Microsoft will use its InfoComm 2026 keynote to stake a bold claim: the conference room is no longer just a place for meetings—it’s an intelligent edge node in the enterprise AI fabric. Ilya Bukshteyn, corporate vice president for Teams Calling, Meetings and Devices, takes the stage on Wednesday, June 17, at 1:30 p.m. to lay out a vision where Teams Rooms devices act as AI endpoints, hosting autonomous agents that transform how hybrid teams collaborate.

This isn’t just another product update. Bukshteyn’s keynote signals a strategic shift for Microsoft, one that merges the physical meeting space with the cloud AI platforms already reshaping business workflows. For the professional AV and IT professionals gathered at InfoComm 2026, the message is clear: the line between AV equipment and IT infrastructure has dissolved, and Teams Rooms is the control point.

The Convergence of AV and IT Comes of Age

For years, the pro AV industry has talked about AV/IT convergence. Microsoft’s keynote makes it tangible. Teams Rooms devices—whether the compact Surface Hub 3, the modular Teams Rooms systems from Crestron and Logitech, or the bring-your-own-device (BYOD) setups—are no longer isolated conferencing tools. They are data‑generating, AI‑processing endpoints that tie directly into Microsoft 365 Copilot, Graph, and the broader Azure AI ecosystem.

Consider what happens in a standard hybrid meeting today: video feeds, shared content, and chat streams flow between on‑premises and remote participants. Microsoft sees a near future where AI agents running locally on the Teams Rooms compute module or intelligently offloaded to the cloud listen, transcribe, and generate meeting notes in real time. They draft action items, flag decisions, and even pull relevant documents from SharePoint without a human lifting a finger. The meeting room becomes a proactive participant, not a passive conduit.

AI Meeting Agents: From Hype to Hardware

The term “AI meeting agent” has been buzzing through tech circles since Copilot for Microsoft 365 arrived. But Bukshteyn’s keynote will argue that agents only reach their full potential when they can perceive the physical meeting space—who is speaking, which whiteboard is in use, and what body language suggests. Teams Rooms cameras with spatial audio and AI‑enhanced video feed this data into the agent fabric. An agent that knows you walked to the whiteboard can automatically capture a high‑resolution shot of your diagram, run handwriting recognition, and inject the cleaned‑up graphic into Loop components shared with the entire team.

This shift requires compute at the edge. Microsoft has been seeding the ground with the recent general availability of Windows 365 on Teams Rooms and the certification of AI‑accelerated mini‑PCs from partners. Those devices pack enough neural processing power to run small language models locally—handling sensitive meeting data without ever sending it to the cloud. The result is a distributed AI architecture that satisfies both the latency demands of real‑time conversation and the compliance requirements of heavily regulated industries.

Hybrid Work’s Second Act

The pandemic forced a rapid pivot to remote work. Now, the conversation has moved to the harder, more expensive problem: making hybrid work feel equal. Microsoft’s own Work Trend Index research shows that 67% of employees want to be in the office at least part‑time, but the number‑one friction point remains the gap between in‑room and remote participant experiences. Teams Rooms with intelligent speakers and multi‑camera setups already narrow that gap, but Bukshteyn’s vision goes further.

Imagine a Teams Room that recognizes you when you enter, signs you in with facial recognition, and immediately surfaces your personalized Copilot dashboard on the front‑of‑room display. Remote attendees get an adaptive gallery where AI separates each in‑room voice into an individual audio stream, labels them by name, and even translates speech in real time. When a whiteboard is used, the remote view receives a crisp, glare‑free digital canvas overlaid on top of the physical person drawing—not a washed‑out camera feed pointed at a shiny surface. All of this happens invisibly, orchestrated by agents that understand the meeting intent and the room’s capabilities.

The Governance Bridge

One phrase in the keynote’s teaser deserves particular attention: “Pro AV IT Governance.” Historically, meeting spaces lived in a grey zone: AV integrators designed and installed the hardware, while IT was responsible for network and security. That division of labor has created security gaps. Unpatched room systems, default passwords, and unauthorized apps are commonplace. By framing Teams Rooms as AI endpoints, Microsoft is making the case that these rooms must fall under the same strict IT governance as any other managed endpoint—conditional access, zero‑trust policies, and real‑time threat detection.

The Teams Rooms Pro Management portal already gives IT admins granular control. The next step, likely to be demonstrated on stage, is integrating those room devices into Microsoft’s Security Copilot. An AI agent continuously monitors room health, automatically applies firmware updates, quarantines devices that show suspicious behavior, and even predicts hardware failures—alerting the facilities team before a projector bulb fails in the middle of a board meeting. For the AV integrator, that doesn’t mean becoming a help‑desk drone; it means shifting from break‑fix to designing intelligent, self‑healing environments.

What to Expect from the Keynote

Bukshteyn is known for blending technical depth with a showman’s touch. Attendees can expect on‑stage demos that connect an actual Teams Room to the same Copilot agents currently available in Office apps. One rumored demo involves a “meeting‑end agent” that, immediately after a call, drafts a summary, creates Project Plan tasks from spoken deadlines, and even pre‑books the next review session—all without human prompting. Another demo may show how multimodal agents can answer questions about content shared during the meeting, such as, “What was the budget figure mentioned by Sarah?” and get back a direct quote with source video timestamp.

Hardware announcements are also in the pipeline. Several OEMs are preparing to launch Teams Rooms kits built around the new Windows on Arm compute modules, promising lower power consumption and silent operation. Because the NPU on these devices can run AI models locally, Microsoft will highlight how sensitive video feeds for people‑counting and room occupancy analytics stay on‑premises, addressing privacy concerns head‑on.

The Competitive Landscape

The InfoComm keynote doesn’t exist in a vacuum. Zoom Rooms has been aggressively pushing its own AI Companion, and Google’s Gemini for Workspace is creeping into meeting hardware through ChromeOS‑based devices. Microsoft’s differentiation is the depth of the Graph—the data graph that connects emails, chats, calendars, and documents. A Teams Rooms agent doesn’t just transcribe; it understands context from the three‑email thread that preceded the meeting and the PowerPoint deck shared two days earlier. That context makes the agent’s suggestions genuinely useful rather than a novelty.

But the real test will be execution. The AV industry is notoriously slow to refresh hardware cycles; many conference rooms still run early‑generation Teams devices with underwhelming processors. Microsoft will need to show a clear migration path that doesn’t leave those rooms behind. The keynote may tease a “Teams Rooms Essentials” agent plan that delivers a subset of AI features to older hardware via cloud processing, while full on‑device agents remain exclusive to certified, AI‑ready systems.

Practical Impact for Windows Enthusiasts and IT Pros

For Windows‑focused IT professionals, this keynote reinforces the centrality of the Windows platform. Teams Rooms on Windows remains the premium experience, and Microsoft’s AI roadmap ties closely to Windows 11’s evolving AI stack. Expect to hear how the same WinML APIs and the Windows Copilot Runtime that power AI experiences on the desktop are being extended to room devices. This convergence means that skills learned managing Windows endpoints translate directly to managing AI‑enabled meeting spaces.

Developers will also get a nod. The Teams Rooms developer platform, which allows third‑party apps to run on room consoles, is expanding. Bukshteyn’s team has been working on a SDK that lets ISVs build meeting‑agent extensions—custom agents that, say, can connect to a proprietary CRM and surface customer insights during sales calls. This opens a new app ecosystem for the meeting room, one that could be as lucrative as mobile apps were fifteen years ago.

Looking Ahead

The keynote will likely close with a road‑map tease: features rolling out in waves through late 2026 and into 2027. The most ambitious piece—fully autonomous agents that can not only summarize but also proactively schedule, negotiate, and execute follow‑up actions—remains on the horizon, but the building blocks are shipping now. For the 40,000 attendees of InfoComm, Bukshteyn’s central thesis will be hard to miss: if you’re not thinking of your meeting spaces as AI endpoints, you’re already behind.

Microsoft is betting that Teams Rooms can become the operating system for the physical workplace, just as Windows became the OS for the personal computer. The hardware, the software, and the cloud intelligence are finally aligned. The stage is set for June 17 at 1:30 p.m. to see if Bukshteyn can convince the industry that the smart room era has truly arrived.