An IBM exec built an AI agent to prep for meetings — and said it saved hours every week

• IBM's Dave McCann is using an AI agent to prepare for client meetings.
• He said "Digital Dave" helps him save five hours a week by eliminating 30-minute prep calls.
• Companies from tiny startups to the biggest firms are building AI agents to take on routine work.

Dave McCann oversees thousands of humans, and also an AI agent he named after himself: Digital Dave.

One of the most valuable things it does is conduct research, including on the company's customers. That's a big help for McCann, who, as a global managing partner for transformation at IBM Consulting, is responsible for thousands of clients, including Nestlé, Ericsson, and Riyadh Air.

The agent — it's actually a collection of AI agents and assistants — scans McCann's calendar for client meetings and drafts a list of 10 things he needs to know for each one. The goal, McCann told Business Insider, was to free up time he and his staff spent preparing for the meetings.

Under the old setup, people on his team would put together a briefing document with him and typically have a 30-minute prep call ahead of the client meeting.

"All that's now gone," said McCann, who is tasked with helping transform the global IBM Consulting business, which has nearly 150,000 employees.

He generally talks with about 10 clients a week, so the agent saves him roughly five hours of prep time, McCann said.

"All the time I used to invest in client prep, I can now see more clients," he said.

Freeing up the team

McCann's research agent, which he and his team began building this past fall, is based on a tool that a group at the company had started to develop as part of an annual internal competition.

The agent reviews in-house data, what IBM and the client are doing in the market, external data, and account details — such as project status and services sold and purchased, McCann said. It can also identify industry trends and client needs by, for example, reviewing a firm's annual report and identifying a corresponding service IBM could provide.

Digital Dave also saves McCann's team time, he said, because the three or four staffers who used to spend hours pulling together insights for the prep calls are now free to do other work.

"It's not just about driving efficiencies, but it's really about transforming how work gets done," McCann said.

The agent's research abilities aren't limited to client reports. McCann has also begun using it to help him assess the hundreds of IBM Consulting partners he evaluates each year. The goal, he said, is making informed decisions and giving the execs good advice about their strengths and weaknesses as part of their performance evaluation.

McCann said the data the agent reviews can include the services execs sold and the profits they generated, the training they provided, and the impact they had on their teams' development. Before the agent, he said, his approach involved perhaps a dozen spreadsheets culminating in "the worst pivot table of your life."

Now, McCann said, all of the data goes into a model, and he can ask pointed questions and make queries about top performers in a particular parameter.

'That multiplier effect'

One benefit of building agents, McCann said, is that IBMers who develop them can share them with others on their team or more broadly within the company, "so it immediately creates that multiplier effect."

Many of the people who report to him have created agents, he said. There's a healthy competition, McCann said, to engineer the most robust digital sidekicks, especially because workers can build off of what their colleagues created.

Across industries, companies are developing AI agents to take on knowledge work — especially tedious tasks — once handled by humans. From one-person startups to consulting giants, firms are using agents across functions such as HR, IT, finance, communications, and training.

Agents can handle a range of functions, including gathering information, processing paperwork, drafting communications, taking meeting minutes, and pulling research. It's still early, but these systems are quickly becoming a major focus of corporate AI efforts as companies look to turn generative AI into something that can actually take work off employees' plates.

One challenge McCann sees with clients and building agents is having access to data. IBM Consulting's most advanced clients — maybe a Fortune 100 or Fortune 500 company — might have given access to data to several hundred people, but not to 5,000 people in their finance division or 10,000 people in HR, McCann said. Concerns about security and how to manage the innovation can be roadblocks.

Until you unlock the data, individual workers might be able to get more done, but "you don't get that multiplier effect of the productivity," he said.

For McCann's work, Digital Dave means that he gets critical time back on his calendar.

"I can have much more focused attention out with our clients and with our humans while I have the operations of the business now being run more digitally," he said.