Universities Aren't Training AI Controllers: Should They? Or Is This Just Growing Pains?

General
1

After seeing Alice’s thread about the impossible hire, I got curious and spent a few hours researching what accounting programs are actually teaching in 2026.

What I Found: Traditional Curriculum Hasn’t Changed Much

I looked at accounting degree requirements at 10 major universities (including several “top” programs). Here’s what they ALL teach:

Core Requirements:

  • Financial Accounting (debits/credits, journal entries, financial statements)
  • Managerial Accounting (cost accounting, budgeting)
  • Tax Accounting (individual and corporate)
  • Audit and Assurance
  • GAAP and Financial Reporting Standards

Tech Coverage (if any):

  • Excel (spreadsheet basics)
  • QuickBooks or similar (how to use accounting software)
  • Maybe one “Accounting Information Systems” elective

What’s Missing:

  • Workflow automation and process engineering
  • ML/AI literacy and model evaluation
  • API integration and data pipelines
  • Python/scripting for accountants
  • Version control and collaboration tools

The Question: Should Universities Add “AI Controller” Training?

I can see two very different paths forward:

Path 1: Create Dedicated “Accounting Technology” Programs

Universities could create specialized degree tracks:

  • Accounting Technology BS/MS: Blend accounting fundamentals + computer science
  • Curriculum: 50% accounting (GAAP, tax, audit) + 50% technical (Python, ML, workflow design)
  • Target students: Those who want to be “AI Controllers” from day one
  • Outcome: Graduates ready for hybrid roles, command premium salaries

Pros: Creates pipeline of qualified candidates firms desperately need
Cons: Further fragments already declining accounting enrollment, may cannibalize traditional programs

Path 2: Integrate Technology into Core Accounting Curriculum

Make technical skills part of EVERY accountant’s baseline education:

  • First year: Excel + basic scripting alongside debits/credits
  • Second year: Database queries + data analysis alongside financial accounting
  • Third year: ML basics + workflow automation alongside cost accounting
  • Fourth year: Capstone: build end-to-end automated accounting system

Pros: Every accounting graduate has hybrid skills, no “AI Controller” premium needed
Cons: Makes already difficult degree even harder, could further reduce enrollment

My Tech Industry Perspective

In tech, we went through similar evolution:

1990s: “Web developer” was specialized role (most developers didn’t do web stuff)
2000s: Every developer needed web skills (became baseline, not specialty)
2010s: “Mobile developer” was specialized role
2020s: Every developer needs mobile skills (became baseline)

The pattern: Specialized skills eventually become baseline expectations.

I think “AI Controller” is following same trajectory:

  • 2026: Specialized role, commands premium
  • 2030: Expected for mid-level accountants
  • 2035: Baseline requirement for entry-level positions

If that’s true, universities should integrate (Path 2) rather than create separate track (Path 1).

But Here’s the Counterargument

CPA candidates are down 27% over past decade. Accounting graduates fell 6.6% in 2023-2024.

Accounting programs ALREADY struggle with:

  • 150-credit CPA requirement (5-year burden)
  • Perception of “boring” career
  • Better-paying alternatives (tech, finance, consulting)

Adding MORE technical requirements might:

  • Make degree even harder (scare away marginal students)
  • Require faculty retraining (most accounting professors can’t code)
  • Increase costs (need computer science resources)

Could backfire and further reduce enrollment.

Alternative: Industry Training vs University Training

Maybe the answer isn’t universities at all. Maybe it’s:

  1. Bootcamps: 12-week intensive programs teaching automation to working accountants
  2. Continuing education: CPE credits for learning Python, ML, workflow design
  3. Employer training: Firms invest in upskilling current staff (like tax_tina suggested)
  4. Self-teaching: Platforms like Beancount + online courses let motivated people learn on their own

This keeps universities focused on fundamentals, pushes technical skills to where people already work and understand context.

Questions for the Community

I’m genuinely curious what you all think:

  1. Should universities add AI/automation to accounting curriculum? Or is this skill better learned on the job?

  2. If universities DO add it, should it be separate track or integrated?

  3. Will adding technical requirements REDUCE enrollment further (by making degree harder)?

  4. Or will it INCREASE enrollment (by making graduates more employable)?

  5. Should CPA exam test technical skills? If not, what incentive do students have to learn them?

From a tech person watching the accounting industry, it feels like you’re at an inflection point similar to where software development was in the 1990s when the web emerged.

Curious how accounting professionals think this should evolve.

Sources:

2

Great research, Fred! I strongly vote for Path 2: Integration into core curriculum.

Why Separate Tracks Don’t Work

Creating a dedicated “Accounting Technology” track has several problems:

  1. Stigmatizes traditional accounting: Implies “regular” accountants don’t need tech skills (wrong message to send in 2026)

  2. Creates talent bifurcation: You end up with two types of graduates - technical hybrids and traditional accountants. The traditional ones become unemployable in 10 years when baseline expectations shift.

  3. Enrollment cannibalization: Students choosing tech track means fewer in traditional track, accelerating the decline problem

Integration is the Only Sustainable Answer

I think the Excel comparison is perfect. In 1995, universities could have created “Excel Accounting” specializations. But instead, Excel became integrated into EVERY accounting course:

  • Financial Accounting: build balance sheets in Excel
  • Cost Accounting: model job costing in Excel
  • Tax: calculate deductions using Excel formulas
  • Audit: analyze datasets with pivot tables

By 2005, every accounting graduate knew Excel because they used it throughout their education, not because they took one “Excel elective.”

Same approach should apply to Python/automation:

  • Intro Accounting: Learn debits/credits by writing Python scripts that validate journal entries
  • Financial Accounting: Build financial statement generator using plain text ledgers (hello Beancount!)
  • Cost Accounting: Write importers that pull data from manufacturing systems
  • Tax: Create automation that flags potential deductions based on transaction patterns

The Faculty Challenge

You’re right that most accounting professors can’t code. But:

  1. Partner with CS departments: Co-teach courses, blend domain expertise
  2. Hire hybrid faculty: New PhD programs should require both accounting + programming
  3. Industry practitioners: Bring in professionals who use these tools daily (many would love to teach)

Regarding Enrollment Concerns

I actually think integration could INCREASE enrollment by making accounting more appealing:

Current perception: “Accounting is boring data entry that will be automated away”

New perception with tech integration: “Accounting is designing intelligent systems that automate financial operations” (sounds way cooler!)

Students choosing CS/data science might consider accounting if curriculum included automation, ML, and workflow design. It’s a real differentiator from pure tech roles - you’re solving business problems with coding, not just coding for its own sake.

What About CPA Exam?

This is the forcing function. If CPA exam doesn’t test technical skills, universities won’t prioritize teaching them.

AICPA should update CPA exam to include:

  • Basic scripting/automation section
  • Data analysis and interpretation
  • AI/ML literacy (not building models, but evaluating outputs)

Once exam tests these skills, every program will teach them (just like they all teach audit procedures because it’s on the exam).

Fred, your tech industry perspective is valuable. How long did it take for “web development” to go from specialized to baseline in CS programs? I’m guessing 5-10 years?

3

I appreciate Fred’s research and Alice’s response, but I want to offer a more cautious perspective.

Universities Should Teach Fundamentals Well, Not Chase Every Trend

My concern with both Path 1 (separate track) and Path 2 (integration) is they assume “AI Controller” skills will remain relevant long-term.

But what if the tools change completely in 5 years?

Consider:

  • 2015: Everyone learned QuickBooks Desktop in school → Most professionals now use QBO (different paradigm)
  • 2018: Excel macros were cutting-edge automation → Now people use Python/RPA instead
  • 2022: Everyone rushed to learn GPT-3 prompt engineering → GPT-4 made most of that knowledge obsolete

If universities redesign curriculum to teach “Python + Beancount + ML categorization” in 2026, will those specific skills matter in 2031 when graduates enter workforce?

What Lasts: Principles, Not Tools

Instead of teaching specific technical skills, universities should focus on principles that transcend tools:

  1. Double-entry thinking: Understanding why accounting systems must balance, regardless of software
  2. Data integrity: Recognizing data quality issues, validation rules, reconciliation needs
  3. Process design: Mapping workflows, identifying automation opportunities, exception handling
  4. Critical evaluation: Questioning outputs, understanding limitations, professional skepticism

These principles apply whether you’re using:

  • Beancount or QuickBooks or [future tool that doesn’t exist yet]
  • Python or [future language]
  • Current ML models or [future AI systems]

Students Can Learn Tools Later

I learned accounting fundamentals in university (1990s). Never touched Python in school.

When I needed Python for Beancount (2020), I taught myself using online resources. Took 3-4 months of evenings/weekends.

Was that hard? Yes, but manageable because I had strong accounting foundation.

Would I have preferred learning Python in university? Not really - the Python I learned in 2020 is different from Python in 1995. If my university taught me programming, it would’ve been outdated anyway.

The Self-Teaching Alternative

I think Fred’s point about industry training, bootcamps, and self-teaching is the right model:

Universities should:

  • Teach accounting fundamentals REALLY WELL (better than they do now, honestly)
  • Expose students to automation concepts (show them it exists, explain why it matters)
  • Encourage technical curiosity (“if you want to learn coding, here are resources”)

Then let:

  • Employers provide specific technical training (they know their tech stack)
  • Bootcamps teach working professionals who understand accounting context
  • Self-learners use platforms like Beancount to develop skills at their own pace

Don’t Make Degree Harder When Enrollment is Declining

Alice thinks technical requirements might INCREASE enrollment. I’m skeptical.

Students considering accounting are generally:

  • Risk-averse (want stable career)
  • Not naturally technical (many specifically AVOIDED tech/CS)
  • Intimidated by 150-credit CPA requirement already

Adding Python/ML requirements risks:

  • Scaring away these students → Further enrollment decline
  • Overwhelming those who enroll → Higher dropout rates
  • Creating barrier where none needed → Reduces diversity (privileges students with tech background)

My Recommendation

Keep core accounting education focused on accounting. Teach principles that transcend tools. Let technical skills be learned:

  • On the job (employer-sponsored training)
  • Through continuing education (CPE credits)
  • Via self-teaching (motivated people will learn)

This approach:

  • Preserves university role (teaching fundamentals)
  • Adapts to changing tools (principles outlast specific technologies)
  • Maintains accessibility (doesn’t add barriers that reduce enrollment)

I know this is less exciting than “every accountant will code!” But I think it’s more sustainable long-term.

4

Practical question from someone who hires: If universities add Python/automation requirements, does that make graduates MORE expensive to hire?

The Cost Concern

Right now I hire accounting graduates at $45k-55k for entry-level bookkeeping positions (Austin, TX market).

If those same graduates now have:

  • Accounting degree
  • Python skills
  • ML literacy
  • Workflow automation experience

Won’t they expect $70k-80k starting salary? That’s market rate for junior developers with business knowledge.

And if they have those skills, why would they work for a small bookkeeping practice when tech companies will pay more?

The Small Practice Dilemma (Again)

This connects to my concern from Alice’s thread: large firms can afford to hire graduates with premium skills at premium salaries. Small practices like mine can’t.

If universities train accounting graduates with strong technical skills:

  • Winners: Large firms, Big Four, enterprise companies (can afford to pay for hybrid talent)
  • Losers: Small practices, solo practitioners, local CPA firms (get priced out of hiring qualified people)

Maybe Different Training for Different Career Paths?

What if instead of one-size-fits-all, we had:

Path A - Traditional Accounting (for small practice/bookkeeping roles):

  • Core accounting fundamentals
  • Excel proficiency
  • Basic software usage (QuickBooks, Xero)
  • Client relationship skills
  • $45k-60k starting salary

Path B - Accounting Technology (for corporate/automation roles):

  • Same fundamentals + technical skills
  • Python, APIs, ML, workflow design
  • Process engineering focus
  • $70k-90k starting salary

This way:

  • Students choose based on career goals + aptitude
  • Employers know what they’re getting
  • Salary expectations align with skills

Alternative: Make Technical Skills Optional

Universities could offer:

  • Required core: Traditional accounting curriculum (everyone takes)
  • Optional electives: Programming, automation, ML (for interested students)

This way:

  • Students with technical interest can pursue it
  • Students without technical aptitude aren’t forced into it
  • Small practices can still hire entry-level people at affordable rates
  • Large firms can recruit from the subset with technical skills

The Reality Check

I realize this is unpopular opinion, but: Not every accountant needs to code.

Someone processing payroll for 5 small businesses doesn’t need Python skills. They need to:

  • Understand payroll regulations
  • Use payroll software correctly
  • Catch errors and discrepancies
  • Communicate with clients

Should they be aware automation exists? Yes.
Should they know how to use basic tools? Yes.
Should they write custom importers in Python? Probably not necessary.

Maybe the answer is tiered expectations rather than “everyone must have all skills.”

5

The elephant in the room: CPA exam doesn’t test technical skills, so universities have no incentive to teach them.

Professional Licensure Drives Curriculum

Let’s be honest about why accounting programs teach what they teach:

Programs are designed to prepare students for CPA exam.

CPA exam tests:

  • Auditing and Attestation (AUD)
  • Business Environment and Concepts (BEC)
  • Financial Accounting and Reporting (FAR)
  • Regulation (REG - tax and business law)

CPA exam does NOT test:

  • Python programming
  • ML model evaluation
  • API integration
  • Workflow automation

So why would universities invest resources teaching skills that don’t help students pass the licensing exam?

Follow the Incentives

If AICPA wants accountants to have technical skills, they need to:

  1. Add technical section to CPA exam (even basic - write simple Python function, interpret ML confidence scores, design workflow diagram)

  2. Make it count (can’t just be optional or tiny weight - needs to be 10-15% of score)

  3. Provide study resources (Becker, Roger CPA, etc. will create prep materials)

Once it’s on the exam:

  • Universities WILL teach it (students demand prep for exam sections)
  • Professors WILL learn it (can’t teach what’s on exam without knowing it)
  • Students WILL study it (can’t pass exam without it)

Without Exam Changes, Nothing Will Change

Fred’s research shows universities aren’t teaching these skills. Alice argues they should integrate them.

But universities won’t change unless forced to by licensure requirements.

That’s just reality of how professional education works.

My Prediction

What will actually happen:

  • Short term (2026-2029): Universities don’t change curriculum, continue teaching traditional accounting
  • Market pressure (2029-2032): Employers increasingly complain graduates lack technical skills, hiring difficulty continues
  • AICPA responds (2033-2035): Finally updates CPA exam to include technology/automation section
  • Universities adapt (2036-2040): Curriculum changes to prepare for new exam format

So we’re looking at 10-15 year lag between when industry needs skills and when education system delivers them.

What Can Be Done Now?

Since universities won’t change quickly:

  1. Continuing education: Offer CPE credits for technical skills (Python, automation, ML literacy)
  2. Industry certifications: Create “Certified Accounting Technologist” credential separate from CPA
  3. Employer training: Firms invest in upskilling current staff rather than waiting for universities
  4. Bootcamps for accountants: 12-week intensive programs teaching automation to practicing CPAs

These can move faster than universities + licensure system.

Bottom line: As long as CPA exam doesn’t test technical skills, universities won’t prioritize teaching them. Change the exam first, curriculum will follow.