Something strange is happening in accounting right now. The profession is growing, with employment for accountants projected to increase by 10% through 2026, faster than most other careers. Yet new accounting graduates are struggling to find jobs. How can both things be true at the same time?
In a recent episode of The Accounting Podcast, hosts Blake Oliver and David Leary tackled this paradox head-on, sparked by a sobering prediction from Anthropic CEO Dario Amodei: AI could displace 50% of entry-level white-collar jobs in the next one to five years.
“Accountants coming out of school are having a hard time getting jobs right now, which is strange,” Blake said during the discussion. “The profession is growing, but entry-level jobs are declining.”
The reason behind that paradox is AI is automating exactly the kind of work that new accountants traditionally cut their teeth on. And that’s creating a bigger problem than just unemployment. It’s threatening the way accountants have learned their craft for generations.
The Work That’s Disappearing
Think about what entry-level accountants actually do (or used to do). They gather documents for audits and organize them into folders. They copy last year’s work papers and update them with new numbers. They reconcile accounts, enter data, and basically do the grunt work that teaches them how financial information flows through a business.
These aren’t exciting tasks, but they serve a purpose. By doing this work, new accountants develop an eye for what looks right and what doesn’t. They learn where errors hide and build intuition.
Now, AI is making all of that work disappear.
During the episode, Blake described Anthropic’s new Claude Cowork feature. It’s AI that can literally click around on your computer as a human would. In one example, a journalist asked it to organize a messy folder of business receipts. The AI asked a few questions about how to sort them, then went to work. Five minutes later, it produced a clean Excel spreadsheet listing two years of expenses.
“File management, organizing folders, batch renaming, generating summary spreadsheets,” Blake noted. This is exactly the kind of work that used to take hours of a staff accountant’s time.
The capabilities keep expanding. Claude now has an Excel integration that lets you skip learning formulas entirely. Need to split a column with full names into separate first and last name columns? Just type what you want in plain English.
“All the Excel wizards are going to be in trouble,” David joked, “because I’ll be able to type in ‘put two more columns, first name, last name separately,’ and it’ll go do the formula for me.”
Another striking example is that Gusto now lets you run payroll directly through ChatGPT. Just type “@Gusto, help me run payroll” and have a conversation. No separate login or clicking through menus—just chat.
Where Do New Graduates Go Now?
This automation creates a problem that goes way beyond individual job losses. Most accountants start their careers in public accounting, doing exactly this kind of entry-level work. That’s how they learn the skills they’ll need to become managers and partners later on.
“Where do accounting graduates go to get their first jobs if they don’t go into public accounting?” Blake asked. “Who’s going to want to hire them?”
The challenge is that new graduates don’t have the experience to do the mid-career jobs that are actually in demand. They can’t review work they’ve never done themselves. As David put it, “They don’t have the skills to monitor AI yet, or know when AI is wrong.”
Workers sense this shift happening. According to a survey mentioned by CFO.com, 60% of U.S. adults believe AI will eliminate more jobs than it creates by 2026. Only 16% think AI could never replace what they do.
This fear is already changing behavior. Last year, 51% of workers said they’d quit if their company demanded a return to the office. This year, that number crashed to just 7%.
“The pendulum has swung back to employers,” Blake observed. “Employers have the power in the market.”
The Industry Starts to Respond
Some companies are beginning to tackle this problem, though it’s unclear if their efforts will be enough.
Intuit announced its most ambitious initiative so far: a program to upskill 1 million accounting students over the next five years. They’re focusing on “digital data and advisory skills,” basically teaching students to work alongside AI rather than compete with it.
The program includes online learning, mentorship, and certifications like QuickBooks ProAdvisor. It kicked off with a virtual event in early February about “Skills for the New Era of Accounting.”
But beyond formal programs, the hosts suggested that individual accountants need to change how they think about their work entirely.
“The most important skill in the AI era is going to be curiosity,” Blake argued.
He shared a practical tip for working with AI. Instead of trying to write the perfect prompt, have the AI ask you questions. His go-to addition to any request is, “First, ask me questions to clarify exactly what I’m looking for. After each of my answers, reply with your next question. Include a confidence score as a percentage. When you achieve 100% confidence, ask me to confirm I am ready for you to do the task.”
David added another crucial insight. “That ‘please wait until I say go’ is the most important prompt.” Without explicit instructions to wait, AI tools tend to race ahead with incomplete information.
A Paradox for Payroll Companies
The episode raised an interesting question about companies like ADP, which announced new AI agents for payroll and HR tasks. These agents can audit for errors, flag missing tax IDs, and answer HR questions using company handbooks.
But as David pointed out, there’s something odd about payroll companies embracing AI so enthusiastically. “If you believe AI is going to be important, that’s going to kill millions and millions of these jobs that get paid through ADP,” he said. If AI eliminates 50% of entry-level jobs, that’s potentially half of ADP’s revenue from running payroll.
Paychex mentioned AI 50 times in their recent earnings call. Investors are clearly asking hard questions about how payroll companies will survive if their customer base of employed humans shrinks dramatically.
What Happens Next?
The tools reshaping accounting are here now. Claude can organize receipts. ChatGPT can run payroll. AI agents can audit for errors. Every month brings new capabilities that used to require human hands and human hours.
The mismatch is striking. Technology advances in months, while professional training evolves over years or decades. Universities design programs on multi-year cycles. Firms have hiring patterns built over generations. Everyone assumes the entry-level work that has trained accountants forever will continue to exist.
That assumption is crumbling. The document gathering, data entry, and reconciliations are becoming prompts typed into AI interfaces rather than skills learned through repetition.
For firms, this means rethinking how new hires develop competence. For educators, it means teaching students to verify and question rather than just execute. For students and early-career professionals, it means learning to supervise technology rather than compete with it.
As Blake put it near the end of the episode, “I can be ten times as productive now doing everything I do with AI, and I don’t need to hire a huge team to do it.”
That’s great news for experienced professionals who already know their craft. For those just starting out, it’s an entirely different story—one the profession is only beginning to write.
Listen to the full episode of The Accounting Podcast for more insights on AI tools reshaping accounting workflows, practical prompting techniques, and updates from Intuit, Xero, ADP, and other platforms. Understanding how AI affects entry-level work is table stakes for anyone making decisions about talent and training in the years ahead.