What happened
Salesforce previewed the Enterprise AI Harness and AI Control Plane. The Harness covers six feature areas including data accuracy, security and intelligent model routing; some technologies are already in its cloud services, with the rest rolling out early in fiscal 2028.
Why it matters
Enterprise AI Harness aims to address agents' nondeterministic nature by adding deterministic reliability controls, while AI Control Plane centralizes monitoring of agents' performance and inference costs. Salesforce president Rohan Kumar says trusted, proprietary context will differentiate enterprises.
What to watch
The rollout hinges on the early fiscal 2028 timing, which begins next February. Watch how the six feature collections—including Trusted Context and data lineage—are adopted by cybersecurity teams regulating agent data access.
WHO IT HITSEnterprise IT and cybersecurity teams deploying AI agents will gain tools to control which agents use which datasets, monitor inference costs, and apply reliability controls—areas the body identifies as requiring better management.
Ask the AI about this article →
Summaries like this, in your inbox every morning.
Salesforce's preview arrives as companies increasingly deploy AI agents—large language models extended with prompts, database connectors and other technical assets collectively known as a harness. The company says many of the technologies underpinning the Enterprise AI Harness are already available in its cloud services, suggesting it is building on existing capabilities rather than starting from scratch. The offering's six feature collections include Trusted Context, which brings together data assets and real-time signals, and tools like data lineage and quality controls.
The AI Control Plane complements the Harness by giving administrators a central view of agent performance and inference costs and helping enforce cybersecurity policies. Salesforce's president and chief platform and engineering officer, Rohan Kumar, framed the strategy around trusted, proprietary context: "Models will continue to change," he said, and what will differentiate an enterprise is how it securely turns that context into action.
The outcome hinges on customer adoption of these features, particularly among cybersecurity teams that will regulate which agents can access which datasets. The early fiscal 2028 timeline, beginning next February, is the next milestone to watch. It remains to be seen whether the model-routing and reliability controls deliver the promised cost and accuracy benefits in practice.
For example, today's edition would include:
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. Q&As are published on this page for other readers too.
Much of the attention on AI infrastructure buildouts is now tied to sheer compute power, with dominance define…
Barron's reported September 10 that Kepler Computing emerged from stealth with a memory architecture using fer…
Dynatrace acquired Arize AI, adding AI observability, evaluation and agent monitoring to its application obser…
Reuters reported September 10 that inference-chip startup d-Matrix will use Nvidia's NVLink Fusion to connect…
A Daily Dose of Data Science test kept LoRA adapters separate from a shared 7B base model, cutting 100 fine-tu…
A report by Spencer Kitts, Thomas Larsen and Sydney Von Arx says an OpenAI agent swarm very likely ran an atta…
