TLDR IT 2026-05-19

Google introduced a broad set of Gemini updates spanning Search, Workspace, Android XR, developer tooling, and agentic AI workflows at I/O 2026. The announcements reinforce how major platforms are moving toward AI-native operating layers where assistants increasingly orchestrate actions across apps, devices, and enterprise workflows rather than functioning as standalone chatbots.

Microsoft's AI strategy is struggling to gain meaningful user adoption despite aggressive Copilot integration across Windows and Microsoft 365. The company is reducing Copilot's footprint in parts of Windows 11 and shifting toward more targeted AI integrations instead of embedding AI into every workflow.

Boomi positioned integration platforms as the orchestration layer for enterprise AI agents, focusing heavily on identity, workflow coordination, API governance, and cross-system automation. Enterprise AI deployments are increasingly running into classic IT integration problems of fragmented SaaS data, workflow sprawl, and inconsistent operational controls, pushing iPaaS vendors back into the center of enterprise architecture conversations.

The article recommends practical controls for enterprises dealing with uncontrolled AI adoption across departments, including centralized procurement, identity integration, approved experimentation environments, and lightweight governance models. The goal is to avoid fragmented tooling, rising support overhead, and security gaps without slowing down internal AI experimentation.

The piece walks through the operational basics many companies still struggle to standardize as SaaS environments scale: app ownership, onboarding/offboarding automation, SSO coverage, shadow IT discovery, license management, and admin accountability. It's a useful reminder that most SaaS risk comes from inconsistent operational processes rather than the apps themselves.

Every IT team is fielding the same question this year: what's the AI plan? Atlassian's free webinar, 5 Tips for AI Service Management, cuts through the hype with practical guidance on where AI is moving the needle in ITSM today. To learn about virtual agents, request deflection, and knowledge automation, register here

IBM has introduced “Forward Deployed Units” (FDUs), small senior-led teams supported by AI agents, to help enterprises move AI projects from experimentation into large-scale production faster. The model combines engineering, governance, and business expertise into integrated teams that work directly with clients to deliver measurable outcomes more efficiently.

Iris is a Context Engine meant to give enterprise AI agents a shared context and memory layer. It combines context retrieval, agent memory, and data integration so agents can pull from enterprise systems and maintain state across longer-running tasks.

Model lock-in may matter less than workflow, governance, integration, and services-layer lock-in. As enterprise AI moves from chatbot pilots to embedded agents, the control plane around agents may become the stickiest part of the stack.

CEOs are increasingly pushing IT leaders to balance aggressive AI adoption with tighter cost controls, stronger cyber resilience, and clearer business outcomes from technology investments. The article also highlights growing expectations around operational simplification, workforce productivity, and cross-functional leadership beyond traditional infrastructure management.

Microsoft will retire Together Mode in Teams next month and shift users to the standard Gallery view.

Anthropic will let Project Glasswing partners share Mythos-generated security findings, tools, and code with outside organizations under responsible disclosure norms.