
What happened
Salesforce introduced an Enterprise AI Harness built on six capabilities — context, agency, action, governance, security and models — plus a new AI Control Plane.
Why it matters
Instead of each agent needing its own separate stack, enterprises can now use one composable system across Salesforce and third-party models and agents.
What to watch
New capabilities and a unified experience are planned to begin rolling out in early fiscal FY28. Specifics on availability, pricing and upgrade paths will come closer to general availability.
WHO IT HITSEnterprise IT and platform teams that currently stitch together separate systems for each AI agent will be able to manage context, actions, governance and security centrally. Existing Salesforce customers get an upgrade path to unlock new capabilities as they become available.
Summaries like this, every morning.
Salesforce frames the Enterprise AI Harness as a response to a specific enterprise problem: no single system holds the full answer when an agent is asked something like "Can we fulfill this order today?" CRM knows the customer, ERP knows inventory, contracts hold commitments, analytics provides business definitions, and policies determine what can be promised. The Harness is designed to connect that reasoning to the business rules and controls required for predictable execution.
The six capabilities are meant to be reusable across the enterprise — context shared across agents and models, actions invoked wherever needed, and governance and security applied consistently. Salesforce says the architecture draws on Data 360, Informatica, MuleSoft and Agent Fabric, Tableau, Agentforce, Salesforce Guardian and the Salesforce Platform. The AI Control Plane is positioned as a common place to see, manage and control agents across Salesforce and third-party AI.
The competitive bet, as Rohan Kumar puts it, is that enterprises will differentiate on proprietary context rather than model choice. The outcome hinges on whether customers can upgrade eligible Salesforce investments smoothly and whether the promised unified experience arrives on schedule. For data and security leaders, the test is whether the control plane delivers the consistent visibility and policy enforcement the announcement describes.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Oracle Corp. will make Google LLC's TPU AI accelerators available through its public cloud, marking the first…
ITmedia reports NTT is adjusting plans to build a data center in Osaka for Anthropic, costing up to 最大1兆円規模 wi…

NVIDIA and AMD each said they will expand AI partnerships with Meta, supporting its AI agent Muse and related…

Bona Film Group's Sanxingdui: Future Past, China's first fully AI-generated feature to reach cinemas, took in…

Qun Wan and Meiqin Xu sued OpenAI, alleging GPT-4o drafted lamotrigine taper schedules taking their daughter B…

Anthropic, a public benefit company, gives its Long-Term Benefit Trust the right to elect a majority of its bo…
