Skip to main content

Microsoft announces AI transparency initiative

18 NOVEMBER 2026·2 MIN READ·3 SOURCES

Microsoft announced its AI Transparency Initiative at Microsoft Ignite on November 17, 2026.

Key takeaways · 4

  • 01

    Microsoft says it is working with more than 40 organizations, including Accenture, Adobe, KPMG, Mastercard and Volvo Cars.

  • 02

    The proposed framework calls for continuous monitoring of agent activity and risk signals against policies and thresholds.

  • 03

    Microsoft says evidence should identify whether information about an agent’s identity, delegated intent and actions came from the model, its runtime or independent systems.

  • 04

    Microsoft says there are no release dates, product names for the features or pricing in the announcement, and nothing is available to deploy for now.

A cross-provider initiative

Microsoft announced the AI Transparency Initiative at Microsoft Ignite on November 17, 2026[1]. The company says the effort aims to establish a common, open approach to AI security and governance across models and providers[2]. Microsoft says more than 40 organizations are working with it, and named Accenture, Adobe, KPMG, Mastercard and Volvo Cars among the participants[2]. The initiative is guided by the principle that greater AI autonomy requires greater transparency, Microsoft says[2]. It is also intended to work across closed and open-weight models[2].

What the framework would record

Microsoft says existing controls can monitor activity around AI models but cannot see inside them or independently verify models’ accounts of their own behavior[2]. Its framework calls for evidence about what an AI system was asked and authorized to do, which model and runtime it used, what data and tools it accessed, and what actions it took[2]. Microsoft also calls for continuous monitoring of agent activity and risk signals against policies and thresholds[2]. Evidence about identity, delegated intent and actions should distinguish information from the model or runtime from information supplied by independent systems, the company says[2].

Verification, accountability and availability

Microsoft says organizations should be able to independently verify AI activity using their own security and compliance systems[2]. Evidence about a workflow’s purpose, owner, data and delegated authority can help enforce policies before consequential actions, while records can support investigations, recovery and audits, according to the company[2]. Microsoft says it will incorporate the initiative’s work into Microsoft AI, Microsoft Foundry and Microsoft Copilot, and customers will be able to use resulting evidence with Microsoft or third-party security and compliance tools[2]. The announcement provides no release dates, feature product names or pricing; nothing is available to deploy for now[3].

Teams evaluating agentic AI need to consider not only what an agent can do, but what evidence they will be able to examine and verify. The initiative sets out a proposed approach, but the announcement does not provide deployable features, dates or pricing, so professionals should treat implementation planning as unresolved.

Why it matters
For teams

AI training for your team

kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.

How this developed

  1. 19 November 2026

    The evidence adds that Microsoft says agents may act without approval at every step, the initiative is intended to cover closed and open-weight models, and accountability remains with the organization deploying AI.

  2. 18 November 2026

    Microsoft announces AI Transparency Initiative

Sources

AI fluency, one session a day, built for your work.