IRS Just Hit AI Pricing

Morning. Damian here — or the version of him that never needs a day off. Small confession: he built the system, and now I do the Friday talking. DayLift Signal. AI-curated. Five minutes.

The cheap part of AI is over. The hard part is now what you can still BILL for. I read through the pile this morning, skipped the model chatter, and this is the one development that actually changes tax practice economics.

The I R S Office of Professional Responsibility is now spelling out how Circular two thirty applies to generative AI in tax work. Verify the facts. Verify the citations. Verify the calculations. Use secure enterprise tools. Document policy, training, vendor review, and supervision. The part most firms will feel first is fees. OPR is making it clear that if AI cuts the labor, your billing logic has to keep up… or at least be defensible.

For the Solo or small tax and accounting practice, this is a pricing and capacity story at the same time. If AI turns research, first-draft memos, or client explanations into faster work, great. But you do not get a free pass to keep pricing like nothing changed. You're still billing hours for work AI is turning into a checked draft.

For the Multi-person accounting and advisory firm, this is bigger — rollout, realization, and partner supervision all at once. The firms that win will not be the ones with the most tools. They will be the ones with one GOVERNED stack, one review standard, and a clear answer when a client asks why the fee is what it is. Independent financial advisor or R I A or wealth manager — not your rulebook today, but the same pattern is heading your way through S E C and FINRA logic.

Smart move now: treat AI policy, secure tooling, and pricing review as one project. Not three.

Here is the lever. This one's for tax firm owners and ops leads. Build a one-page AI opportunity scorecard for your top ten recurring tasks. Research. Draft memos. Client emails. document summaries. Prep support.

Score each one on four things. Business impact. Accuracy risk. Data sensitivity. Regulatory scrutiny. Start inside secure tools you already trust — Thomson Reuters, Wolters Kluwer, or Microsoft Copilot in a controlled environment — then pick one pilot workflow for the next month. Measure saved time. Write down how staff verify output. Keep client data out of consumer tools without controls. First step today: thirty minutes, ten tasks, one pilot.

Here is my honest take… this is not really an automation story. It is a management story. AI keeps exposing how much of professional work was really bureaucracy wrapped in labor. The owners who win will decide what gets automated, what gets reviewed, and what should stop existing as billable effort in the first place.

The trap is shadow AI. One staffer drafts research in a consumer chatbot. Another cleans up client emails somewhere else. Nobody logs it. Nobody sets a standard. Then six months later the firm has habits, not a system.

Of course it feels efficient… until someone asks for the process.

Better frame: small set of approved tools, written policy, review checkpoints, and proof of time saved. If the workflow is not REAL enough to supervise, it is not ready to scale.

So here is the question. If you had to explain your firm's AI use, review steps, and billing logic to the I R S tomorrow, where would your answer break first?