AI Is Rewriting the Financial Analyst Job. Here's Exactly How.

The Analyst Who Bet on the Inside

In 2023, Alexander Vasylenko had the résumé and the interviews. What he didn't have was a job.

The Ukrainian-born financial analyst had worked for CIBC in Canada, done investment research on Seeking Alpha, and was grinding through equity valuation interviews in New York. "A lot of interviews, but no offers," he would later tell Business Insider. Then a recruiter messaged him on LinkedIn — not about a finance role, but about teaching AI models how finance works.

He was skeptical. He took the gig anyway.

"The workload in finance will be different in a couple of years," Vasylenko said, "and it's necessary to think about how you are going to fit into that."

That sentence is the whole argument. The question for every financial analyst reading this isn't whether AI is changing the job — it is, in specific and measurable ways. The question is whether you're positioned to understand the shift from the inside, or to discover it too late from the outside.

To understand what Vasylenko bet on — and whether that bet applies to you — you need to see exactly which parts of the analyst job AI is already doing, and which parts it genuinely cannot.

What AI Is Actually Doing to the Analyst's Workflow

Start with the part that should unsettle you.

OpenAI's Project Mercury quietly recruited over 100 former investment bankers from Goldman Sachs, JPMorgan, and Morgan Stanley — paying them $150 an hour to teach AI systems how to build LBO models, IPO structures, and pitchbooks. The stated goal was explicit: replace the hours of grunt work performed by junior bankers across the industry. Goldman Sachs CEO David Solomon has said AI can now draft 95% of an IPO prospectus in minutes.

That's not a projection. That's what's happening now, with real people, at real firms, for real money.

And yet the most honest response to that fact is also the most counterintuitive one. One former Morgan Stanley analyst who reviewed the Project Mercury work told Quartz: "Past a point, financial modeling is kind of rote and it's a waste of time for some of our brightest minds to be doing it if it's unnecessary." A second analyst offered the necessary counterweight: "A model doesn't do anything magic. What makes it valuable is the validity of the assumptions."

Both things are true simultaneously. The mechanics are automating. The judgment that makes the mechanics meaningful is not.

The technological progress over the past two years is just astonishing. It's a bit scary when you wonder how useful you'll be going forward.
by Alexander Vasylenko, Financial Analyst

The same pattern holds in research and documentation. BlackRock has deployed "Asimov," a virtual investment analyst, inside its fundamental equity business to synthesize research that previously took analysts weeks. UBS is using AI avatars of 36 of its analysts to scale video research notes from roughly 1,000 videos per year to a target of 5,000. But UBS had to slow the avatar rollout for some analysts because the AI flattened certain accents, losing the individual voice that makes research credible to clients. BlackRock's deployment relies on internal "tech translators" embedded with investment teams to make sure the AI extracts useful heuristics. Augmentation at scale still requires human architecture.

The corporate finance picture shows the bifurcation most starkly. Nearly 43% of FP&A job postings now require AI or machine learning skills — up from 33% a year ago. Controller upper-range salaries have collapsed 21% year-over-year. The operational finance roles most exposed to automation are feeling it first.

Meanwhile, CFO lower-range salaries rose 9% over the same period. McKinsey found that in finance functions where AI has been robustly adopted, professionals spend 20 to 30 percent less time on data processing — and that time is being redirected toward scenario design, stakeholder communication, and strategic recommendation. Not eliminated. Redirected.

The automation risk is real and concentrated in specific tasks: the first 60% of most workflows, where data is gathered, models are populated, and first drafts are generated. The judgment-intensive back half — the assumption that makes the model meaningful, the recommendation that earns client trust — is not yet automatable, and that's where the time AI saves should be going.

What the Job Market Is Already Pricing In

Knowing which tasks are under threat is clarifying. The salary and hiring data makes it personal.

The Bureau of Labor Statistics projects 5.7% employment growth for financial and investment analysts through 2034. The profession isn't vanishing. But the criteria for what makes an analyst hireable and well-compensated are shifting fast enough to split the field into two distinct tracks.

Nearly one in three finance job postings — 31% — now explicitly require AI or machine learning capabilities, up from one in four just a year ago. For FP&A roles specifically, it's 43%. AI mentions in accountant job postings surged 67% year-over-year. The market is repricing what a finance professional is expected to know, and it's doing so faster than most annual performance review cycles will catch.

The dollar figure attached to that repricing: job postings that include AI skills offer a 28% salary premium — nearly $18,000 more per year — than comparable postings without AI requirements, according to Lightcast research based on analysis of over 1.3 billion job postings.

This is where Vasylenko's story becomes technically useful rather than just motivating. Two years ago, he told Business Insider, he would have had to walk an AI model through every single step to calculate free cash flow — specifying inputs, sources, formulas. Now, he can hand it five PDF filings, reference three outside sources, and ask it to synthesize assumptions into a calculation. "The technological progress over the past two years is just astonishing," he said. "It's a bit scary when you wonder how useful you'll be going forward."

He said that as someone earning $90 to $160 an hour precisely because he can spot where the model goes wrong — a skill that compounds as the models get better.

Mary Callahan Erdoes, JPMorgan's head of Asset and Wealth Management, described the institutional version of the same dynamic at the firm's 2026 investor day. Her division used AI to transform a controls review process that previously required 200 people to individually read and compare 50-plus pages of documentation. After building the AI-enabled solution, JPMorgan identified 3,000 to 5,000 additional employees who could benefit from the same tool. The goal, Erdoes said, was to eliminate "the no-joy-work in our employees' daily lives, so that they can get on to higher-level added value."

The 31% posting requirement and the 28% salary premium aren't role-specific signals. They're market-wide. Whether you're in buy-side equity, corporate FP&A, credit, or sell-side research, the repricing is already underway.

How Worried Should You Actually Be About Layoffs

The honest answer is: less than the headlines suggest, but not so little that you should stop reading.

AI was cited as the reason for 25% of all U.S. employer-announced job cuts in March 2026, accounting for 15,341 layoffs in that month alone. CFOs surveyed by Fortune are projecting AI-related layoffs nine times higher in 2026 than in 2025. Headline figures from HSBC considering 20,000 roles and Citi's CFO flagging continued headcount reduction are real and ongoing.

The people who will evolve in this environment are those who know their subject matter well and those who know how to combine their knowledge with AI.
by Alexander Vasylenko, Financial Analyst

But a working paper from the National Bureau of Economic Research — based on a survey of 750 CFOs conducted with the Atlanta and Richmond Federal Reserve Banks — tells a more calibrated story. Only 44% of firms plan any AI-related job cuts. When the researchers calculated the aggregate employment impact across the entire U.S. economy, they arrived at roughly 0.4% — about 502,000 roles out of 125 million. Finance sector firms in the same survey reported the strongest labor productivity gains from AI of any sector, but negligible impact on total headcount.

Multiple economists note the complicating factor: "AI is often a scapegoat for things — it's easier to blame AI than softening consumer demand, or uncertainty because of tariffs, or maybe poor HR strategy the past few years in terms of overhiring coming out of COVID." Challenger, Gray & Christmas found that AI was cited in only 4.5% of all 2025 job losses. The fear is real. The scale, so far, is not matching it.

For financial analysts specifically, the genuine risk isn't that the whole role vanishes. It's that the roles built almost entirely around data-gathering and model-formatting mechanics will shrink or be restructured. The protective factor, consistently, is the ability to do what AI cannot: synthesize ambiguous information, judge which assumptions are defensible, and communicate the interpretation that turns model output into a decision someone will act on.

The practical question worth sitting with: what share of your current weekly work would have appeared in OpenAI's Project Mercury training set?

What You Can Do Starting This Week

Vasylenko now holds a full-time analyst position at a major steel producer and earns up to $160 an hour on evenings and weekends reviewing AI financial models — not because he's a technologist, but because he understood early that knowing where the model fails is a skill the market will pay for. "The people who will evolve in this environment," he told Business Insider, "are those who know their subject matter well and those who know how to combine their knowledge with AI."

The bet underneath his schedule isn't the template to copy. The reasoning is.

Four graded actions, lowest commitment to highest:

This week, take one standard deliverable — a variance report, a model refresh, a research summary — and run it through a free LLM. Note specifically where the output fails or requires correction. That failure-spotting instinct is the exact skill that commands $90 to $120 an hour in the AI training market. You're not chasing a side gig. You're learning the boundary of the tool.

This month, pull up three recent job postings for roles at your level. Count how many mention AI or machine learning. If fewer than a third do, that number is about to change — 31% of finance postings already require it, and the share is rising. Know what's coming before it arrives in your next performance review.

Next quarter, take one finance-specific AI fluency module. Not a general prompt-engineering course — something applied to financial modeling, FP&A workflows, or research synthesis. Lightcast's data shows a 28% salary premium attached to AI skills in finance postings. That's nearly $18,000 a year. The return on three months of deliberate upskilling is unusually clear.

Ongoing, track which of your weekly tasks shrink in hours as you use AI tools. That's your personal augmentation map. The tasks that disappear from your plate are the ones you don't want to build your career identity around. The ones that remain — or expand — are where you compound.

The question was never whether AI would change the financial analyst job. It already has. The question is whether you'll understand it from the inside — or find out from the outside, when it's someone else's bet that paid off.

Explore Further