While some will call AI a fad, real value can be derived from it. In unified communications, AI takes different...
shapes and sizes. One area where AI is taking shape is in video meeting rooms where it's all about productivity and experimentation.
Meeting room technology has oscillated between high-end dedicated hardware to software-based products. In this market cycle, it seems like we're back to hardware-based video meeting rooms. The nuance today is that these are smaller meeting rooms -- often called huddle rooms -- and vendors are using devices like Intel NUC and Google Chromebox.
The main focus of the newest meeting room technology is on collaboration and productivity. People meet each other virtually in huddle rooms to do actual work together, not just to make decisions. As we move toward that focus, AI services are focused on four areas to improve productivity in these virtual spaces.
1. Voice analytics
Voice analytics is about AI looking at a conversation taking place in the meeting room and understanding it. The underlying meeting room technology here is speech-to-text. Most video conferencing vendors today are working toward introducing automatic meeting transcription capabilities. Transcription capabilities are used in two ways:
- Transcription through speaker diarization, which takes meeting audio and separates individual speakers, making it easier to understand who said what.
- Transcription that allows users to pick a specific sentence in a meeting and have the video transcription jump to that sentence.
Additional capabilities vendors have coming down the pipeline include the ability to summarize meetings, take action items and identify key points in the meeting. While the focus for these capabilities is video, the value still lies heavily in voice and what is said in the meeting.
2. Voice bots
Think Amazon Alexa for meeting room technology. Video conferencing room systems already have a microphone as part of their product, which can double as virtual assistants of sorts.
Many vendors are investing in adding voice bot capabilities into their products. As a first step, this means adding a wake word that the video conferencing system actively waits for. This is the enterprise equivalent of the "Alexa" and "Hey Google" commands. Wake words are important as they are processed locally without sending continuous voice streams to cloud services.
Next, real-time transcription and understanding captures what is spoken after the wake word with the purpose of identifying the intent of the user. Creating accurate intent is time- and resource-consuming, which is why voice bots in video meeting rooms are focused on replacing the remote control and simplifying the user experience.
These capabilities aren't meant to compete with Amazon Alexa or Google Home. They are there to provide a better remote control.
3. Computer vision
Computer vision, also known as machine vision, is a nascent area of AI that uses cameras to allow computers to "see." While most computer vision talent is focused on autonomous cars, there are use cases for video meeting rooms.
Computer vision for video meetings focuses on the ability to identify people in the room. Identification includes understanding where people are in the captured video and applying automatic zoom or counting participants. Knowing how many people are sitting in meetings and in which rooms can give some valuable insights as to the needs of the company, such as whether more meeting rooms are needed or what size rooms are required.
Some vendors are looking to expand computer vision capabilities to include gesture control as a remote control replacement and the ability to identify and focus on a whiteboard automatically.
4. Quality improvements
AI services can play a role in improving the quality of media in a meeting as well. Where we see this most in video meetings is by providing better noise suppression techniques within built-in speakers. Noise suppression is important when dealing with modern meeting rooms that have a lot of glass walls or when trying to join meetings from open spaces or public areas.
While machine learning can be employed to improve media quality, at the moment it seems vendors are not investing as much in machine learning in these areas.
Remote work, cross-company collaboration and distributed teams all increase the need for video meetings. These evolving working scenarios are changing the focus of meeting room technology. Productivity is key and AI is taking an active role in improving productivity.