Cisco Spark Assistant bringing voice commands to meeting hardware

Anyone who has used modern meeting software knows it’s still fraught with challenges trying to get everyone into the meeting, futzing with the hardware or software and smoothly integrating external documents like PowerPoint presentations. Cisco is trying to improve and simplify the meeting experience with voice commands, and today it introduced Cisco Spark Assistant, a voice controlled interface for Cisco Spark meeting hardware.

The artificial intelligence underlying the voice recognition technology is courtesy of MindMeld, a conversational AI company that Cisco purchased in May for $125 million. The idea is to reduce the effort required to manipulate the meeting hardware by introducing a limited set of meeting-related voice commands, says Timothy Tuttle, CTO of the Cognitive Collaboration Group, and former CEO and founder of MindMeld.

He says they are not trying to replace popular consumer voice assistants like the Amazon Echo or Google Home. Instead, they wanted to maintain a steady focus on the meeting room and the kinds of commands that make sense in that environment.

“Cisco acquired MindMeld in May of this year. We have taken and applied [our technology] to Spark Assistant. The initial focus is to use voice commands to streamline the things people do every single day in meetings involving joining and controlling device functions,” Tuttle explained.

That could be actions like entering your private meeting room, inviting people to your meeting, joining meetings or adjusting the volume. To get the system to listen, you use the trigger phrase, “Hey Spark,” and then the given command such as “Hey Spark, turn up the volume.” The team wanted to create a trigger command that you couldn’t use by accident.

Tuttle says that tuning MindMeld’s technology to work with Cisco hardware has been challenging, but by limiting the command set and taking advantage of the quality of the hardware setups, such as multiple microphones, has made it easier to pick up voices in a room and implement the voice recognition algorithms.

Over time, they hope to add other functionality to call up documents or analytics from various enterprise repositories using voice commands. Other future features could include a full automated meeting transcript (a feature which rival Zoom announced earlier this year).

The product is being announced this week, but won’t be available until some time early next year with additional commands being layered on over time.