Early in my career—long before AI, cloud ERPs, or any notion that accounting could be automated out of existence – I found myself staring at a wall of inventory reports at Johnson & Johnson. Finished goods. LIFO. Dozens of layers accumulated year after year because inventory levels always seemed to rise. Cost indexes went up, layers piled on, and the whole thing became a kind of archaeological dig through decades of cost inflation.

Then our division decided to do something unusual: clean up our warehouses.

This wasn’t an accounting event; it was an operations initiative. Reduce working capital, free up cash, improve turns. Basic blocking and tackling. But for the first time in many years, inventory didn’t go up. It went down. And that triggered something I had only seen in textbooks: a LIFO liquidation.

When inventory quantities fall, you peel back LIFO layers. Not metaphorically—literally. You liquidate the most recent layers first, and if the reduction is large enough, you unearth older layers priced at much lower historical costs. That’s the rule. There’s no smoothing, no averaging, no managerial judgment. It’s mechanical GAAP.

And sure enough, when I ran the roll-forwards: COGS dropped, margins jumped, and income rose sharply.

Operations cheered. Management blinked. Auditors frowned.

I wrote white papers defending the accounting – layer by layer, index by index. I walked the audit team through ASC 330 (then ARB 43, APB 29). I demonstrated that the income wasn’t an “error,” “timing issue,” or “model bug.” It was simply LIFO functioning exactly as designed. Reduce quantities → liquidate older, cheaper layers → lower COGS → higher income. Full stop.

That was my initiation into technical accounting. I wasn’t just booking entries anymore; I was explaining why the entries existed. I was translating operational decisions into financial consequences.

Pretty cool.  The quiet technical accountant was asked to put down the books, set aside technical jargon and present what the heck was going on. Surprise a roomful of disinterested business owners, doing a presentation showing how accounting is important and even interesting. Bet your AI Bot cannot do that without human assistance!  Accounting was the driver of big EBITDA gains. Do we tell this to prospective students shying away from a career in accounting?

Somewhere in that moment, someone called me a “technical accountant.” I call it more casual technical accountant – there are many much more technically astute and deserving of the technical label. But it stuck.  And it played a role in career shifts from Inventory to Revenue Recognition, Acquisition Accounting, SOX and Internal Controls across Services, SaaS and now ARM.

The interesting irony: the exact kind of casual technical work that built my reputation is now the part of the profession being automated fastest.  Or is it?

The Technical Accountant Then: Research, Codification, Memos

For most of the past few decades, technical accounting meant:

• Interpreting the codification
• Drafting memos
• Defending positions to auditors
• Navigating gray areas in ASC 606, 842, 805, 740
• Building models (often manually)
• Validating complex roll-forwards (like LIFO)
• Designing controls around those judgments

It was specialized, it was detail-oriented, and it rewarded people who were patient enough to think in footnotes and subparagraphs. It was a discipline.

But the profession is shifting. This is not just FCPA, SOX or similar sea changes of the past.  No denying – AI is a fundamental shift.

The Technical Accountant Now: Still Needed, but Changing Shape

In seconds, AI can now do:

• Codification searches
• Policy comparisons
• Revenue recognition frameworks
• Inventory costing models
• Error-free LIFO layer roll-forwards
• Draft memos
• Disclosure checklists

I’m not minimizing the work. I’m just acknowledging what’s happening: the “mysterious technical layer” has become reproducible.

So where does the value move? Into the spaces where AI can’t operate alone – places where context, judgment, governance, and cross-functional understanding matter more than memorizing ASC 330.

Let me use an example. Today I work in one of the simplest revenue models you can imagine: collections. We earn a contingency commission rate on dollars collected. Straightforward. ASC 606 should apply cleanly. Single performance obligation. Collect money. When collection happens, revenue happens. Simple.

Except—it wasn’t. Buried deep in our system was a design flaw. The system priced placements upon placement, not upon collection. This worked fine years ago when pricing was fixed. But as the business introduced variable pricing, the system started producing incorrect revenue and lots more work to get revenue recognition right.

In other words, the technical problem wasn’t GAAP. It was the system’s simplified interpretation of reality and how it fit GAAP.

Another example, much more relevant lies in one of the fastest-growing specialties.

Systems & AI Governance: The New Frontier of Technical Accounting

This didn’t exist when I was reconciling LIFO pools. Today it’s becoming essential.  What does it entail?

1. Ensuring AI follows GAAP correctly

AI is now generating:

• revenue schedules
• lease classifications
• inventory valuations
• COGS calculations
• impairment indicators

Someone must verify that the AI applies GAAP principles consistently; decisions are explainable, model outputs match policy, and the system can withstand audit scrutiny.  Hallucinations are real. Be very careful. Take the warnings at the bottom of every Chat site seriously. There is an old maxim that applies: to err is human; to really foul things up requires a computer.

This isn’t “research.” It’s governance.

2. Designing controls over AI-assisted financial reporting

SOX is concerned with process and in that sense eternally essential.  But where once it revolved around human approvals, now it revolves around what the model did; why it did it; how exceptions were handled; how drift is detected, and who approves AI logic changes.

It’s ICFR (Internal Control over Financial Reporting) evolving into ICAFR – Internal Controls over AI-Assisted Financial Reporting.

3. Understanding how data flows through systems

Accounting teams have always been concerned with process – can we rely on internal controls? More than ever now with AI, accountants increasingly have to understand how ERPs feed data into AI models; where accounting logic resides; how data transforms across environments; how to detect when something breaks or misaligns.

It’s accounting + systems architecture + risk management.

4. Governance and disclosure

Companies will soon need to disclose where AI is embedded in reporting; what judgments it influences; what oversight exists; how risks and biases are mitigated. The footnotes to the financials are the richest element in financial statements.

This is not all traditional accounting work, but it draws heavily on the mindset technical accountants have always had.

So What Happens to Technical Accounting?

It doesn’t disappear. It evolves. Fewer roles will be pure GAAP research. More will be:

• operational finance partners,
• AI governance specialists,
• data-aware accountants,
• integration leads,
• SEC reporting experts,
• tax structuring advisors,
• M&A and purchase accounting specialists.

In other words, the fun stuff! The work moves up the value chain. We need to make sure aspiring accountants understand this.

The same way legal research was automated, accounting research is following. And just as lawyers shifted toward judgment, negotiation, and advocacy, accountants are shifting toward systems, interpretation, design, and oversight.

Looking Back at That First LIFO Memo

When I think back to that first LIFO liquidation memo—the one that “made me a casual technical accountant” – I realize it represented something bigger than I understood at the time.

I was doing three things:

1. Explaining technical mechanics
2. Translating operations into financial consequences
3. Defending the reasoning to auditors

Today, AI can handle #1. But #2 and #3? That’s still very human, to say nothing of the judgements and fuzzy logic required in forecasting.

And that may be the best way to explain where the profession is going. The technical is becoming automated. White collar focused on just that will perish, as will the monotony of posting journals, reconciliations and the like – Thank Goodness! Embrace it.  Therein lies Accounting’s next great opportunity.