Perspectives on the Profession
Building the Next Generation of CPAs: AI, the Pipeline, and the Case for Calling Us Professionals
By Jessica Irving Marschall, CPA
June 17th, 2026
The accounting profession is facing a talent shortage that, by most honest accounts, has reached a crisis stage. Writing in The CPA Journal, observers no longer describe the pipeline as merely leaking but as a profession in crisis, pointing to a Wall Street Journal report that more than 300,000 accountants and auditors left the profession or labor force between 2019 and 2022. The AICPA has been candid about the stakes: in its Plan to Accelerate Talent Pipeline Solutions, it acknowledges that meaningful, quick action is needed and that the response must be anchored in data that identifies the root causes rather than treating symptoms.
The numbers from the employer side are just as stark. AACSB summarized survey findings indicating that approximately 83 percent of financial leaders reported difficulty finding qualified accounting talent. The causes are familiar to anyone inside the profession: rigid licensure requirements, outdated perceptions of what accountants actually do, demographic decline, and the simple fact that many accounting graduates choose more flexible or more lucrative paths and never sit for the CPA exam at all.
I want to take these challenges in turn, from the vantage point of someone who has lived through the profession’s transformation, and connect three threads that are usually discussed in isolation: how artificial intelligence is reshaping our daily work, why the way we classify and finance this profession is broken, and how we train the next generation of experts when the rudimentary work that once forged them is being automated away.
What AI Actually Does in Our Practice
Across the four companies I help lead, we have integrated AI deliberately and without apology. We did not do it to replace judgment but instead implemented custom-systems to let the technology do what it genuinely does best, which is to catch human error and accelerate rudimentary tasks. Most of our software providers have already embedded these capabilities directly into the tools we use every day (like QuickBooks), so in many cases the question is no longer whether to adopt AI but how thoughtfully to supervise it.
Any CPA who has been in practice as long as I have can tell you what the work used to demand. CPAs who entered the profession before widespread automation will remember what it was like to reconcile a bank statement with a thousand transactions by hand in the old QuickBooks Desktop. It was slow, it was hard on the eyes, and it was precisely the kind of repetitive, high-volume matching where a tired human makes mistakes and a machine does not. That is the work AI should be doing, and increasingly it is.
This is not a niche trend. Thomson Reuters reports that the Big Four have led adoption, but smaller firms are close behind, applying AI to tax research, return preparation, advisory work, bookkeeping automation, and document summarization. In its 2025 survey data, approximately one in five tax firms already identified as using generative AI, with about half more planning or considering it, and the share of firms with no plans at all had fallen sharply from the prior year showing that the direction of travel is unmistakable.
However, the most useful evidence is the kind we accountants love most, which would be hard data. A study by researchers at Stanford Graduate School of Business and MIT, reported by the Journal of Accountancy, found that accountants who used generative AI reallocated roughly 8.5 percent of their time away from routine data entry toward higher-value work, a gain of about three and a half hours in a 40-hour week. In the study, AI adopters recorded a 55 percent increase in weekly client support, posted 21 percent higher billable hours, and closed the month-end books about seven and a half days sooner than peers who did not use AI. Notably, quality did not suffer and the AI users recorded a 12 percent increase in general ledger granularity, capturing transactions in more specific accounts rather than lumping them together. The authors framed the finding exactly as I would. AI is augmenting professional expertise, not replacing it.
Why We Must Be Recognized as Professionals
Here is where I want to plant a flag. We need the Department of Education to formally reclassify CPAs as professionals, and we need to fix the financing of the path into this profession before we lose another generation to it.
The reason this matters became painfully clear in the way the Department implemented the new graduate loan limits enacted in last year’s One Big Beautiful Bill Act. In carrying out that law, the Department narrowed the set of graduate degrees that qualify as professional, the category that allows students to borrow up to $50,000 a year and $200,000 overall. In implementing the law, the Department relied on a longstanding regulatory definition of professional degrees that currently includes eleven fields: chiropractic, clinical psychology, dentistry, law, medicine, optometry, osteopathic medicine, pharmacy, podiatry, theology, and veterinary medicine. As NPR reported, that list was drawn from a regulation that had not been meaningfully updated since the 1950s, an era when many modern graduate programs barely existed. Accounting does not appear on that list, and many nursing pathways do not either.
College already costs too much, and the road to CPA licensure runs through 150 credit hours of education. If students cannot finance that education on reasonable terms, the pipeline narrows at exactly the point where we most need it to widen. Nursing is in the same boat, and the parallel is instructive. A coalition of two dozen states and the District of Columbia has sued over the rule, arguing it will shut talented people out of critical professions at a time when communities already face provider shortages. The American Nurses Association said it was profoundly dismayed. When a federal rule treats two essential, exam-gated professions as something less than professional, the rule, not the professions, is what needs revisiting.
There is a second front, and it cuts the other way. Even as the Department declines to call us professionals for financing purposes, CPA licensure itself is under pressure in state legislatures. On the Journal of Accountancy podcast, James Cox of the AICPA, who also leads the Alliance for Responsible Professional Licensing, described a wave of deregulation efforts that weaken the qualifications for licensure and target the very boards on which licensing depends. It is, he noted, a bipartisan phenomenon driven by state budget pressures and free-market sentiment, and the attack on licensing boards is the most pervasive trend he sees. The irony is hard to miss as we are deemed not professional enough to merit ordinary student financing, yet professional enough that the public stakes its trust on our work.
And the public does value what licensure protects. Cox cited polling showing that roughly three in four voters consider licensing essential to maintaining professional standards, a similar share believe it keeps the public safe, and nine in ten business owners say it protects and enhances their reputation. A school board trusts that a licensed CPA audited its financials. Main Street trusts that the numbers are true. That trust is the franchise, and it is worth defending on both fronts at once.
Training Experts When the Basics Are Automated
This brings me back to AI and to the hardest question the profession faces. If the machine now does the rudimentary work, how does the next generation ever become expert?
I learned my craft through the basics, and I do not say that with nostalgia so much as with conviction. I reconciled those thousand-transaction bank statements by hand. I built depreciation and amortization schedules manually, line by line and year by year, until the patterns lived in my head. The work was tedious and it was also exactly how I became an expert. The repetitive nature of the fundamentals is what builds the instinct that lets a seasoned CPA glance at a trial balance, 1120S, Schedule E, or budget and sense, before any tool flags it, that something is off.
The Stanford research should give us pause precisely on this point. It found that senior accountants gain the most from AI because they treat it as a collaborator, stepping in when the system’s confidence drops and applying judgment where it matters. Junior staff, by contrast, are more likely to accept AI-generated output at face value even when it is flagged as uncertain, and they realize smaller gains as a result. The study went so far as to warn about AI-generated errors flowing through human-in-the-loop systems. The CPA Journal authors similarly observed, when its authors ran live tests through a leading AI model, that it was useful for tedious work, and quick to fall short where context and judgment are required.
The answer is not to keep AI away from young accountants. The answer is to make sure they still do enough of the foundational work to develop the judgment that lets them supervise the machine rather than rubber-stamp it. This is where intergenerational mentorship, which The CPA Journal rightly emphasizes, earns its keep. Experienced CPAs transfer hard-won knowledge and context, while younger colleagues bring genuine fluency with the new tools. AI should free our people for higher-value work. It must not be allowed to hollow out the apprenticeship that makes expertise possible in the first place.
The Pipeline Is Worth Building
For all these challenges, my conclusion is that we must continue to build out the CPA pipeline, and we should do it with confidence. Accounting is one of the most versatile majors a young person can choose. It is the language of business, and it travels everywhere business goes. It has allowed me, a mother of five, to build and run four companies spanning tax, valuation, sustainability, and advisory work. Very few degrees offer that range of possibility.
But versatility alone will not fill the pipeline. If we are serious, the profession needs to act on a few specific commitments:
- Pay starting CPAs more. The apprentice-style model of low early pay justified by a distant partnership belongs to a different era and is actively driving talent away.
- Make raises and bonuses strategic and targeted, rewarding the specific skills and judgment the profession actually needs rather than rote tenure.
- Insist, to the Department of Education and to anyone else who will listen, that passing one of the country’s most rigorous professional licensing examinations in the country does indeed make us professionals.
We are the people who certify that the numbers are true, and a society that runs on trust in those numbers cannot afford to starve the pipeline that produces us. The work is worth defending, worth financing, and worth building for the generation that comes next.
Sources
1. AICPA & CIMA, Plan to Accelerate Talent Pipeline Solutions. https://www.aicpa-cima.com/resources/article/draft-plan-to-accelerate-talent-pipeline-solutions
2. The CPA Journal, Intergenerational Solutions to Address the Crisis of the Leaking Accounting Pipeline (April 2026). https://www.cpajournal.com/2026/04/01/intergenerational-solutions-to-address-the-crisis-of-the-leaking-accounting-pipeline-2/
3. AACSB, Rebuilding the Pipeline for Accounting Talent (June 2025). https://www.aacsb.edu/insights/articles/2025/06/rebuilding-the-pipeline-for-accounting-talent
4. NPR, States sue over new student loan limits on certain nursing and healthcare degrees (May 2026). https://www.npr.org/2026/05/19/nx-s1-5826688/lawsuit-student-loans-nursing-healthcare-graduate-degree
5. Journal of Accountancy, Deregulation’s state of play and the threats it poses to CPA licensure (May 2026). https://www.journalofaccountancy.com/podcast/2026/may/deregulations-state-of-play-and-the-threats-it-poses-to-cpa-licensure/
6. The CPA Journal, How Artificial Intelligence May Impact the Accounting Profession (September 2025). https://www.cpajournal.com/2025/09/08/how-artificial-intelligence-may-impact-the-accounting-profession/
7. Stanford Graduate School of Business, AI Is Reshaping Accounting Jobs by Doing the Boring Stuff (June 2025). https://www.gsb.stanford.edu/insights/ai-reshaping-accounting-jobs-doing-boring-stuff
8. Thomson Reuters, How do different accounting firms use AI? (November 2025). https://tax.thomsonreuters.com/blog/how-do-different-accounting-firms-use-ai-tri/
9. Journal of Accountancy, Calculating AI’s impact on CPAs: New study quantifies time savings (August 2025). https://www.journalofaccountancy.com/news/2025/aug/calculating-ais-impact-on-cpas-new-study-quantifies-time-savings/
