” said Rohan Kumar, President, Chief Platform and Engineering Officer.
“That’s what we’re building with Salesforce’s Enterprise AI Harness: a system where every interaction and action can create new context and intelligence, making the enterprise’s context richer over time while operating with the security, governance, and control businesses need to trust AI at scale.”
A New AI Control Plane for the Enterprise
As agents and AI spread across organizations, businesses need a consistent way to understand what AI is operating across their enterprise, how it is performing, what it is authorized to do, and how much it costs.
Salesforce’s new AI Control Plane is designed to provide a common place for businesses to see, manage, and control agents and AI across the enterprise.
It enables companies to:
- Discover and register agents and AI capabilities
- Establish identity and policy
- Manage AI and agent lifecycles
- Evaluate performance
- Observe behavior and outcomes
- Control costs
These capabilities extend across Salesforce and third-party AI, giving enterprises a consistent layer of visibility and control as AI expands across teams, applications, models, and systems.
The approach is intended to eliminate the need to manage every agent or AI experience separately.
Built for Choice and Openness
Salesforce’s Enterprise AI Harness is designed to meet customers wherever—and however—they choose to build and use AI.
A new unified experience is intended to make it easier for different teams to work with the capabilities most relevant to their roles, from context and governance for data leaders, to identity and policy for security leaders, and agents, models, and actions for builders.
The Enterprise AI Harness is also being built headlessly from the ground up, with capabilities accessible through technologies including MCP, APIs, Skills, and Plug-ins.
This approach allows Salesforce capabilities to extend beyond traditional Salesforce applications and into the AI experiences where people already work—including Claude, Slack, Microsoft Teams, Agentforce, and other AI experiences.
It also advances Salesforce’s broader AIforce strategy, which aims to bring Salesforce capabilities to any AI, agent, app, or surface.
“AI is moving faster than any technology we’ve seen — which is exactly why composability matters so much to us. We don’t want to bet our future on one closed stack; we want the freedom to adapt as the landscape shifts. That’s what resonated with us about Salesforce’s new Enterprise AI Harness,” said Shawn Malhotra, CTO, Rocket Mortgage.
Availability
Many of the technologies that form the foundation of Salesforce’s Trusted Enterprise AI Harness are available today, while new capabilities and the unified experience are planned to begin rolling out in early fiscal FY28.
Existing customers will be able to upgrade eligible Salesforce investments to unlock new capabilities as they become available.
Salesforce said it will announce additional details regarding availability, packaging, pricing, and upgrade paths as the Enterprise AI Harness moves closer to general availability.
Pricing and packaging remain subject to change. Availability may vary by region and is governed by customer agreements. Customers should base purchasing decisions on products and services currently available.
