There is a steady drumbeat right now of predictions that AI is about to wipe out huge swaths of knowledge work. Entire finance and operations teams cleared out; analysts, FP&A managers, controllers, and the people who own the monthly close all replaced by a model that does the job faster and cheaper.
I’m not fully convinced.
If these jobs were really just about economic efficiency, a lot of them could have been automated, outsourced, or converted to contract work years ago. Offshoring has been around for decades. So has workflow software, RPA, and every other wave of tooling that was supposed to finally streamline the back office.
And yet the org charts kept growing. Whatever was supposed to happen in all of those previous efficiency cycles mostly did not happen, and that is worth considering before we assume this wave will somehow be different.
Why efficiency has never been enough
The best explanation I have read for this is the late David Graeber’s Bullshit Jobs. His central claim is that a surprising share of modern work exists not because it is economically necessary, but because it is politically or socially useful inside the organization. Layers of reporting, box-checking, status-signaling, and meetings about meetings serve a purpose. Just not the one written in the job description.
Some of it is about signaling seriousness. A director with five direct reports looks more important than a director with one, regardless of whether those reports are producing something anyone actually uses. Some of it is about internal politics. A variance analysis gets written because a controller wants cover on a line item, or because a department head needs to justify a budget. Some of it is about legibility. A team’s work is easier to defend in the next planning cycle if it generates artifacts, even if nobody opens them again.
None of this really concerns economic output in the most direct way. And that is exactly why efficiency alone has never cleared these roles out, and why I suspect AI alone will not either. You cannot automate away something whose real purpose is to make a manager feel in control, or to give a department something to show at the quarterly review. The performative part of the work sticks around because it is doing something other than what its description says.
Why this matters to me as a trainer
I want to be honest that I find the “AI will just wipe them out” framing frustrating, because it skips over what actually matters.
Most of the FP&A, accounting, and operations people I meet are sharp. They are curious, capable, and usually pretty clear-eyed about what is going on around them. They know which reports nobody reads. They know which meetings could be a three-line email. They are often the first to admit when a process is more ceremony than substance.
Getting good at Excel, Power Query, and modern analytics, to me, is about giving those people and their teams a little more agency inside their roles. The skill is leverage, in the sense that one sharp analyst with the right tooling can do in an afternoon what a less-equipped team would take a week to produce. It is also, to be candid, dignity. It means being the person in the room who can actually answer the question the CFO is asking, instead of the person reformatting the deck for the fourth time.
The exit ramp, if you want one
This is where I actually get optimistic, and it is not the “AI is going to fix everything” kind of optimism.
The hopeful part is narrower than that. AI is giving individuals a real exit ramp when they want one. If your current role has been quietly draining you, the tools now exist to build something of your own, sharpen a craft, or move into work where your output actually matters. An FP&A analyst who has been thinking about fractional CFO work can spin up a presence in a weekend. A controller who has wanted to teach the next generation of finance pros can build and deliver a course without hiring a team. You can learn a serious technical skill, in depth, at a pace and price that would have been out of reach a decade back.
None of that is automatic. It takes initiative, patience, and some willingness to look foolish for a while. But the path exists now in a way it did not before, and that seems genuinely good to me.
The companies that get it have their chance too
The exit ramp framing is about individuals, but the same moment is open to organizations.
The companies that come out of this window well are not the ones that treat AI as an excuse to cut their finance and ops headcount in half. They are the ones that take their FP&A, accounting, and operations teams seriously as analytical partners and give them the tools, the time, and the training to actually operate at that level.
When a mid-market company invests in getting its finance team fluent in Power Query, Power BI, and AI-assisted workflows, the return is not just a faster monthly close. It is a team that can investigate an anomaly the week it happens instead of the quarter after. It is forecasts that actually inform decisions. It is the ability to say yes to analyses the business has wanted for years but nobody had the bandwidth to produce. It is a controller or FP&A lead who has space to think again, instead of spending half the month stitching spreadsheets together.
That is the version of this I am most hopeful about. It looks like a company deciding that the people closest to the numbers should have the sharpest tools, and then actually equipping them that way.
The part where you actually have to do something
So I am skeptical of the mass-displacement story, optimistic about the exit ramps, and genuinely excited for the individuals and companies who decide to use this moment on purpose.
“AI gives you leverage” is not the same as “AI will save you.” The leverage is real, and it is also something you have to use deliberately. For an individual, that means getting serious about the tools that are already on your desk and learning to use AI in a way that makes your judgment sharper rather than thinner. For an organization, it means investing in your finance and ops team’s capability instead of treating them as a cost center that somehow deserves fewer tools than the sales org gets.
If you have the agency and the wherewithal, there is a version of this where AI makes things better for your career and your organization, not worse. That version just asks you to think a little differently about what your finance and ops people are actually for.
If that sounds like the shift you want to make, whether as an individual or as a team, that is most of what I work on. You can read more about how I work and where to start.