Anderson Hall. Photo by ShameraKane, CC BY-SA 4.0.
If you’re a junior or senior in the College of Business, you’ve spent the last two years being told that accounting and finance are first in line for automation. A study published this year, built on job ads rather than predictions, points the other way, and it’s worth understanding what it does and doesn’t say before anyone changes a major over it.
What the study looked at
Coresignal, a company that collects public web data, worked from 2.6 million U.S. entry-level job ads posted in the first five months of 2025 and the first five months of 2026. Nobody was surveyed. The method is simply to read the title on each posting and file it by the kind of work it names, with one file reserved for tasks software has gotten good at, like pulling reports together or squaring numbers that should match. Those get tagged AI-exposed, and the tag’s share is tracked inside each of fourteen job areas.
The finance number went down
Entry-level finance and accounting ads in the AI-exposed category went from 45.24% in 2025 to 43.27% in 2026. That’s a drop of 1.97 percentage points, and it’s the second-largest decline anywhere in the table. The share of finance ads describing manual work stayed close to flat, moving from 0.96% to 0.92%.
Two things are true at once here and both matter. Finance is still one of the most AI-exposed job areas there is, sitting above 43%, well ahead of areas like operations or sales that score in low single digits. But the direction of travel over the past year was down, in a year when consulting rose 4.47 points and marketing rose 3.02.
Why a high number isn’t a countdown
This is where the label gets misread most often. “AI-exposed” describes what the work involves, not what happens to the job. Reconciling accounts and pulling together a variance report score high on a measure like this because software can help with them, and that was true before anyone said “large language model” out loud. A high share means the category contains a lot of that kind of task. It is not a forecast that employers stop hiring.
The study is also counting shares of job advertisements, not people who got hired. If finance postings grew in absolute terms while the mix shifted, a falling share is compatible with more openings, not fewer. The data can’t distinguish those cases and doesn’t claim to.
There’s a caution inside the study that’s easy to skip past. When Coresignal broke down the nationwide drop in postings for hands-on work, most of the movement turned out not to be jobs ending. Only about a third of that decline reflected employers genuinely changing what they hire for. The bulk of it was job titles sliding into blander wording, which a title-reading classifier files somewhere new even though the work is the same. Some of what looks like upheaval is a rewrite.
What the hiring numbers say
Job-ad data describes what employers advertise. For what they actually do, the National Association of Colleges and Employers surveys employers directly, and its Job Outlook 2026 Spring Update found employers projecting a 5.6% increase in hiring for the Class of 2026. That’s a revision upward from the flat 1.6% NACE projected back in November 2025, and it follows two weak years for new graduates.
Neither dataset says the entry-level market is easy. Both say it’s steadier than the commentary around it.
The practical read for anyone about to graduate into an accounting or finance role is that the category isn’t emptying out, the tasks inside it are shifting, and the students who can work alongside the tooling will be sorting themselves from the ones who can’t. That’s a less exciting sentence than the ones in your feed, and it’s better supported.
Coresignal draws no causal conclusion from any of this, and flags that its own way of spotting an AI-building company is rough. The whole department table, finance row included, sits in its study of entry-level hiring across 14 job areas.
