THIS IS A PROMOTED POST

We’ve been talking about “digital transformation” in AV for, what, a decade now? Maybe longer? And let’s be honest, most of it has been marketing fluff. A few cloud dashboards here, a firmware manager there, and every booth claiming they’re “IoT-ready.” But here’s the truth: We haven’t actually transformed anything yet. Not until every mic, speaker, display, switch, sensor and control processor is connected to the cloud and to AI. That’s what Xyte calls Physical AI, and frankly, that’s the first time I’ve heard a term that makes sense. Because that’s when AV stops being a pile of boxes and starts acting like a system-of-systems.

And the wild part? The next two years, not ten, two, are going to define the next twenty. We’ve been in cruise control for four or five years now. Another soundbar, another display, another incremental upgrade that solves the same problems in a slightly shinier way. But 2026 and 2027? That’s the fork in the road. Either AV embraces Physical AI and real cloud platforms, or we slowly slide into “commodity hardware” territory, where the only differentiator left is who can ship the box faster. And if you’ve been following the IT industry for more than five minutes, you already know where that story ends.

In the Physical AI era, devices are not installed and then forgotten. They’re alive. Always on, always connected, always improving. A display or mic isn’t a one-time transaction anymore — it’s a service. It gets better over time with new AI capabilities, new features, new insights, and new automation. And for manufacturers, that’s not just nicer for the user, it’s a completely different business model. You’re not living and dying by the hardware margin anymore. You’re delivering ongoing value.

And then there’s support. If you think the future is still a human reading a ticket and Googling a manual, you’re dreaming. AI support agents are about to become your first-line techs, and they’re going to live in the cloud, not in some outsourced help desk. Once every device is connected to AI, the system can analyze telemetry, cross-reference it with documentation and past incidents, and resolve 30 to 40% of problems before a human ever sees them. That cuts OEM support costs, slashes MTTR for end users, and finally stops the endless back-and-forth of “send me a log file.”

Speaking of telemetry, it’s the new gold. But right now 99% of it evaporates into nothing. Every device spits out data, and nobody knows what to do with it. That’s the gap. We don’t need more dashboards. We need platforms that actually do something with the data, normalizing it, correlating it and turning it into automated actions or clear insight. That’s what a cloud AI layer is supposed to do.

And here’s the important part: A neutral cloud platform is the only realistic way to break vendor silos without killing ecosystems. Ecosystems aren’t going anywhere — let’s not kid ourselves — but a neutral data and AI layer allows anonymized insights to flow across brands, enabling a microphone from one company, a switch from another, and a display from a third to actually collaborate in diagnosing a problem together. That’s the dream we’ve been pretending to sell for years, but Physical AI is the first time it’s technically possible.

And Physical AI doesn’t mean everything has to move to the cloud tomorrow. The big misconception is that people think it’s cloud versus edge. It’s not. It’s both. Some intelligence will always need to live on the device for latency, resiliency and the basic “the network hiccuped and the room didn’t die” logic. But the real magic comes from the cloud, from the ability to learn from millions of events across thousands of deployments and orchestrate AI actions across a multi-vendor environment. The edge handles the moment; the cloud handles the collective intelligence.

Now, for AV and IT teams who are being told by their leadership, “Go do AI,” the place to start is simple: Pain. Uptime, support, lifecycle. Automated tier-one troubleshooting. Proactive maintenance. Smarter firmware management. Usage analytics that aren’t vanity metrics but actual decision-making inputs. Start there, and suddenly AI isn’t scary — it’s helpful.

And let’s talk about end users for a second. They don’t want ecosystems. They don’t care about ecosystems. They want everything to just work. If the mic is the issue, they don’t want to hear that it’s the switch. If the switch is the issue, they don’t want to hear it’s the display. Physical AI, when the system is connected to a common platform with a common language, such as OpenAV Cloud, enables devices to work together, identify the issue, escalate it to the right OEM, and resolve it. Sometimes before anyone even logs a ticket. That’s how you finally deliver “it just works” without lying.

And once AV is visible to AI, the entire category jumps from “nice-to-have tech” to a strategic business asset. When rooms, devices and usage patterns are part of the enterprise’s data fabric, AV starts informing real estate decisions, workplace strategy and employee experience — the stuff the board actually cares about.

So as we roll into ISE, here’s what I’ll be watching for: Proof that Physical AI is real. Not new boxes. Not shinier pixels. But cloud-connected devices, OpenAV Cloud interoperability, and AI workflows that make AV feel less like hardware and more like an intelligent system-of-systems. If we see that, we’re not just entering the next chapter — we’re finally entering the era AV has been promising for a decade.