There was a time when a damaged assembly room meant calling IT, ready for somebody to indicate up, and hoping they’d determine whether or not it was the monitor, the cable, the ClickShare, or simply somebody who modified the enter on Friday and forgot to vary it again. You misplaced 20 minutes. Possibly you rescheduled. Possibly you gave up and used your laptop computer within the hallway.
Now, apparently, an AI agent named Lena can stroll into the room earlier than you do, test each machine, match configurations in opposition to a known-good state, repair what’s damaged, and alert an admin if one thing wants human consideration. The assembly begins on time. IT by no means will get the decision. And no person has to guess which HDMI enter the CEO left it on after watching Formulation 1 highlights over the weekend.
The constructing doesn’t simply exist anymore. It troubleshoots itself.
Sam Kennedy, EVP at NetSpeek, oversees a verticalized AI platform referred to as Lena that focuses particularly on the executive and operational aspect of AV methods. The platform makes use of an agentic structure—a number of AI brokers, every skilled on particular producers like Neat, Barco, Samsung, Poly, and BrightSign—to watch, troubleshoot, and restore assembly room configurations mechanically.
What makes this completely different from generic AI instruments is specificity. Lena doesn’t guess. It’s skilled on producer launch notes, API documentation, and configuration requirements, saved in a separate database referred to as Iris that eliminates hallucinations. When one thing breaks, Lena is aware of what “working” appears to be like like for that particular machine in that particular room—and might put it again.
One early buyer advised Kennedy that room-check alone would save them 1,000,000 {dollars}. The product is already in market, prospects are utilizing it, and NetSpeek is prepared for proofs of idea.
The Million-Greenback Room Examine
Kennedy’s first demo is straightforward: automated room checks. Somebody walks right into a convention room over the weekend, adjustments the monitor enter, adjusts the amount, mutes the system, and walks out. Monday morning, staff arrive for a gathering and the room doesn’t work.
In the present day, organizations remedy this by sending individuals into rooms bodily to confirm configurations. NetSpeek has automated that course of.
Lena runs scheduled checks and compares the present state of each machine in opposition to a known-good configuration. If the monitor is on the fallacious enter, Lena switches it. If the amount is muted, Lena unmutes it. If the system is offline, Lena flags it for an administrator.
“One in every of our early prospects has advised us simply this characteristic alone would save them 1,000,000 {dollars}. The explanation we picked room test as the primary use case is as a result of it’s a common factor. Everybody I talked to is having individuals go to those rooms.”
The configuration interface is simple. An administrator specifies which units are in a room, how they’re linked, and what state they need to be in for various modes: video convention, presentation, digital signage. From there, Lena handles the remaining.
The system works throughout producers. Kennedy demonstrated it with Barco ClickShare, however the identical logic applies to Neat units, Poly methods, Samsung shows, and some other supported {hardware}. The aim isn’t vendor lock-in, however seamless interoperability on the operational layer.
Associated Tales
Pure Language Troubleshooting (With Guardrails)
The second functionality Kennedy demonstrated is extra bold: pure language troubleshooting. An admin can describe an issue in plain English—”ClickShare isn’t working”—and Lena will diagnose and repair it.
In Kennedy’s demo, the precise downside was that the monitor was turned off, not that ClickShare had failed. However the person didn’t know that, and neither did the assistance desk. Lena checked the ClickShare (working wonderful), checked the show (off), and provided to show the monitor on. Downside solved remotely, no truck roll, no wasted time diagnosing the fallacious difficulty.
Kennedy was requested whether or not this functionality might lengthen to finish customers—whether or not staff in a damaged room might ask Lena immediately to repair it, relatively than calling IT. His reply was cautious however clear: not but, and never with out strict entry controls.
“You don’t need them to go, ‘reboot all of those units.’ I imply, it’s one of many demos. I exploit pure language to reboot all the pieces. You don’t need that. So that you’ve bought to construct quite a lot of controls in there. And that’s the place we haven’t completed it but.”
The technical functionality exists. The governance mannequin doesn’t. And that distinction issues, as a result of when you give a system the flexibility to behave on pure language instructions, you want guidelines about who can ask what—and what occurs when somebody asks the fallacious factor.
Why It Doesn’t Hallucinate
Kennedy hung out on structure, which is uncommon for a vendor pitch however revealing. The issue with utilizing client AI instruments for AV troubleshooting is that they don’t know the area. They guess. They hallucinate. They pull from Reddit threads and outdated boards.
NetSpeek constructed a separate data system referred to as Iris that ingests documentation immediately from producers—launch notes, API specs, configuration guides—scrapes and codecs it, and feeds it right into a trusted database. A human engineer evaluations it to make sure accuracy. That database trains Lena.
The result’s domain-specific intelligence that doesn’t invent solutions. When Lena troubleshoots a Neat machine or a Samsung show, she’s working from the producer’s personal documentation, not a language mannequin skilled on the open web.
This issues as a result of the AV trade is fragmented. Organizations don’t standardize on one vendor. A troubleshooting software that solely works for one producer doesn’t remedy the operational downside. NetSpeek’s pitch is that Lena works throughout all of them.
When Effectivity Turns into Autonomous
Proper now, Lena is a software. An administrator asks a query, Lena diagnoses and suggests a repair, the administrator approves it, the repair occurs. There’s a human within the loop.
However Kennedy additionally demonstrated that Lena can execute instructions in pure language: “reboot all units,” “activate the monitor,” “change to HDMI 3.” The aptitude exists. What doesn’t exist but is the entry management framework that will enable finish customers to work together with Lena immediately.
However the trajectory is obvious. If Lena can diagnose sooner than a assist desk, repair issues remotely, and function throughout a number of distributors, the subsequent logical step is to take away the intermediary. Let the room repair itself. Let the person ask immediately.
None of this requires dangerous intentions. It simply requires optimization. And when you’ve automated prognosis and restore, it’s onerous to justify conserving a human within the loop for routine fixes—particularly when that human prices money and time.
The query isn’t whether or not assembly rooms will get smarter. They already are. The query is what occurs when the system that fixes your room additionally decides what “mounted” means—and whether or not “working appropriately” is outlined by person expertise, or by operational effectivity, uptime metrics, and price discount.
Be part of the Dialog
That is the place AV operations are heading—whether or not assist desks are prepared or not. When you’re managing assembly rooms at scale, you’re not alone.
Be part of the UC In the present day publication for weekly insights on collaboration expertise, AI within the office, and the way forward for AV operations.
Be part of the UC In the present day LinkedIn group to attach with IT and AV leaders navigating the identical operational challenges.
Comply with Rob Scott on LinkedIn for extra on AV operations, office expertise, and AI within the enterprise.
Steadily Requested Questions
What’s NetSpeek’s Lena platform?
Lena is an AI platform constructed for AV operations. It makes use of a number of AI brokers skilled on particular producers (Neat, Barco, Samsung, Poly, BrightSign) to watch, troubleshoot, and restore assembly rooms mechanically. The platform is skilled on trusted producer documentation to keep away from hallucinations.
How does automated room-check work?
Automated room-check runs on a schedule and compares every machine’s present state in opposition to a known-good configuration. If one thing is fallacious—monitor on fallacious enter, quantity muted, system offline—Lena both fixes it mechanically or alerts an administrator.
Can finish customers work together with Lena immediately?
Not but. NetSpeek has not enabled direct end-user entry as a result of customers might difficulty harmful instructions like “reboot all units.” Strict entry controls are being constructed. For now, Lena is admin-facing solely.
What occurs when AI can repair assembly rooms autonomously?
The infrastructure exists for Lena to maneuver from assisted troubleshooting to autonomous operation. This improves effectivity however shifts decision-making from individuals to algorithms. The danger is that “working appropriately” turns into outlined by uptime metrics and price optimization relatively than person expertise.
