Demand for CPAs Will Increase: In the World of AI, Human Judgement Becomes a Premium

Written by Andrew Ross

Artificial intelligence is beginning to change the way work gets done inside CPA firms. Tasks that once required hours of effort, such as reconciling accounts, summarizing client records, drafting memos, classifying transactions, preparing schedules, and identifying exceptions, can now be completed or accelerated with the support of AI. Some believe this will eliminate the need for CPAs; however, I believe the opposite is true. AI will make accounting work faster, but it will also place a greater emphasis on the value of the CPA and the professional judgement they bring to the process.

The CPA’s role will increasingly shift from manual preparation to reviewing AI-assisted work, validating the reasoning behind it, identifying what may be missing, and applying professional judgement before anything is relied upon by a client, a lender, a regulator, or a business owner. Clients will demand it. With AI solutions being used to prepare, analyze, and recommend, clients will continue to require a qualified professional to determine whether the output can be trusted.

That is the central issue: trust. Trust does not come from speed. It does not come from automation. It comes from knowing that the right facts were considered, the assumptions were reasonable, the evidence supports the conclusion, and the final recommendation can stand up to professional scrutiny. This is the value CPAs bring, and this becomes even more important in the world of AI. The future of the profession will belong to those who understand this clearly: AI may change the work, but human judgement will define the value.

This change is happening at the same time that public practice is already facing a shortage of CPAs. Firms need more professionals who can work with AI, exercise judgement, supervise technology, communicate clearly with clients, and protect trust in the financial information businesses depend on. The challenge, therefore, is not just how CPA firms adopt AI, but how firms develop the next generation of CPAs in an environment where AI may produce the first draft, but the professional remains responsible for the final judgement. That question should shape how firms train junior staff, how bookkeepers and accounting technicians evolve, how universities and colleges prepare students, and how current CPA professionals get ready for the next stage of public practice.

Human Judgement and Trust

Accounting work is not valuable simply because it is completed quickly. It is valuable because clients, lenders, regulators, boards, and business owners can rely on it.

The starting point for any discussion about AI in public practice should be trust. Accounting work is not valuable simply because it is completed quickly. It is valuable because clients, lenders, regulators, boards, and business owners can rely on it. That reliance depends on the process behind the answer and someone to confirm a complete and correct answer.

AI is producing the first draft with impressive speed, summarizing a lease, drafting a variance explanation, classifying a transaction, reading a stack of invoices, or preparing an initial reporting package. Those capabilities will save time, but AI still needs the “right” data, with “right” context. Someone is still needed to determine whether the output is right, complete, and fit for use.

A bank reconciliation that looks complete may still miss a stale-dated cheque issue. A lease summary may ignore an amendment. A tax memo may cite the right rule but apply it to the wrong facts. A transaction classification may be statistically likely and still be commercially wrong for that specific client. These are not small details. They are the difference between information that appears useful and information that can be relied upon.

This is why the CPA becomes more important, not less. The CPA is the professional who reviews the output, tests the reasoning, confirms the facts, challenges the assumptions, and decides whether the conclusion is defensible. In an AI-assisted firm, the CPA is no longer only reviewing whether the number ties. The CPA is reviewing whether the system understood the transaction, considered the right evidence, applied the right policy, identified the right risk, and reached a conclusion the firm can stand behind.

CPA.com’s 2025 AI in Accounting Report describes this shift clearly. Firms are investing in AI-powered workflow automation for areas such as bank reconciliations, transaction coding, month-end close, and reporting. At the same time, human-in-the-loop verification remains essential, especially in high-stakes areas such as tax, assurance, and advisory work.

That is the human judgement premium. As AI produces more work, the scarce value shifts to the person who can determine whether the work can be trusted. The output may come from AI. The accountability cannot.

Developing CPAs

The need for trusted judgement is rising at the same time that public practice is under pressure to attract and develop enough professionals. That matters because AI is often discussed as if it is entering a profession with excess capacity. It is not. CPA firms are already dealing with increasing client expectations, more complex business models, changing tax rules, technology adoption, retirements, and a tighter talent pipeline.

In the United States, the Bureau of Labor Statistics projects employment for accountants and auditors to grow 5% from 2024 to 2034, with about 124,200 openings per year on average over the decade. The same outlook notes that while routine tasks may be automated, advisory and analytical duties are expected to become more prominent rather than reduce the need for the profession.

Similarly, the AICPA’s 2025 Trends Report shows the pressure from another angle. Bachelor’s degree completions in accounting declined 3.3% in 2023-2024 after a larger decline the prior year, while master’s degree completions declined 15%. At the same time, 75% of responding firms that hired accounting graduates in 2024 expected to hire the same number or more in 2025.

Canada faces its own version of the same issue. The Canadian Occupational Projection System projects 83,100 job openings for financial auditors and accountants over the 2024-2033 period, driven by both job creation and replacement demand.

AI is not arriving to solve a problem of too many CPAs, rather, it is arriving in a market that needs more professional capacity and more professional judgement. Firms need AI to increase capacity which creates a bigger risk; developing staff and having the staff supervise the AI work.

For decades, CPA firms trained professionals through preparation. A junior staff member prepared the working paper. A senior reviewed it. The junior received review notes, corrected the work, and learned. Over time, repetition built an understanding and comprehension that developed their judgement. A junior accountant learned what a clean reconciliation looked like by preparing many imperfect ones. They learned what a weak explanation looked like because a reviewer challenged it. They learned how one missing document could change a conclusion because they had to go back into the file and fix the issue.

AI changes that learning path. If AI prepares more first drafts, junior professionals may have fewer opportunities to learn through manual repetition. The Journal of Accountancy has described this as one of the profession’s key training questions: if automation and AI are taking over the repetitive, lower-risk work young accountants historically used to learn systems, controls, and professional skepticism, how do firms train accountants when the training work starts to disappear?

The answer is not to preserve inefficient manual work just because it once served as a training ground. That would be like asking new pilots to ignore modern instruments because earlier pilots learned without them. The answer is to redesign apprenticeship around the work that will matter most in an AI-assisted environment. Future CPAs need to learn how to challenge AI-generated work. They still need the technical foundation in accounting, tax, assurance, finance, and business from solid education and then they need to learn how to inspect the reasoning, validate the evidence, identify missing context, and explain why they accepted, edited, or rejected an AI suggestion.

A junior accountant may no longer spend three hours building the first draft of a working paper from scratch. Instead, they may spend one hour reviewing an AI-generated draft, checking source documents, documenting exceptions, and preparing a short conclusion for manager review. That can still be powerful training if the firm is deliberate about what the junior is learning.

The review note of the future should not simply say, “Fix this calculation.” It should ask: Did you verify the source? Did you understand the assumption? Did you consider the prior-year treatment? Did you identify the exception? Did you explain why you accepted or rejected the AI suggestion?

That is how judgement gets built in the next generation of CPAs. The learning moves from doing every step manually to understanding the work deeply enough to review, challenge, and improve it.

Formal Education

If the work changes, education has to change as well. Universities and colleges cannot prepare students for AI-supported accounting by simply adding one lecture on AI ethics or one assignment using a chatbot. AI has to be integrated into how accounting is taught. That requires a different kind of learning. Students should be given AI-generated workpapers and asked to find what is wrong. They should review AI-generated tax memos and identify missing facts. They should inspect bookkeeping-agent outputs and explain which items should be accepted, edited, rejected, or escalated. They should learn how to document their reasoning, not just arrive at an answer.

CPA education in Canada is also moving in this direction. The new CPA Professional Program, launching in 2027, is being developed to align with Competency Map 2.0 and to equip CPAs to lead in a fast-moving world. CPA Ontario’s education partners have described the coming program as blending technical excellence, ethics, real-world experience, and future-focused skills that evolve with the demands of a world being reshaped by AI.

Colleges and universities should also think carefully about the bookkeeping and accounting technician layer. If bookkeepers are going to become agent managers, accounting programs must teach more than debits, credits, and software navigation. They must teach data quality, workflow design, exception handling, source-document reliability, system controls, and AI supervision. A student should graduate knowing not only how to prepare a bank reconciliation, but also how to supervise an AI system that prepares one.

How CPA Professionals Can Get Ready

For current CPA professionals and firm leaders, the next step is not to wait until AI is perfect. It will not be perfect. The better approach is to start building controlled workflows that allow the firm to learn safely.

Start with one workflow. Choose something repetitive, meaningful, and reviewable: transaction classification, bank reconciliation support, variance commentary, document intake, tax memo preparation, or month-end reporting. Then define what AI is allowed to draft, what a bookkeeper or technician must validate, what a junior accountant must review, and what requires CPA judgement.

This is also where firms need clear governance. Deloitte’s poll of finance and accounting professionals found that trust was the leading barrier to agentic AI adoption, followed by integration into existing systems and lack of skilled personnel. The same poll found that 59.7% of respondents trusted AI agents to make decisions only within a defined framework, while judgement calls should continue to be made by people.

That finding should not surprise anyone in the accounting profession. Trust is not created by saying the AI is accurate. Trust is created by workflow design, review discipline, access controls, source traceability, documentation, and accountability.

In practical terms, firms should be able to answer a few basic questions before AI output is relied upon: Who approved the AI output? What data did the system use? What did the human reviewer change? What exceptions were escalated? Where is the audit trail? What work is the AI allowed to do, and what still requires professional review?

The safest operating model is not “AI decides.” The safer model is: AI proposes, the professional decides. That model protects the firm, the client, and the profession. It also creates the right learning environment for staff. Junior professionals can see the AI output, review the evidence, make a recommendation, receive feedback, and improve. Bookkeepers can monitor agents and manage exceptions. CPAs can focus more of their time on judgement, advisory, assurance, client communication, and risk.

Wolters Kluwer’s 2025 Future Ready Accountant Report found that AI adoption among firms rose from 9% in 2024 to 41% in 2025, with 77% planning to increase AI investment and 31% citing advanced technical skill development as a top staffing challenge. The message for CPA firms is clear: technology adoption and people development have to move together.

Current CPA professionals can prepare by becoming more intentional about how they use AI. They should learn how AI tools produce outputs, where those tools fail, how to request source-backed reasoning, how to document review procedures, and how to explain AI-assisted conclusions to clients. They do not need to become software engineers. They do need to become better supervisors of technology.

Conclusion: CPAs Owns the Final Judgement

AI will change the accounting career ladder. It will change the work of bookkeepers, how junior staff learn, what universities and colleges need to teach, and how CPA firms design their operating model.

But it will not remove the need for CPAs.

In fact, as AI becomes more capable, the need for trusted professional judgement will increase. Clients will still need someone accountable. Firms will still need someone to sign off. Business owners will still need someone who understands the facts, the risks, the numbers, and the human context behind the decision.

The firms that succeed will not be the ones that simply automate the most work. They will be the ones that build the strongest connection between AI capability and professional judgement. They will use AI to create capacity, but they will also build the review discipline, training model, governance structure, and education partnerships required to protect trust.

That is the opportunity in front of the profession. CPA firms should not respond to AI with fear, blind adoption, or nostalgia for the old training model. They should respond by building a better model.

AI can generate the first draft. The profession still owns the final judgement. And in the world of AI, that judgement becomes the premium.

Auciera was built on this principle: AI handles the work that does not require judgement, so the CPA can focus entirely on the work that does.

Author

References

1. CPA.com. “2025 AI in Accounting Report.” The report discusses AI-powered workflow automation, human-in-the-loop verification, agentic AI, bookkeeping automation, tax review, audit risk, advisory services, and the need for firms to build governance and staff capabilities. https://www.cpa.com/sites/cpa/files/2025-06/CPAcom-2025-AI-in-Accounting-Report.pdf
2. U.S. Bureau of Labor Statistics. “Accountants and Auditors.” Occupational Outlook Handbook. The BLS projects 5% employment growth for accountants and auditors from 2024 to 2034, about 124,200 annual openings, and notes that automation is expected to make advisory and analytical duties more prominent rather than reduce overall demand. https://www.bls.gov/ooh/business-and-financial/accountants-and-auditors.htm
3. AICPA & CIMA. “2025 Trends Report.” The report identifies trends in U.S. accounting graduations, hiring of new graduates, and firm expectations. It reports declining bachelor’s and master’s accounting completions but continued optimism among responding firms about hiring accounting graduates. https://www.aicpa-cima.com/professional-insights/download/2025-trends-report
4. Employment and Social Development Canada. “Financial Auditors and Accountants – Canadian Occupational Projection System.” The projection identifies 83,100 job openings for financial auditors and accountants in Canada over the 2024-2033 period. https://occupations.esdc.gc.ca/sppc-cops/occupationsummarydetail.jsp?lang=eng&tid=16
5. Journal of Accountancy. “How will accountants learn new skills when AI does the work?” The article directly addresses the apprenticeship challenge created when AI takes over repetitive training work and argues for conceptual mastery, human judgement, technological fluency, and adaptability. https://www.journalofaccountancy.com/issues/2026/mar/how-will-accountants-learn-new-skills-when-ai-does-the-work/
6. Deloitte. “Trust Emerges as Main Barrier to Agentic AI Adoption in Finance and Accounting.” Deloitte’s poll found strong optimism around AI agents, but also identified trust, integration, and lack of skilled personnel as adoption barriers, with most respondents preferring AI decisions only within defined frameworks while judgement calls remain with people. https://www.deloitte.com/us/en/about/press-room/trust-main-barrier-to-agentic-ai-adoption-in-finance-and-accounting.html
7. Wolters Kluwer. “2025 Future Ready Accountant Report.” The report found AI adoption among firms rose from 9% in 2024 to 41% in 2025, with 77% planning to increase AI investment and 31% citing advanced technical skill development as a top staffing challenge. https://www.wolterskluwer.com/en/news/wolters-kluwer-releases-its-2025-future-ready-accountant-report
8. CPA Competency Map 2.0. The Canadian CPA Competency Map identifies data and data governance, real-time decision-making, technological innovation, automation, and evolving entry-level roles as major themes shaping the future of the profession. https://assets.cpaontario.ca/students/pdf/cpa-competency-map.pdf
9. CPA Western School of Business. “Introducing the CPA Professional Program.” The Canadian CPA profession is updating certification to align with Competency Map 2.0, including refreshed education, examinations, and practical experience requirements. https://www.cpawsb.ca/cpa-professional-program/
10. CPA Ontario Centre for Accounting and the Public Interest, Ivey Business School. “Begin your journey to a CPA designation.” The page describes the new CPA Professional Program as blending technical excellence, ethics, real-world experience, and future-focused skills for a world redefined by AI. https://www.ivey.uwo.ca/accountingcentre/for-students/prospective-students/