We're Teaching the Wrong Part of the Job
Stop protecting compliance and start protecting the skills clients actually pay for
Every conference I attend, someone eventually asks the same question: “If AI does the compliance work, how will the next generation ever learn?”
It’s a fair concern, but it always makes me wonder if we’re trying to preserve the wrong part of the profession.
Doing the Work Isn’t the Same as Learning the Work
I spent five years in tax compliance before going back for my master’s in taxation. Looking back, I don’t think I learned tax by preparing hundreds of returns. I learned tax when I started understanding why the rules worked and when I could see how changing one fact changed the answer. The thing that helped me the most was listening to experienced professionals explain how they were thinking, not just what they were doing. Those are the things that made everything click.
On the other hand, I’ve also met people who came into the profession with advanced tax degrees, but very little real-world experience. They understood tax code, but they hadn’t learned how to apply that knowledge when a client couldn’t answer every question, when records were incomplete, when two good options existed or when the “right” answer on paper wasn’t necessarily the best business decision.
My path had the opposite problem. I had seen hundreds of returns before I truly understood the theory behind them. I could complete the work, but it took much longer to develop the deeper understanding of why one approach was better than another.
Neither path was complete because one emphasized doing, while the other emphasized knowing. Great advisors need both.
Compliance Was the Exercise, Not the Lesson
That’s why I wonder why we keep framing AI as a threat to learning because it reduces manual compliance work. Was compliance ever the thing we were actually trying to teach? Or was it simply the way we got exposure?
The value wasn’t typing numbers into tax software. It was:
Seeing hundreds of different client situations
Recognizing patterns across those situations
Watching experienced advisors work through ambiguity
Hearing the questions they asked
Understanding why one answer made more sense than another
If the goal has always been to develop professionals who can think critically, advise confidently and solve complex client problems, then perhaps this is our opportunity to redesign how we teach those skills instead of assuming they’ll emerge after enough years of repetitive work.
Instead of asking, “How do we make sure junior accountants still prepare enough tax returns?” maybe we should be asking, “How do we help them develop judgment faster?”
AI Can Be More Than an Autopilot
What if AI didn’t just prepare the return, but it also:
Explained why it classified something a certain way?
Highlighted the assumptions it made and alternatives it considered?
Showed how changing one fact changes the outcome?
Explained why certain language in a client email builds trust?
Surfaced patterns across hundreds of engagements so junior staff can learn to recognize them, too?
What if it became less like an autopilot and more like a mentor sitting beside you?
Protect the Skills Clients Actually Pay For
The skills that create good advisors include critical thinking, pattern recognition, strategic decision-making, client communication and professional judgment.
Compliance has always been one way to develop those skills. AI gives us the opportunity to separate the lesson from the exercise.
The real goal has never been to create accountants who are great at compliance. It’s to create accountants who can use what they know to help clients make better decisions. If AI can take care of more of the mechanics, our responsibility isn’t to recreate the old apprenticeship model just so the next generation learns the way we did. It’s to build a better one.
That means intentionally designing tools and training that develop the skills clients actually pay for: judgment, critical thinking, communication and the ability to turn knowledge into advice.
We don’t need to protect the old work. We need to protect what the old work was supposed to teach us.
