Zenity, the main end-to-end safety and governance platform for AI brokers, introduced the overall availability of runtime safety controls for AI brokers constructed on Microsoft Foundry as a part of an expanded partnership with Microsoft. Constructing on the inline prevention capabilities introduced this previous November, this integration permits enterprises to soundly deploy and scale Foundry functions and brokers by delivering steady, inline safety in opposition to runtime threats.
Additionally Learn: AiThority Interview with Glenn Jocher, Founder & CEO, Ultralytics
“This expanded partnership with Microsoft permits clients to construct and run AI brokers in manufacturing with confidence, with capabilities that block agent misuse in real-time, earlier than information strikes or instruments execute.”
As organizations transfer from experimentation to manufacturing, safety dangers are shifting from mannequin improvement to real-time execution. Brokers more and more make selections, chain actions, and invoke instruments throughout enterprise environments. These behaviors introduce new courses of threat that conventional, prompt-level or post-execution controls weren’t designed to deal with.
“With brokers, dangers emerge and persist at runtime,” stated Michael Bargury, CTO and co-founder of Zenity. “This expanded partnership with Microsoft permits clients to construct and run AI brokers in manufacturing with confidence, with capabilities that block agent misuse in real-time, earlier than information strikes or instruments execute.”
Rising buyer adoption of homegrown brokers has elevated demand for manufacturing‑prepared protections that perceive agentic conduct and intent. The expanded Zenity and Microsoft Foundry integration delivers:
- Finish-to-end agent runtime protection throughout fashions, instruments, information and enterprise methods
- Inline prevention for a number of risk courses, together with delicate information leakage, secret and credential publicity, jailbreak makes an attempt and power misuse (coming quickly)
- Agent-aware context and behavioral enforcement that evaluates selections and chained actions quite than remoted prompts
Collectively, these capabilities assist organizations transfer past level options towards a extra complete strategy to securing brokers in manufacturing.
Prospects are constructing and deploying brokers on Microsoft Foundry throughout IT operations, buyer help, finance, healthcare, manufacturing, and the general public sector. These brokers usually join on to enterprise assets comparable to SharePoint, OneDrive, databases, SaaS platforms, and inner APIs. Whereas this connectivity unlocks vital enterprise worth, it additionally will increase publicity when brokers are misconfigured, manipulated, or abused at runtime.
Zenity integrates natively into the agent execution path inside Foundry, enabling safety groups to detect, disrupt, and stop threats as they happen, earlier than information is uncovered, instruments are misused, or methods are impacted.
“Inline safety solely works if it’s a part of the execution path,” Bargury added. “Zenity integrates the place brokers truly function, guaranteeing that autonomy doesn’t come at the price of safety or compliance.”
Additionally Learn: The Infrastructure Warfare Behind the AI Increase
[To share your insights with us, please write to psen@itechseries.com]
