The accounting profession faces a stark reality-check as entry-level auditor positions have declined by 43% since January, and a third of accountants admit they cannot identify AI-generated fake receipts.
In episode 455 of The Accounting Podcast, hosts Blake Oliver and David Leary address the growing evidence that AI is disrupting accounting more rapidly than most firms can keep up with. From vanishing entry-level jobs to the rise of “work slop” (low-quality AI output that wastes time and money), the profession is struggling with changes that are both promising and perilous.
The Tech Stack Problem Nobody Wants to Talk About
Before diving into AI’s disruption, Oliver shared a surprising statistic: only 37% of accounting firms require their clients to use their technology stack. That means 63% let clients choose their own tools, creating a mess of incompatible systems and inefficient workflows.
“It’s one of the things we did in my firm that was a differentiator and allowed us to scale quickly,” Oliver explained. “It reduced training time. It increased the speed at which we worked.”
The reluctance to standardize reveals a deeper problem in the profession: firms are so afraid of losing clients that they sacrifice efficiency and scalability. Yet Oliver found the opposite: “The ones that were willing to make that shift ended up listening to us about other things, too. So you might want to consider requiring clients to switch as, like a testing mechanism to see if they’re actually going to be a good fit.”
This standardization challenge becomes even more critical as firms try to implement AI. Without consistent data inputs and workflows, automation becomes nearly impossible.
The Vanishing Entry Level: A 43% Wake-Up Call
The most alarming news Oliver shared was the 43% drop in entry-level auditor job postings since January, based on a study of 126 million job postings. Meanwhile, senior positions requiring ten or more years of experience increased by 6%.
“These firms are extremely shortsighted,” Oliver argued. “They are just trying to juice profits and revenue in the short term. And the easiest way to do that is to replace your entry-level people with AI.”
The vulnerability of these positions is clear. As Oliver explained, “The stuff an entry level auditor does is so basic, like cash confirmations. You can have an AI agent doing cash confirmations all day long. It’s not complicated.”
The fear extends beyond auditing. Nearly half (45%) of accounts payable professionals now fear layoffs in 2025, up from 27% last year. These workers see AI agents matching invoices, approving bills, and processing expense reports—tasks that once required human oversight.
Leary raised an important question: Are firms actually succeeding with AI, or are they cutting staff first and hoping to automate later? In Oliver’s view, the automation is working well enough to justify the cuts. But this creates a long-term problem. Without entry-level positions to train tomorrow’s senior accountants, where will future leaders come from?
Work Slop: The $200 Hidden Cost of Bad AI
A new Harvard Business Review study coined a term for low-quality AI output: “work slop.” And work slop is expensive. Each incident wastes nearly two hours and costs about $186 per worker per month.
Forty percent of workers report receiving work slop in the past month. More than half feel annoyed when they get it, and 42% view the senders as less trustworthy.
“Every time one coworker gives another coworker slop, it costs your company 200 bucks,” Leary emphasized. But, “Employees who turn out work slop probably already did work slop before. They just did it at a much slower volume.”
The hosts shared their own experiences with work slop. Job applicants submit unedited ChatGPT responses. Guest pitches reference the wrong podcast. Some candidates even feed interview questions into AI during live video calls.
“It looks good,” Oliver said about typical work slop. “Like if you look at the email, it’s nicely formatted and it looks good and then you actually read it and you realize that it’s garbage.”
The paradox is striking: 97% of firms admit they’re not using technology efficiently, yet 86% believe AI-using firms have a competitive advantage. The gap between aspiration and execution means firms produce more low-quality work faster rather than better work more efficiently.
The Fraud Detection Crisis
Perhaps most concerning is accountants’ declining ability to spot fraud. Thirty-two percent admit they can’t recognize AI-generated fake receipts. Another 30% are seeing more fraudulent receipts than last year, and 42% suspect colleagues have submitted fake or altered receipts.
“If you want to see just how difficult it is or how easy it is to make one, just go and ask ChatGPT to make you a receipt,” Oliver challenged listeners.
Leary noted that expense fraud isn’t new. After all, people used to pick a receipt up off the ground at McDonald’s. But AI changed the game. Now anyone can generate perfect forgeries on demand.
Oliver explained that current AI models don’t understand physics, so shadows and lighting in fake images often don’t match reality. But detecting these requires expertise most accountants don’t have.
“When nothing is physical anymore, how do you, as an auditor or an accountant, rely on a scanned document?” Oliver asked, highlighting a fundamental challenge for the profession.
Solutions Emerging from the Chaos
Despite the challenges, practical solutions are emerging. Zapier announced a “human in the loop” feature that pauses automated workflows for human review at critical points. “Don’t try to automate the whole workflow,” Oliver advised. “Try to automate one task in the workflow.”
Keeper launched a new AI product that converts payroll reports and settlement statements into journal entries—a task that previously required complex spreadsheets and manual work. At $50 per client per month, it represents the kind of targeted automation that actually works.
Even Drake Software, long criticized for being behind the times, launched cloud-based tax software. While limited to certain forms, it signals that even legacy providers recognize the need to modernize.
These tools show that successful AI implementation isn’t replacing humans entirely. Instead, it augments specific tasks while maintaining human oversight for quality and judgment.
Looking Ahead: A Profession at a Crossroads
The accounting profession faces interconnected challenges that require more than technological solutions. The 43% drop in entry-level positions poses a threat to the talent pipeline. Work slop erodes trust and efficiency. Fraud detection capabilities are falling behind those of fraudsters.
Yet there are opportunities within these challenges. Firms that thoughtfully integrate AI, maintain human oversight, and invest in training the next generation will have an advantage over those who chase short-term profits by cutting entry-level positions and blindly implementing AI.
As Oliver noted about his own firm’s success, standardizing technology, requiring client buy-in, and focusing on quality over quantity created real competitive advantages. The same principles apply to AI adoption. Success requires strategy, not just software.
To hear Oliver and Leary’s complete analysis of these shifts in accounting, including their discussion of H-1B visa changes, Trump’s latest tariff threats, and more practical insights for navigating AI’s impact, listen to the full episode of The Accounting Podcast. Their unfiltered weekly discussions provide essential perspective for anyone trying to understand where the profession is heading and how to thrive despite the uncertainty.