Improving meetings with visual AI

Improving meetings with visual AI

AI-driven video solutions help to deliver high-quality and inclusive meeting experiences for both in-room and remote participants 

Recently, the terms ‘hybrid work’ and ‘AI’ have received considerable attention. Crestron’s Joel Mulpeter raises the question: can the latter help the former?  

Guest contributor |


You have likely heard quite a bit about virtual assistants and the ability of artificial intelligence to summarise a meeting and report on everything from attendees’ names to the mood of the session. What is and is not allowed in that regard can vary between jurisdictions, but there are other aspects of AI that aren’t nearly as contentious. 

Hybrid working has revealed several problems, many of which are technological. Tools used by in-office workers to collaborate with remote workers often provide poorly framed video, which makes it difficult to see and hear participants. All of these issues affected how engaged remote workers were with the meeting. For example, as the clarity of an image diminishes, the gestures, expressions and all the other nonverbal cues critical to maintaining engagement become compromised. And, when remote meeting attendees cannot see their colleagues properly, meeting fatigue can set in and negatively impact productivity.  

Intelligent video systems offer an elegant solution, dramatically improving the experience for the remote viewer while allowing in-person participants to conduct a meeting or contribute without worrying about disappearing from the screen. Intelligent video systems automatically create a broadcast-quality video, eliminating the need for manual video switching or camera operation. Driven by visual AI, intelligent video follows people when they are presenting and frames groups to minimise dead space in the shot. It also automatically reframes as people move around in seats and tracks speakers to follow the natural conversation around the room. 

In-room attendees can usually see their remote counterparts’ faces clearly. However, it can be difficult for virtual attendees to see the faces of in-room attendees, especially if the room is large. A major advancement is an AI-driven intelligent video function called group framing, which keeps everyone in the frame of the shot being transmitted to remote viewers. Such systems will adjust as people enter, exit or move around the room to ensure that everyone stays in the shot.  

Presenter tracking effectively keeps moving presenters in the frame. It detects when people enter a presentation area, frames them and continues to track them in real time. It is ideal for rooms with presentation areas like lecture halls, training venues and large conference rooms where a presenter needs to be in motion.  

Meanwhile, speaker tracking switches the view as the conversation moves from person to person in the space. Multi-camera speaker tracking solutions are superior to single-camera options, as primary and secondary cameras allow for clean cuts between speakers. The first camera remains on the active speaker, while the second moves to the next speaker’s position, allowing the system to switch seamlessly. By eliminating panning, tilting and zooming, the system provides a higher-quality viewing experience for the remote participant. As a conversation moves around the room, the system follows the natural flow of the meeting.  

Systems such as those offered by Crestron provide an additional benefit: they are so intuitive that users never need to concern themselves with the underlying technology. They can focus all their attention on presenting, collaborating, conversing and achieving results.  

Joel Mulpeter

Joel Mulpeter is director of product marketing at Crestron 

Discover more insights like this in the Autumn 2024 issue of Technology Record. Don’t miss out – subscribe for free today and get future issues delivered straight to your inbox.      

Subscribe to the Technology Record newsletter


  • ©2024 Tudor Rose. All Rights Reserved. Technology Record is published by Tudor Rose with the support and guidance of Microsoft.