Consulting firms have built thousands of AI agents. Now they're trying to figure out their worth.

• McKinsey, PwC, EY, and BCG have raced to adopt AI. They've launched thousands of AI agents.
• Now those consulting firms are trying to understand the technology's true value.
• Firms are measuring how much time AI agents save, and how that saved time is used.

Big questions are swirling around AI's real impact — and consultants are racing to supply the answers.

Over the past year, consulting firms have begun deploying armies of AI agents as they work to transform their own operations and advise clients to do the same — automating research, building task-specific tools, and building proprietary AI models.

McKinsey & Company CEO Bob Sternfels said last month that his firm has launched tens of thousands of internal AI agents in recent years, and eventually plans to have one for all of the company's 40,000 employees.

Amid the rapid rollout, consultants are now asking themselves a tough question: Is it worth it? They are working to measure if AI is truly improving performance, boosting revenue, and freeing consultants to focus on higher-value work.

"I think we are now in the age of confusion," Mina Alaghband, a former McKinsey partner, now the chief customer officer at Writer, a full-stack enterprise AI platform built for agentic AI, told Business Insider.

Alaghband said that a year ago, most companies were focused on adoption, tracking metrics such as how often a tool was used.

Now, she says said the emphasis should be on measuring the value that's created — like the amount of human labor reassigned to higher-value work, or improvements in revenue.

PwC's chief AI officer, Dan Priest, recently told Business Insider that PwC is now less concerned with how many agents it deploys, and more with how many human users each agent has.

Priest said his firm starts by targeting an "impact zone," such as improving the customer experience.

Within these impact zones, the firm looks to deploy "specialized AI agents" that have earned that designation because they're good at what they do, Priest said. "When we deploy agents, we want to see a high rate of human adoption, which means more humans are using them," he said.

EY also prioritizes quality over quantity, Steve Newman, EY's global engineering chief, told Business Insider. The firm tracks the value created by its AI agents through key performance indicators for productivity, quality, and cost efficiency on a month-to-month basis.

If the defining promises of the AI boom are speed and efficiency, then the metric that may matter more isn’t usage, but time reclaimed.

Boston Consulting Group tracks its agents by that metric — and whether that time is then reinvested in higher-value work, Scott Wilder, a partner and managing director based in Dallas, told Business Insider.

Wilder said humans at the firm now spend about 15% less time on low-value activities, like making slideshows, and that those people are reinvesting about 70% of their saved time into higher-value activities, such as deeper analysis.

Time saved doesn't always mean more work. At BCG, it can mean more free time. Wilder said BCG has found that employees keep about 30% of the time AI saves. "They get a little more sleep or get to go to a yoga class or whatever someone wants to do," he said.

Nearly a century ago, economist John Keynes predicted that as productivity rose, the balance between work and leisure would inevitably change.

"I would predict that the standard of life in progressive countries one hundred years hence will be between four and eight times as high as it is," he wrote in his 1930 essay "Economic Possibilities for our Grandchildren."

It's almost 2030, but in small ways, that vision may already be surfacing.

"It's benefiting them — and this is a tough job, so every hour of free time matters," Wilder said.

Something to share about how consultants are using AI? Business Insider would like to hear from you. Email Lakshmi Varanasi at lvaranasi@businessinsider.com or contact her on Signal at lvaranasi.70.