5 use cases for AI in the workplace

Artificial intelligence (AI) has come a long way in the past few years.  What was once something only witnessed in science fiction has now become very real with AIs playing poker, telling us when to leave for the airport and letting us know what the weather will be like tomorrow.  In the business world, AI has been used to improve cybersecurity and help contact center agents be smarter but it has yet to do is make workers more productive in any significant way but that will change soon.  

This week in San Francisco, collaboration market leader, Cisco, held an Advanced Technology Fair to look at what happens when AI and collaboration are brought together.  Instead of issuing a press release, Cisco chose to show what was possible through a series of demos.  I was fortunate to be at the event and wanted to share what happens when AI is applied to meetings. 

[ Cut through the hype with our practical guide to machine learning in business and find out whether your organization is truly ready for taking on artificial intelligence projects. | Get an inside look at 5 machine learning success stories. | Get the latest insights with our CIO Daily newsletter. ]

Cisco demonstrated that there are multiple ways AI can improve meetings.  The innovation is badly needed as we spend more time in meetings and the experience is getting worse.  It seems every meeting I attend someone is looking for a dongle, cable, trying to find a dial up number or can’t find the latest version of some PowerPoint presentation. My research has found that on average, 15 mins of every meeting is wasted just getting stuff set up and AI can help bring the 15 mins to zero and I’ll explain through the following use cases that Cisco presented at its Advanced Technology Fair. 

1. Voice enabled assistants. Everyone is familiar with a voice assistance from using Siri, Alexa or any number of other consumer agents. Cisco’s Spark assistant works very much like its consumer counter parts but with a business twist. The agent is passive until the keywords “Hey Spark” and then can be used to execute different meeting related tasks.  For example:

>

• Hey Spark, join the meeting – logs the person into a scheduled meeting
• Hey Spark, join my personal meeting room – starts an ad hoc meeting by understanding who is in the room from the proximity pairing between the user and a Spark Board
• Hey Spark, place a call to Zeus Kerravala – calls Zeus Kerravala
• Hey Spark, call Jonathan – Spark would then show a filtered corporate directory of all the Jonathans and then the user can pick the correct person
• Hey Spark, record the meeting / stop recording – starts and stops recording of the meeting

There’s really no limit of the meeting related tasks it the voice assistant can do. The key for a business focused virtual assistant is having the right data set. Consumer assistants need to have a very wide base of knowledge but not very deep so they can answer basic questions about the weather or traffic.  A business virtual assistant would need to have a very narrow but deep data set about employee’s skills, directory information, where data is stored and other information.  Ideally the knowledge base would be a combination of vendor provided data and company institutional knowledge.

2. Improving meeting ergonomics. Video meetings are supposed to improve meeting dynamics because people can see each other but often fails to do so.  Some video rooms are set up where everyone sits along the sides of a long table, which provides no one the ability to have eye contact or read facial gestures. There are also cameras that offer far end remote control so a video participant can pan and tilt the camera to zoom in on the speaker.  The challenge with this is that the act of controlling the camera takes so much concentration that the person doing it can’t really pay attention.  With an AI, it knows when people enter a room and start speaking and can constantly reframe the camera so everyone has a great view of everyone else. 

Cisco also demonstrated how AI can be used to improve the audio portion of a meeting.  One of the common annoyances of audio calls is that it seems there’s always that loud typist whose clickity clack of the keyboard drowns out everyone else.  Or that person working from home with the dog barking.  Cisco is using AI to detect common sounds like keyboards, dog barks and sirens and then automatically mute that line. 

3. Facial recognition. Ever been on a video call and someone is speaking but you don’t know who it is? Who hasn’t? An AI can be used to scan people’s faces and then superimpose their name over top of the screen.  Now no matter who is talking, you can quickly identify who the person is.  This sets up an interesting dynamic where, for large groups, the video experience is better than meeting in person as we don’t walk around with name tags on.  Over time, it’s reasonable to expect the AI to pull information from something like the employee directory or LinkedIn and show things like job title, purchase history or anything else that might be useful to the meeting.

4. Proactive listening. Most voice agents are passive until the keyword is spoken. This is to protect the scenario where every time someone utter the word weather, Alexa interrupts the conversation.  However, Cisco did show an interesting use case of an active listening agent.  It demonstrated something it is currently calling “Team TV” where a group of people are on a perpetual video meeting and going about their day.  Picture a screen in a “quad box” format that has three populated with people and a fourth for the AI.  If someone happens to say, “jet engine” the AI can pull up information and show it in the screen.  The AI is playing the role of an active participant and listening for meaningful cues and then contributing content to the meeting.  I suspect we are a few years away from this being mainstream but shifts virtual assistants from being something passive to active. 

5. Virtual reality. Cisco demonstrated a virtual reality (VR) collaborative session, similar to the experience I wrote about with NVIDIA’s Holodeck. With VR, all of the above use cases could be integrated into the session.  For example, if I’m interested in heart pumps, I can ask the AI to go learn about it and then place a virtual one in the middle of the room.  Then I and the people I am working with can then start to take it apart and learn about it without having to buy one.  

One of the sentiments regarding VR is that people won’t want to walk around with the goggles on all the time and I agree with that but VR’s use as a mainstream collaboration tool is use case driven. People won’t use it to work on PowerPoint presentations but they will to design buildings or cars or other things.

At the event, I had a chance to speak a few minutes to Rowan Trollope, Cisco’s SVP and GM of IoT and Applications and he shared with me a bit of vision of where the company is headed.  It’s his belief, and I tend to agree with him, that we will eventually live in a world where there are microphones, computers and cameras embedded into everything, a term he called “ambient computing”.  As we go about our day and have the need to speak to someone, find some information or any other task that takes time, the ambient AI will understand the need and push data to us, connect us over video or complete some other task that helps us do our jobs better. 

The use cases of AI in the workplace are limitless. It requires an understanding of business process and all the things that makes us inefficient today and then opening one’s mind to how the use of machines can help us do things better. I believe we sit on the cusp of something big and about ten years from now we won’t even think about BOTs, virtual agents and AI as they will be embedded into everything we do. Cisco gave us a glimpse of how AI can be used to improve meetings and collaboration but that’s just the start.

More on AI and machine learning

• 9 IT projects primed for machine learning
• 10 strategic tips for getting started with machine learning
• Which deep learning network is best for you?
• How to build a highly effective AI team
• Why you should invest in AI talent now
• Why AI careers can start with a degree in linguistics
• The year of Alexa and the coming decade of A.I.