PwC Executive Assistant Layoffs: What AI Actually Cut, and Who Got Called Back

You probably saw the headline in May. It landed in an EA group chat, then a LinkedIn thread, then your own head at 11pm: assistants earning six figures at a Big Four firm, gone because of AI. The PwC executive assistant layoffs became the story everyone forwarded, usually with one question attached.

Am I next?

I want to give you a real answer. Not the LinkedIn-optimism version. A real one, with the facts straightened out first, because some of what got shared wasn't quite right.

What happened with the PwC executive assistant layoffs?

PwC's US firm laid off about 600 executive assistants, recruiters, and other support staff in February 2026. Bloomberg reported it on May 21, 2026, framing it as part of a wider shift at consulting firms toward AI tools for scheduling, expenses, and travel. Bloomberg also noted that assistants at these firms can earn more than $100,000 a year with incentives and bonuses.

Two details matter. First, the 600 wasn't only EAs; recruiters and other support roles were in the same number. Second, the work named was specific: calendars, expense reports, travel. Not the whole job. One slice of it.

Did McKinsey cut executive assistants too?

Not in the way many posts claimed. The widely shared "McKinsey cut 200 more" figure comes from a separate story. In November 2025, Bloomberg reported that McKinsey was cutting about 200 technology-focused support roles, and global managing partner Bob Sternfels said non-client roles were under review for possible reductions over the next two years (summary of the Bloomberg report).

So yes, support functions at big firms are being squeezed. But "200 EAs at McKinsey" isn't what was reported, and I'd rather you worry about the right thing than the loudest thing.

The part nobody's talking about: the callback

Here's what didn't make most of the headlines. In March, the St Pete Catalyst reported that an EA recruiter had heard from several former PwC assistants who said the firm had reached out asking whether they'd be interested in coming back. That was roughly a month after the layoffs.

I can't tell you how many were called or how many said yes. Nobody has published that. But the fact that the calls happened at all says something.

AI took the tasks. It couldn't take the context.

My read, from years of doing this job: tools can book the meeting, file the expense, and draft the reply. They can't tell that a particular email should have been a phone call. They don't know which client needs to hear from the partner directly, or remember why that vendor gets handled carefully. That knowledge sits in people, and you usually find out how much of it there was after the people are gone.

Why rules-based EA work was cut first

The roles most exposed to AI are the ones where most of the week is repeatable, rules-based work. Diary management, expense reconciliation, travel booking, and routine scheduling all follow patterns, and patterns are exactly what AI tools handle well. That's not a knock on the people who did that work. It kept firms running for decades. It's just the slice that's easiest to hand to software.

Sound familiar? It should. Your exec almost certainly has access to at least one AI assistant already, and has probably asked it to do something that used to come to you.

The useful move isn't to compete with the tool on speed. It's to sit above it.

Sit above the AI, not below it

Sitting below the AI means doing the leftover bits it can't finish, on its timeline. Sitting above it means you design the process, decide what the tool is allowed to touch, and check what comes out before it reaches anyone who matters.

In my own work as EA to the founder and CEO of an Inc. 5000 SaaS company, that looks like this: AI drafts, triages, and researches. Leadership minutes still go out the same day, and I'm the one who decides what's in them. Access, money, and people decisions never go to AI. The tools make me faster. They don't make the calls.

Done well, this makes you the person your exec trusts to run the tools. Done badly, it makes you a proofreader for a machine. The difference is whether you own the system or just clean up after it.

Three things worth doing this week, not next quarter

1. Audit your week honestly. List what you spent time on and mark anything a decent prompt and a short workflow could handle. My rule of thumb: if that's more than half your week, treat it as a signal. Not a verdict, a signal.
2. Pick one tool and go deep. Not seven tools at surface level. One, properly. Learn where it breaks. The most valuable thing you can know about an AI tool is exactly what it gets wrong.
3. Make your judgment visible. Start naming the catches. "I flagged this because..." and "I changed this because..." in a quick note to your exec. Calmly, not performatively. It documents the part of your job no tool is doing.

The six-figure EA roles built mostly on scheduling and expenses are shrinking. That's real, and I'm not going to soften it. But those roles aren't going because EA work stopped mattering. They're going because the definition of doing it well changed faster than some job descriptions did.

FAQ

How many executive assistants did PwC lay off?

Bloomberg reported that PwC's US firm laid off about 600 executive assistants, recruiters, and other support staff in February 2026. The figure covers several support roles, not only executive assistants.

Did PwC rehire executive assistants after the layoffs?

A March 2026 St Pete Catalyst report cited an EA recruiter who said several former PwC assistants had been contacted about returning. PwC has not published how many were approached or rehired.

Did McKinsey lay off executive assistants because of AI?

Bloomberg reported in November 2025 that McKinsey cut about 200 technology-focused support roles and was reviewing other non-client roles over the next two years. The commonly shared claim of 200 executive assistants cut at McKinsey does not match that reporting.

Which executive assistant jobs are most at risk from AI?

Roles where most of the week is rules-based work, like scheduling, expense reconciliation, and travel booking, are most exposed. Roles centered on judgment, discretion, stakeholder relationships, and running AI workflows are much harder to replace.

The headline was about the jobs that went. The lesson is in the calls that came after.

Clarisse