A first-year accountant in 2015 spent an unglamorous share of their time on bank reconciliations, ticking and tying, agreeing subledgers, and chasing variances that turned out to be timing differences. It was widely considered a waste of a bright graduate's time. It was also, though almost nobody described it this way, how they learned what a wrong number looks like. That training mechanism is being automated away faster than the profession has designed a replacement for it.
Key Takeaway
Two things are simultaneously true and in tension. Demand for experienced accountants is strong: CPA Canada's 2025 Compensation Study reports median compensation of $154,000 for CPAs with three or more years of post-designation experience, up from $143,000 in 2023, with a 7.7% climb between 2022 and 2024 that outpaced inflation. At the same time, firms are reducing entry-level openings as they implement AI, and Stanford researchers found entry-level workers in AI-exposed occupations experienced a 13% employment decline since 2022. The paradox is that the entry-level tasks AI now performs were the mechanism by which juniors developed the professional judgment that makes senior accountants valuable. A profession can automate its apprenticeship, or it can produce experienced practitioners, but it has not yet worked out how to do both. There is one genuinely encouraging counter-signal: AICPA data shows accounting enrolment at four-year institutions rose 8.9% in spring 2026, a third consecutive year of growth.
Two Facts That Sit Awkwardly Together
Most commentary on this subject picks one of two narratives and runs with it. The first is the shortage story: not enough people are entering accounting, experienced practitioners are retiring, and firms cannot fill roles. The second is the displacement story: AI is automating accounting work and entry-level roles are disappearing. Each is supported by real evidence, and they appear to contradict each other, which is why coverage tends to choose one and ignore the other.
They do not actually contradict. They describe different segments of the same labour market: strong and strengthening demand for experienced professional judgment, alongside weakening demand for the routine processing work that juniors traditionally performed. The uncomfortable implication, and the subject of this article, is that those two segments are not independent, because the second has historically been the production pathway for the first.
The Canadian Compensation Numbers
The demand side is well documented in Canadian-specific data. CPA Canada's 2025 Canadian CPA Profession Compensation Study reports that CPAs with three or more years of post-designation experience have median compensation of $154,000, up from $143,000 in 2023[1]. Between 2022 and 2024 that figure climbed 7.7%, outpacing Canada's inflation rate over the same period[1]. CPA Canada frames this directly as evidence of the profession's long-term value and as a signal of the shortage: earnings outpacing inflation is what a labour market does when demand exceeds supply[2].
For a Canadian business owner, this is the practically relevant fact underneath the abstract discussion: the cost of hiring experienced accounting talent has been rising faster than general inflation, and the structural drivers described below suggest that is not a temporary cyclical condition.
How Real Is The Shortage?
The supply-side evidence is substantial though largely US-sourced. The 2024 Finance & Accounting Talent Market Outlook found 83% of senior leaders reporting an accounting talent shortage, up from 70% in 2022[3]. The US pipeline decline is stark in the exam data: first-time CPA exam candidates fell from 48,004 in 2016 to 32,188 in 2021, a 33% decline, with the AICPA reporting a further 7% decrease from 2021 to 2022 reaching the lowest number of exam takers since 2006[4].
Demographics compound this. Approximately 75% of AICPA members are at or near retirement age[4], and current analysis notes the workforce is increasingly concentrated in later-career age groups[5]. The replacement problem is not simply numerical: as commentary on the shortage observes, accounting expertise compounds over time, and replacing a senior tax or audit professional requires years of training, client exposure, and judgment that cannot be rushed[5].
Canada's situation is documented less granularly in the public data reviewed for this article, but CPA Canada acknowledges the Canadian profession faces pipeline challenges and describes a shortage of CPAs[2], while noting that declining enrolments and graduate numbers have been reported worldwide and particularly in the United States, where the AICPA's National Pipeline Advisory Group has called for reforms[1].
The Counter-Trend Most Coverage Misses
Honesty requires giving the contrary evidence equal prominence, because it is genuinely encouraging and it is routinely omitted from shortage coverage. AICPA data shows accounting enrolment at four-year colleges and universities rose 8.9% in spring 2026, marking the third consecutive year of growth[6].
Three consecutive years of enrolment growth is not a rounding error, and it materially qualifies the "structural crisis" framing common in coverage of this topic. What it does not do is solve the near-term problem, for a straightforward reason: an accounting student enrolling in 2026 does not become an experienced, judgment-capable professional until roughly the early-to-mid 2030s. Current analysis makes exactly this point, noting that rebuilding the talent pipeline will take years even with the recent positive trend[6]. The enrolment recovery addresses the inflow. It does not address the training-pathway question this article is actually about, and it arguably makes that question more urgent, since a larger cohort of entrants will need somewhere to learn.
The Stanford Finding
The displacement side has recently acquired a specific, citable empirical anchor. Stanford University researchers found that entry-level workers entering jobs exposed to AI have seen a 13% decline in employment since 2022[3]. This finding is not accounting-specific; it covers AI-exposed occupations generally, and applying it to accounting is an inference rather than a direct measurement, though accounting is uncontroversially among the more AI-exposed professional occupations.
The mechanism within accounting is directly observed rather than merely inferred. Current profession commentary reports that firms are using AI to create efficiency and capacity, and that some firms are reducing entry-level openings as a result of implementing AI, an effort described as creating a new problem for entry-level hiring numbers[3]. This is a rational firm-level decision, and it is the aggregate of many such rational decisions that produces the profession-level problem below.
The Actual Paradox
The core tension is stated with unusual clarity in current profession commentary: not only is it harder to secure entry-level positions, but AI has also taken over the foundational tasks that gave new accounting hires the knowledge to progress to higher-value work[3]. The same analysis describes AI as rendering traditional entry-level training obsolete[3].
This is a genuinely difficult structural problem rather than a transitional inconvenience, and it is worth stating why. The profession's model for producing senior judgment has been essentially an apprenticeship: a junior performs high-volume, low-stakes work under supervision, accumulates pattern recognition through repetition and correction, and gradually earns responsibility for judgment calls. Remove the high-volume, low-stakes work, and the apprenticeship has no substrate. The junior arrives at the judgment-requiring work without having accumulated the patterns that make judgment possible.
The uncomfortable corollary is a timing mismatch. The efficiency benefit of automating entry-level work is immediate and measurable in this year's margin. The cost, a cohort of professionals who reach year five without year-one-through-four's pattern accumulation, arrives years later and is diffuse, unattributable to any specific automation decision, and borne substantially by the profession collectively rather than by the firm that made the efficiency decision. This is a textbook externality structure, and it predicts, correctly, that individual firms will under-invest in the training pathway relative to what the profession as a whole needs.
What The Tedious Work Was Actually Teaching
It is worth being specific about what was actually learned in the work now automated, because vague appeals to "experience" are unpersuasive and invite the reasonable response that tedium has no inherent educational value.
Reconciliation work taught what a normal transaction pattern looks like for a given business, which is the foundation for noticing an abnormal one. Chasing variances taught the ordinary causes of discrepancies, timing, coding errors, unrecorded items, and their relative frequencies, which is the basis for efficient diagnosis later. Preparing schedules taught the structure of the financial statements from the inside out, in a way that reading them from the outside does not replicate. Repeated exposure to many clients' or many periods' records built a comparative baseline, which is what allows an experienced practitioner to look at a trial balance and register that something is off before articulating why.
None of these are things the junior was consciously studying. They were byproducts of volume and correction, which is precisely why removing the volume removes the learning without anyone noticing at the time that education was occurring. The practical implication is that any replacement pathway has to deliberately reproduce what was previously an accidental byproduct, which is harder and more expensive than it sounds.
The Second Shortage: Skills, Not Bodies
A parallel problem compounds the pipeline question and is frequently conflated with it. Current analysis describes firms facing a dual talent crunch: a persistent accountant shortage alongside a critical skills deficit, with leaders increasingly struggling to find professionals who pair core technical accounting with data analytics, AI fluency, and strong advisory skills[6]. The needed blend is described as technical accounting expertise, technology fluency, data skills, business judgment, and client advisory capability[6].
Profession commentary has proposed a corresponding redefinition of the core skill set: what were once considered soft skills, people skills, tech fluency, and analytical thinking, become critical for success in the age of AI and must be fully integrated into accounting and finance education, with firms broadening talent acquisition to hire for accounting literacy plus tech fluency and analytical skills, so newer staff can work confidently alongside AI and engage clients on the reasoning behind the numbers[3].
A Worked Case: The Reviewer Who Couldn't Review
A Canadian mid-market firm implemented AI-assisted transaction categorization and reconciliation across its bookkeeping engagements, reducing junior processing time substantially and, over two hiring cycles, reducing entry-level headcount accordingly. Three years later, the firm encountered a specific problem it had not anticipated: staff at the two-to-three-year mark, who under the previous model would have been ready to take first review on client files, were not.
The deficiency was not attitude or intelligence. These were capable people who had spent their first years on genuinely higher-value work: client communication, advisory support, exception handling. What they lacked was the accumulated baseline that makes review possible, the internalized sense of what a given client's numbers ordinarily look like and therefore what warrants a second look. Asked to review an AI-prepared file, they could confirm the arithmetic was internally consistent and the process had been followed. They could not reliably identify that a plausible-looking result was nonetheless wrong for this particular client, because they had never built the comparative baseline that recognition depends on.
The firm's response was to deliberately reintroduce a structured version of what automation had removed: a rotation in which juniors manually reconstruct a sample of AI-processed files from source documents, not because the manual work is needed for the client deliverable, but because performing it builds the pattern recognition the review role requires. It is training cost with no billable output, deliberately incurred, which is precisely the investment the externality structure described above predicts most firms will not make.
The Barrier-To-Entry Debate
Any discussion of the pipeline eventually reaches the question of whether the profession's own entry requirements are part of the problem, and it is worth handling carefully because it is contested and the Canadian and US positions differ.
In the US, the 150-hour education requirement for CPA licensure is frequently identified as a deterrent, with analysis noting some students perceive the additional educational requirement beyond a standard bachelor's degree as a barrier that discourages candidates unwilling to commit to further study[4]. The same analysis identifies two further attraction problems: a perception of accounting as boring relative to other majors, a misconception it attributes to the assumption that accounting is primarily number crunching and repetitive tasks, and compensation that is often lower than finance or technology at the entry point[4].
The perception problem interacts with automation in a way worth naming directly, because it cuts against the intuitive read. If the deterrent perception is that accounting is repetitive number crunching, then automating the repetitive number crunching arguably improves the profession's attraction proposition rather than damaging it, which may be part of what the three consecutive years of enrolment growth reflect. CPA Canada's own framing points in this direction, identifying opportunities in AI and sustainability as fields that may help draw new entrants[1]. The training paradox and the attraction problem therefore pull in opposite directions: the same automation that damages the apprenticeship pathway may be improving the profession's appeal to people deciding whether to enter it at all.
What Firms Are Actually Being Advised To Do
Current profession guidance converges on several approaches worth noting. Firms are advised to identify future technological and advisory needs and adopt a blended talent model, upskilling existing staff to boost retention while using contract professionals to fill immediate technology or regulatory gaps[6]. On pipeline building specifically, guidance recommends hiring for baseline potential rather than rigid software experience, supporting continuous professional development, and using structured internship programs to convert students into full-time hires early[6].
On the attraction problem, CPA Canada has pointed toward AI and sustainability as fields that may help draw new entrants[1], and the AICPA's National Pipeline Advisory Group has called for reforms making accounting education more engaging and telling a more compelling story about accounting careers[1]. Academic analysis has separately proposed generationally-informed human resource strategies addressing both attraction and retention, noting that different generations enter the workforce with materially different career advancement expectations that recruiting messages should reflect[7].
It is worth observing that comparatively little of this published guidance directly addresses the training-pathway problem this article identifies. Most of it addresses attraction, retention, and skill mix. The question of how a profession produces experienced judgment when the apprenticeship substrate has been automated remains, in the material reviewed here, more frequently named than answered.
What This Means If You Hire An Accountant
For a Canadian business owner, this analysis has three practical consequences worth planning around rather than merely observing.
Experienced accounting talent will likely continue getting more expensive. The compensation data already shows above-inflation growth, and the structural analysis suggests the constraint on experienced practitioners tightens before it loosens, since the enrolment recovery takes roughly a decade to reach the experienced tier.
Years of experience is becoming a less reliable proxy for judgment. If a portion of the current cohort has spent early career years without the pattern accumulation that historically accompanied them, then "five years' experience" carries more variance than it used to. Hiring assessment that probes actual diagnostic reasoning, presenting a plausible-looking but wrong set of figures and asking what seems off, is more informative than it was a decade ago.
The value of an external advisory relationship shifts toward judgment rather than processing. If AI handles processing competently and cheaply, the defensible value of an accounting relationship is precisely the accumulated judgment this article describes as becoming scarcer, which is worth weighing when comparing a low-cost automated bookkeeping option against an advisory relationship on price alone.
The Numbers At A Glance
For quick reference: Canadian CPAs with 3+ years post-designation experience report median compensation of $154,000, up from $143,000 in 2023, a 7.7% climb between 2022 and 2024 that outpaced inflation. Senior leaders reporting an accounting talent shortage: 83% in 2024, up from 70% in 2022. US first-time CPA exam candidates: 48,004 (2016) falling to 32,188 (2021), a 33% decline. Approximately 75% of AICPA members at or near retirement age. Entry-level employment decline in AI-exposed occupations since 2022, per Stanford research: 13%. And the counter-signal: accounting enrolment at four-year institutions up 8.9% in spring 2026, a third consecutive year of growth.
The Limits Of This Analysis
Several caveats matter. Most of the pipeline and displacement data cited is US-sourced, drawn from AICPA and US labour research, and applying it to the Canadian profession is a reasonable but genuine inference; CPA Canada acknowledges Canadian pipeline challenges but the granular Canadian equivalents of the US exam-candidate and enrolment series were not located in the material reviewed for this article. The Stanford 13% figure covers AI-exposed occupations broadly rather than accounting specifically. The enrolment recovery finding is genuinely positive and is presented here at equal prominence rather than buried, and readers should weigh it against the shortage framing accordingly. Finally, the central argument of this article, that automating entry-level tasks degrades the judgment-production pathway, is an analytical claim supported by profession commentary identifying the mechanism, rather than an empirically measured finding; to our knowledge no study has yet quantified a resulting judgment deficit, and it is possible that alternative training pathways will develop faster than this article's framing anticipates.
Frequently Asked Questions
Is there actually an accountant shortage, or is AI replacing them?
What do Canadian CPAs actually earn?
Is the accounting pipeline actually recovering?
What is the "training paradox"?
What should a business owner do differently when hiring?
Will AI eliminate the accounting profession?
References
- CPA Canada. (2025). CPAs Weigh In On How To Attract A New, Much-Needed Crop Of Accountants, reporting 2025 Canadian CPA Profession Compensation Study findings. cpacanada.ca/news/Analysis/pipeline
- CPA Canada. Early-Career CPAs Thrive In A Volatile Job Market. cpacanada.ca/news/accounting/compensation
- CPA Practice Advisor. (2025, October 27). Why The Accounting Profession Needs To Rethink Foundational Skills For Entry-Level Staff, citing the 2024 Finance & Accounting Talent Market Outlook and Stanford University research on entry-level employment in AI-exposed jobs. cpapracticeadvisor.com/.../rethink-foundational-skills
- Angulo, R. Accountant Shortage: Obstacle or Opportunity? Lehigh University, citing AICPA/CIMA annual report data on CPA exam candidate decline. lehigh.edu/~inluac/news/shortage.html
- Ramp. (2026). The Accountant Shortage in 2026: Causes, Impacts & Solutions, citing the 2025 AICPA/NASBA Trends Report. ramp.com/blog/accountant-shortage
- CPA Practice Advisor. (2026, August 3). Beyond The Accountant Shortage: The Skills Shortage Facing CPA Firms, reporting AICPA enrolment growth data. cpapracticeadvisor.com/2026/08/03/beyond-the-accountant-shortage
- The CPA Journal. (2026, February 23). Intergenerational Solutions To Address The Crisis Of The Leaking Accounting Pipeline. cpajournal.com/2026/02/23/intergenerational-solutions-leaking-accounting-pipeline
This article discusses published profession research and labour market data and is provided for general informational purposes. Much of the underlying pipeline data is US-sourced and applied to the Canadian context by inference. It is not hiring, employment, or career advice for any specific situation.