Job interviews aren’t just about resumes anymore. 

Corporate America is at a point where employees no longer need to hide AI usage.

In a growing number of white-collar jobs, AI is becoming a necessity the same way spreadsheets and word processors once did—workers are expected to know it and expected to use it.

Look no further than consulting titan McKinsey, which recently began trials for new hires that require them to use the company’s AI system called Lilli to help solve a case study.

Candidates in the pilot had to use Lilli to analyze a case, refine conclusions, and show they could prompt effectively and apply judgment to the output.

That’s a clear sign that the consulting firm is treating AI as part of the job. 

The point is, AI is changing the way companies view their workforce. It’s no longer just about how many people are on payroll. 

The next frontier of how to measure AI usage may now be based on agent count. Agents are as close to an employee replacement corporations currently have. 

They aren’t fully autonomous, but they are designed to complete a specific task or set of tasks with minimal human input. That distinction matters because “AI usage” is a fuzzy metric. 

A company can claim everyone has access to a chatbot and still get very little productivity out of it. Even McKinsey makes this point in its work on agentic AI. 

Early deployments can lift individual production, yet broad rollouts often fade after the first wave of power users unless leadership embeds the tools into company culture and expectations. 

Agents change the equation because they turn AI from a feature into a process. 

A lot of times, they can take the place of what a junior employee at a company like a consulting firm would do today. It’s not a one-for-one comparison—one agent does not replace one employee. 

Rather, each agent can do one employee task. Something like compiling data, running a type of analysis, or creating a style of document. 

When enough of those tasks are strung together under human supervision, the output starts to look like a team.

That’s why McKinsey has created 25,000 AI Agents over the past two years and why it hopes to grow to at least one agent per human employee in the coming months. The firm’s CEO Bob Sternfels described a “workforce” that includes tens of thousands of agents on top of roughly 40,000 staff, with the explicit goal of reaching one agent per human in the near term.

This is likely to be a new measure for successful businesses. 

Rather than focusing purely on human headcount growth, success will be measured based on total workforce. McKinsey for instance has a “workforce” of over 60,000 between 40,000 employees and 25,000 agents. Its goal, both internally and for its clients, is to achieve at least one agent per employee over the next two to three years.

Once staffing a company is looked at this way, a few old assumptions will have to change. 

First, headcount cuts stop being an automatic red flag. 

McKinsey shrank its human workforce after peaking around 45,000, and it has discussed further reductions tied to AI-driven efficiency. If agent capacity rises at the same time, total output doesn’t have to fall. 

Second, organizational “pyramids” get flatter. Traditional consulting models (and many other businesses) rely on armies of junior employees feeding a smaller group of senior decision makers. 

If agents can absorb the junior layer’s task load, companies can operate with a more evenly distributed workforce across all levels of seniority. 

Third, the competitive gap widens. McKinsey’s own research argues that modernizing processes with agents can speed up work by 40% to 50% and reduce costs by more than 40% in certain settings. 

McKinsey’s new vision for the workforce is what investors have to watch for in the future.

Not many companies are speaking in these terms today, but this will be a way to measure success in the near future. 

This doesn’t imply that headcount doesn’t matter, but it’s not enough information to understand if companies are succeeding or failing. The firms that keep reporting “AI initiatives” without any further details likely aren’t using AI effectively.

Meanwhile, the companies that expand their workforces through a combination of human employees and agents will start pulling ahead of their peers.

Best regards,

Joel Litman & Rob Spivey
Chief Investment Officer &
Director of Research
at Valens Research

Today’s analysis highlights the same insights we share with our FA Alpha Members. If you want to an get in-depth analysis of market trends and uncover undervalued stocks, become an FA Alpha Member today.