Latest
AI Affairs, home

Wednesday 18 November 2026

Corporate

Microsoft launches AI Transparency Initiative with more than 40 organisations

Microsoft plans to put the coalition’s work into its AI, Foundry and Copilot products, giving enterprise customers evidence they can use with security and compliance tools.

Building 92 at Microsoft’s Redmond headquarters, with a Microsoft sign outside.
Photo: Coolcaesar, CC BY-SA 4.0, via Wikimedia Commons (cropped)

Microsoft launched an AI Transparency Initiative with more than 40 organisations on 17 November, saying the group will develop common standards for governing AI models and agents across providers. Chairman and CEO Satya Nadella announced the effort at Microsoft Ignite, the company said in its announcement. Microsoft plans to incorporate the resulting work into its own products so customers can use evidence about AI activity with their security and compliance tools.

Key points

  • Microsoft says participants include Accenture, Adobe, KPMG, Mastercard and Volvo Cars.
  • Microsoft plans to expand its participation in NIST’s SI Agent Standards Initiative and work with the participating organisations on open standards.
  • Microsoft says it will incorporate the work into Microsoft AI, Microsoft Foundry and Microsoft Copilot.

More than 40 organisations join Microsoft effort

Accenture, Adobe, KPMG, Mastercard and Volvo Cars are among the participating organisations, according to Microsoft. The company says it is working with customers and partners on standards that apply regardless of which provider supplies a model. It wants the work to cover transparency, security and governance as enterprises use AI systems connected to their data and tools.

Microsoft says it will expand its participation in the National Institute of Standards and Technology’s SI Agent Standards Initiative. Microsoft plans to use research from the NIST initiative to develop open standards. Collaboration with the more than 40 organisations forms a separate part of the effort, the company says.

The proposed approach is meant to span closed and open-weight models, Microsoft says. Its aim is to let developers change or update the models behind agents while IT and security teams continue to govern them and compliance teams continue to audit their actions. That requirement reaches beyond any one model supplier’s account of how an agent behaved.

Microsoft targets evidence of agent actions

Microsoft says organisations can observe activity around a model but often cannot reconstruct what occurred between a request and its outcome. That limits their ability to verify a model’s behaviour independently, according to the company. It says the difficulty becomes more consequential when agents can access enterprise data, tools, credentials and infrastructure, then act without approval at every step.

Hayete Gallot, executive vice president of Microsoft Security, said effective security controls require greater visibility into the model itself. Microsoft says its initiative is organised around seeing what an AI system accesses and does, verifying its actions independently and controlling what it is permitted to do. The company also wants teams to know which model and runtime an agent used, and what it was authorised to do.

StreetInsider reported that, under Microsoft’s proposed approach, organisations would use their own security systems to verify agent actions rather than rely solely on model providers. Microsoft says teams should monitor agent activity and risk signals continuously against their policies and thresholds.

Aanchal Gupta, chief security officer at Adobe, said security teams need visibility into what agents access, what actions they take and how those actions can be verified. John Israel, global chief information security officer at KPMG International, said clients want to move faster with AI and need transparency and oversight to do so safely. Both companies are among the participants Microsoft named.

Foundry and Copilot are in Microsoft’s plan

Microsoft says the third part of its initiative is to incorporate the standards work into Microsoft AI, Microsoft Foundry and Microsoft Copilot. The company says customers will be able to use the resulting evidence with security and compliance tools from Microsoft or other providers.

Microsoft describes an agent with access to data, tools and credentials as carrying risks comparable to those of a privileged insider. It says organisations remain accountable to customers, employees, boards, auditors and regulators for consequential actions taken through their systems. The company’s proposed controls would let them inspect what an agent was asked to do, what it accessed and what it did.

Sources

Topics: Agents, Enterprise adoption, Safety

What should we cover? Tell us

The weekly briefing

The AI Affairs weekly briefing: the week's 10 most important AI stories in business, government and markets, every Monday. We'll email you a link to confirm. Every email has a one-click unsubscribe. Privacy policy