Using an AI agent to attend meetings at work? Here's what you need to know first

Data from McKinsey shows 65% of U.S. companies are experimenting with AI agents on the job, with more companies tolerating and even encouraging staffers to send an AI agent proxy to a digital meeting on their behalf.

It may not work for meeting presenters (yet), but AI agents can reduce team meeting time and let staffers avoid meetings and focus on what counts. As usual in the early artificial intelligence workplace rollout phase, there are upsides and downsides to leveraging an AI agent as a company meeting proxy, and anyone mulling the practice should know both sides of the issue.

Here are the factors that matter most with AI agent meetings, for executives and staffers alike.

Time saved is the biggest issue

According to Microsoft $MSFT, since February of 2020, users have increased their attendance at Teams meetings by 300% compared to previous years. “Data shows 35% of the time spent in meetings is wasted due to poor planning, vague meeting objectives, and inviting people who do not belong in the meeting,” said Kishore Bitra, collaboration engineering team lead at the City of Baltimore (MD). “Longer duration meetings result in even greater amounts of wasted time.”

Consequently, a big upside of AI meeting proxies is efficiency. “I love how agents can reliably capture key points and follow-ups, which of course will help free time in people's schedules,” said Chris Sorensen, CEO of PhoneBurner and ArmorHQ, two B2B SaaS sales platforms. “They also make information easier to share across teams.”

Guardrails should be set

Sorensen, who allows AI agents to be used at his companies, said he’s “not a fan” of his team becoming too reliant on the practice. “I don’t want to see it all the time, but broadly I do not have issues,” he noted. “I find AI agents work well when 3-4 members of a team don’t sit in on a call.” Sorensen said maybe one team member can be on hand physically and the AI agent can sit in and capture notes, summarize decisions and takeaways, as well as track key action items for the others.”

Sorensen said issues can arise when they quietly replace human presence in discussions “that require judgment, accountability, or relationship-building,”

Upfront transparency is a must

Dr. Shawn DuBravac, CEO and president of the Avrio Institute, an advisory firm that provides digital transformation consulting and research services to C-suite executives, said he personally uses AI agents at organization meetings.

“I use them in the form of notetakers in all virtual meetings, including both ones I attend and ones I do not attend,” he said. “Using AI agents as meeting stand-ins is not inherently unethical, but it does require intent.”

DuBravac said he’s okay with company staffers using an AI agent to listen and record, summarize, organize action items, and generally help coordinate efforts. “That is augmentation, not deception,” he noted. “Problems can definitely arise if AI is used to create the illusion of presence or participation without disclosure.”

Establish usage protocols

Bitra advises company decision-makers to identify approved types of meetings that can involve AI agents. “That could be with regular status updates and information briefings,” he said. “Meetings that humans should conduct include performance reviews, disciplinary actions, legal matters, high-risk strategy discussions, and sensitive union or HR topics.”

Another green light is to motivate individuals to use AI agents to reduce attendance at more mundane meetings, “but not to skip important discussions or relationship-building opportunities that need human interaction and understanding,” Bitra advised.

Additionally, turn off consumer bots that automatically join meetings unless they have been specifically authorized and set up in accordance with governance guidelines. “IT administrators should also block the automatic usage of any bot available in the market; it has to be verified before making it available for the end users,” he noted.

Companies should create a clear policy on use and also socialize the cultural norms around corporate use. “Companies can also start with low-stakes meetings, like routine updates or recurring status calls,” DuBravac said. “Companies should also ensure AI agents and notetakers are accomplishing what they want - better meetings and better/quicker decisions.”

Weigh the key factors when deciding on AI agents at your company meetings

Management should address the issues surrounding AI agents at company meetings and clear the air so everyone understands the pros and cons of moving forward with AI meeting proxies.

“There are many upsides of using AI agents,” DuBravac noted. “AI agents can absorb information, generate clean summaries, extract decisions, and track commitments, allowing us to spend more time on judgment, creativity, and execution.”

They can also reduce meeting fatigue, improve documentation, and create a searchable institutional memory that most organizations lack. “I have found this later part to be very beneficial with my firm,” he said.

There are also potential downsides.

“AI can miss nuances that humans pick up on, and we’ve found that AI agents are good at understanding what was said, but not why it was said,” DuBravac added. “AI agents, at least now, miss the political dynamics. Overreliance can also weaken team relationships if people feel team members or leaders are absent.”