Pull to refresh
Logo
Salesforce unveils enterprise AI harness to govern agents across the business

Salesforce unveils enterprise AI harness to govern agents across the business

New Capabilities

Six-capability harness previewed ahead of Dreamforce with seven named agents; stock up 23% into the conference, security and cost tooling still in progress

September 14th, 2026: Salesforce launches 7 named AI agents before Dreamforce

Overview

Updated Sep 15

Enterprises are putting AI agents to work across their business systems, which creates a new problem: controlling what those agents can see, do, and spend. Salesforce's answer, unveiled September 10, 2026, is the Trusted Enterprise AI Harness, a software layer that wraps AI in six capabilities: context, agency, action, governance, security, and models.

Salesforce previewed the harness ahead of Dreamforce 2026 and launched seven named Agentforce agents it will manage. The pitch lands as 85% of enterprises run 3.1 agent platforms at once, per TheOutpost. Salesforce stock rose about 23% in the three weeks before the conference, then gave back 6%.

Why it matters

Enterprises deploying AI agents face a governance decision: adopt Salesforce's harness, a rival like ServiceNow, or assemble point tools from startups.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

6
Core capabilities in the harness
Context, agency, action, governance, security, and models form the six pillars.
8
Existing Salesforce platforms contributing tech
The harness draws on Data 360, Informatica, MuleSoft, Agent Fabric, Tableau, Agentforce, Salesforce Guardian, and the Salesforce Platform.
Feb 2027
Unified experience target
Rollout begins early fiscal year 2028; some foundational technologies are available now.
MCP + APIs
Headless 360 delivery channels
Salesforce's headless layer exposes the platform through Model Context Protocol tools and APIs so the harness can reach Claude, Slack, and Microsoft Teams.
7
Named Agentforce agents launched before Dreamforce
Salesforce released seven named agents ahead of Dreamforce 2026; the harness is the management layer for them.
85%
Enterprises running multiple agent platforms
TheOutpost reports 85% of enterprises run 3.1 agent platforms simultaneously, the sprawl problem the harness targets.
23%
Stock run-up into Dreamforce
AInvest reports Salesforce shares rose about 23% in the three weeks before the conference, then gave back 6%.

Voices

Curated perspectives — historical figures and your fellow readers.

Ever wondered what historical figures would say about today's headlines?

Sign up to generate historical perspectives on this story.

People Involved

Organizations Involved

Timeline

March 2018 September 2026

7 events Latest: September 14th, 2026 · 1 month ago
Tap a bar to jump to that date
  1. Salesforce launches 7 named AI agents before Dreamforce

    Latest Product launch

    Salesforce released seven named Agentforce agents ahead of Dreamforce 2026. The Enterprise AI Harness, previewed but not yet generally available, is the management layer for these agents.

  2. Salesforce stock runs up 23% into Dreamforce, then gives back 6%

    Market reaction

    AInvest reported Salesforce shares rose about 23% in the three weeks before Dreamforce, trading near 30x forward earnings, before giving back about 6% in the final week.

  3. Salesforce publishes harness details; analysts flag security and cost gaps

    Follow-up

    Salesforce published the official press release detailing the six capabilities, the AI Control Plane, and headless delivery through MCP and APIs to surfaces like Claude, Slack, and Microsoft Teams. Launch coverage from diginomica quoted Kumar saying trusted security and FinOps capabilities still need to be built.

  4. TechTarget frames harness as Salesforce's challenge to ServiceNow

    Analysis

    TechTarget reported the Enterprise AI Harness is Salesforce's latest salvo aimed at ServiceNow. It noted Salesforce previewed MuleSoft Agent Fabric at Dreamforce 2025 and has since released it.

  5. Salesforce unveils Trusted Enterprise AI Harness

    Product launch

    Salesforce announced the six-capability AI harness and control plane; unified rollout starts early fiscal 2028.

  6. Agentforce launches at Dreamforce 2024

    Product launch

    Salesforce launched Agentforce, its agentic AI platform, positioning the company for autonomous agents.

  7. Salesforce announces MuleSoft acquisition for $6.5 billion

    Acquisition

    The deal brought MuleSoft's API management and integration tools, which anchor the harness's action layer.

Scenarios

1

Salesforce harness becomes the enterprise standard for agent governance

Possible Resolves by Q3 2027

Discussed by: Salesforce's announcement materials and Kumar's model-agnostic framing

Salesforce bundles the harness with Agentforce, MuleSoft, and Data Cloud and lands large enterprise deals. Intelligent model routing and the AI control plane win customers running multiple large language models. Pricing announced closer to general availability, likely early fiscal 2028, will determine how quickly adoption accelerates.

2

Adoption splits as customers mix the harness with rival point tools

Likely Resolves by End of 2027

Discussed by: diginomica's analysis flagging governance and cost-control gaps in the early formulation

Enterprises already using LangChain, Microsoft Copilot Studio, or Amazon Bedrock for some workloads keep those in place, turning to Salesforce only for CRM-adjacent agents. The harness becomes one layer among several rather than the single control plane Salesforce envisions. Independent analysts will measure whether the harness dominates or merely competes.

3

Harness becomes a neutral orchestrator across rival models

Possible Resolves by Jan 31, 2028

Discussed by: SiliconANGLE's coverage of intelligent model routing; Salesforce's headless and MCP strategy

Salesforce leans into open standards — Model Context Protocol (MCP) support, routing across third-party models, and headless delivery to Claude, Slack, or Teams. The harness functions as a routing and governance layer for any model, making model choice secondary. The company has said capabilities will extend beyond its own applications into other AI experiences.

4

Security and cost gaps slow the harness rollout

Possible Resolves by Apr 30, 2027

Discussed by: diginomica's report on Kumar's briefing remarks

Kumar conceded the security and FinOps pieces are not all built. If the unified experience ships in early fiscal 2028 without them, enterprises running regulated workloads may keep their own governance stacks and use the harness only for CRM-adjacent agents. Salesforce says it is working with customers on R&D to align AI costs with enterprise value, and plans a bigger role for Tableau in monitoring the data and AI estate.

5

ServiceNow rivalry defines the agent governance market

Likely Resolves by End of 2027

Discussed by: TechTarget's coverage of the launch

TechTarget framed the Enterprise AI Harness as Salesforce's latest salvo aimed at ServiceNow, which has been selling its own AI agents. Both vendors pitch themselves as the control layer for enterprise AI. The contest may come down to which installed base trusts its vendor for mission-critical agent workflows.

6

AI Control Plane Market Splits Across Salesforce, ServiceNow, Microsoft, AWS, and SAP

Likely Resolves by Q2 2027

Discussed by: The AI Economy substack analysis

The AI Economy notes the AI Control Plane functions similarly to ServiceNow's AI Control Tower, Microsoft's Agent 365, Amazon Bedrock AgentCore, and SAP's AI Agent Hub. Enterprises running multiple clouds may adopt several control planes rather than one, with Salesforce winning CRM-adjacent agents and hyperscalers winning their own ecosystems.

7

Dreamforce fails to produce customer proof points

Deadline passed (Q3 2026)

Discussed by: AInvest's market analysis noting the stock's 23% run-up into the conference

The stock ran up 23% into Dreamforce on harness enthusiasm. If the conference produces no named customers, no pricing, and no demonstration of third-party model management, the give-back could accelerate. Kumar has already conceded security and FinOps tooling are unfinished.

Historical Context

3 moments from history that rhyme with this story — and how they unfolded.

2008-2018

The API management wave (2008-2018)

When companies broke software into microservices and adopted software-as-a-service, APIs multiplied faster than teams could manage them. Vendors like MuleSoft, Apigee, and Kong sold governance layers to track, secure, and control access to those connections.

Then

Enterprises bought or built API gateways to manage connections between applications.

Now

MuleSoft became an enterprise standard and was acquired by Salesforce for $6.5 billion in 2018.

Why this matters now

AI agents are multiplying across enterprises the way APIs did. The harness is Salesforce's attempt to own the governance layer for agents the way MuleSoft did for APIs.

2015-2020

Cloud cost management and FinOps (2015-2020)

As businesses moved workloads to AWS, Azure, and Google Cloud, tracking and controlling cloud spending became a real problem. A wave of startups — CloudHealth, Turbonomic, Apptio — sold cost management tools to fix it.

Then

Cost management tools became standard practice for cloud-heavy enterprises.

Now

Cloud providers built native tooling and bought or displaced the startups.

Why this matters now

The AI Control Plane explicitly manages inference costs and agent performance, applying the FinOps model to AI agents.

1999-2005

Microsoft Active Directory (1999-2000s)

As Windows servers spread through corporate networks in the late 1990s, Microsoft introduced Active Directory to centralize identity and policy. It became the layer that governed who could access what.

Then

Enterprises standardized on Active Directory for identity management.

Now

The control layer anchored Microsoft's dominance of enterprise IT management.

Why this matters now

Whoever owns the identity and governance layer for AI agents gains a controlling position in the enterprise AI stack — the dynamic Salesforce is chasing.

Sources

(18)