How AI can make meetings smarter

If you watched Iron Man, you may remember JARVIS — Tony Stark’s intelligent assistant. Machines as smart as that remain in the realm of science fiction, but I’ve identified several iOS-compatible tools that can put some real-world artificial intelligence (AI) inside your next meeting.

Meeting AI

By rights, the pandemic should have inspired a rapid evolution in “Meeting AI.” Cisco sees the space as a five-stage process, with voice control morphing to semantic understanding and predictive intelligence as the sector evolves.

Think of it this way: you might be in a hybrid meeting, featuring a mix of in-person and remote speakers. How might AI augment that environment?

Here are a few ideas of how this might work, some of the solutions already available, and a few ideas as to how this technology may evolve.

‘Hey Siri… Plus’

First up, voice control. We’re kind of already there on that one. It’s already possible to use Siri to schedule meetings, invite others to meetings, send emails and more. You can also build Siri shortcuts to launch and join Zoom or WebEx meetings and more.

While this is relatively useful, it’s not precisely cutting-edge stuff — and I don’t think it really taxes the Neural Engine Apple has put inside its chips, though it is good to know that in iOS 15 the request should be handled entirely on the device.

Of course, meetings these days are best set up using Doodle, to establish availability.

Everything data

Some products help turn a meeting itself into data. One interesting example is Calendar, which tracks your meetings, spots attendance records, and tries to give you valuable insights, such as how long you spent in meetings in the last few weeks.

After a meeting, AI can help summarize the key topics, deliver schedules or work plans, and more. I make a lot of use of Otter.ai for this, as I find the transcripts it delivers are becoming increasingly accurate and its ability to distinguish between different speakers useful.

Otter.ai can also identify topics discussed during a meeting, generating around ten keywords that it thinks are key to the session. You can then search through the transcript for mentions of those keywords, or any other word in the search tool.

Video conferencing tool Headroom has some of this with the added ability to figure out and share action points.

The big players in the space are undoubtably exploring such solutions, too. Webex Assistant can listen for trigger words such as “take a note” and act on them, highlight comments in transcripts, and more.

Smart meeting support tools

Another solution, Adam.ai, provides a set of tools to support meetings, making them more efficient. Use this to upload files and share them across your teams using multiple collaboration apps, to create agendas, record votes, and to track actions decided on during a meeting.

That last feature is useful, as the app will recognize which actions you are committed to and will also attempt to identify deadlines for those actions, reminding you where you are with those tasks.

Another tool that seems worth a look, Klaxoon, combines project management with in-meeting and post-meeting efficiency.

Tools to augment the digital meeting experiences are also proliferating — you’ll also find plenty of these on the Zoom App Marketplace. Hendrix, Avoma, or Fireflies provide a range of smart-meeting support services designed to integrate with widely used video collaboration tools.

Microsoft is already exploring here. Who else can recall the whiteboard feature it introduced in 2019 that uses machine vision intelligence to display the part of the board that is obscured by a person writing on it during a meeting?

Making meaning from meeting data

Otter really is useful, but the capacity to identify key moments in a meeting a stepping stone towards the next evolution in this space. This will be when AI becomes a kind of research assistant during the meeting, finding documents, data, and expertise from across the company to assist you.

In theory, it might work like this: You’re in a meeting discussing the available budget for your local park and the different needs users have. An intelligent assistant might be able to source this information for you and present it as a chart that shows one particular use is hoovering up all the funding in that space. Now the team knows where it should adjust spending to meet a wider set of needs.

Technologies to identify key topics and moments of high engagement continue to emerge as businesses seek to mine, not miss, such “meeting gold.”

Emotional and contextual understanding

The next evolution will be smarter. Cyrano.AI is already developing tools to analyze a person’s conversation in order to provide insights into how they are motivated, what their values are, and how they communicate.

Aimed at sales and tech support people, the company thinks its tools will enhance communication and help reach deals. This kind of emotion recognition is an emerging field in the meeting AI space.

The kind of smart contextual understanding heralded by Otter.AI’s capacity to identify keywords from within a meeting is a first step toward the creation of post-meeting bots. These will use those keywords to deliver each transcript, meeting minutes, and action items to every attendee, but will also be able to proactively seek all the supporting data, documents, physical, and human resources attendees might need to finish tasks they signed up for during the meeting.

The inevitable direction of travel, of course, is toward Robotic Process Automation, in which intelligent meeting assistants will track all of the above while adding the capacity to analyze and identify the decision flow, manage business administration and more. Though we’ve some way to go before we become Tony Stark with a helpful robot in every home office.

Please follow me on Twitter, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.