Microsoft AI Updates: May 22, 2026

1. EY and Microsoft commit $1B over five years to push enterprise AI past pilots

Microsoft. EY and Microsoft announced a global initiative investing more than $1 billion over five years to help organizations scale AI and deliver enterprise-wide outcomes, pairing Microsoft Forward Deployed Engineers with EY industry professionals under Microsoft’s AI-native Hypervelocity Engineering approach. EY is acting as Client Zero, having already deployed Copilot to 150,000 users and recorded a 15 percent productivity boost that is being scaled to more than 400,000 employees on the new Microsoft 365 E7 Frontier Suite. Initial focus areas span finance, tax, risk, HR, and supply chain across financial services, industrials and energy, consumer and retail, government, and healthcare. Source

2. Microsoft Marketplace lands 59 new offers and refreshes terms of use

Microsoft. Microsoft published its May 21 Marketplace roundup, listing 59 new offers spanning cloud solutions, AI apps, and agents available for transact, try, and buy through the unified storefront. The post also flags that the Microsoft Marketplace Terms of Use update takes effect automatically on May 27, 2026 with no partner action required, alongside reminders that Agent 365 is now generally available with previews for discovering and managing shadow AI agents including local clients like OpenClaw and Claude Code. The cadence underscores Marketplace’s role as the primary transactable channel for Copilot- and Foundry-adjacent ISV inventory. Source

3. Purview gains Claude visibility as Windows 365 for Agents extends governance

Microsoft. Microsoft’s May security recap highlights new Claude integration with Microsoft Purview, giving enterprises centralized visibility and data governance over Anthropic model usage alongside existing OpenAI and in-house agent telemetry. Microsoft also paired Windows 365 for Agents with Agent 365 to provide a consistent, secure environment to run and govern agents, targeting the blind spots created as autonomous workloads spread across SaaS, browser, and local endpoints. The updates extend Microsoft’s bet that AI security must operate at the agent layer rather than the app or identity layer alone. Source