Billable-Hour Economics of Complex Tax Questions
A straightforward 1040 with W-2 income and a standard deduction doesn't require much research. The questions that cost you money are the ones that make you stop, open a database (or worse, Google), and spend 30 to 90 minutes looking for an answer you can defend.
Those questions have a cost, and most practitioners don't calculate it. They should.
What Complex Questions Actually Cost
Let's put numbers on five common research scenarios. I'm using a $300/hour blended rate, which is reasonable for an experienced CPA doing individual and small business work. Adjust for your own rate.
| Research question | Typical time (traditional research) | Cost at $300/hr |
|---|---|---|
| Does this rental activity qualify as a trade or business under IRC 199A? | 45 min | $225 |
| Can the client take the PTE tax election, and does the state conform? | 30 min | $150 |
| Section 1031 exchange boot calculation with mortgage relief and cash received | 60 min | $300 |
| Is this S-Corp shareholder's health insurance deductible under IRC 162(l), and how does it interact with the premium tax credit? | 45 min | $225 |
| Client has income from 3 states, each with different conformity to federal Section 179 limits | 90 min | $450 |
The multi-state question at the bottom is the one that kills solo practitioners. You're not just researching one rule. You're researching three sets of state conformity provisions, three sets of depreciation schedules, and three filing requirements. Each state has its own Department of Revenue guidance, its own conformity date, and its own exceptions.
With citation-backed AI research, those same questions take 3 to 10 minutes. The multi-state question still takes longer because you're verifying three jurisdictions, but 10 minutes versus 90 minutes is a different cost structure entirely.
The Questions That Eat Your Margin
Not all research questions are equal. Simple lookups (what's the standard mileage rate, what's the filing deadline for an extension) cost almost nothing. They take 30 seconds regardless of your tool.
The margin killers are questions with three characteristics:
Multiple authorities. The answer requires cross-referencing IRC sections, Treasury Regulations, and case law. Section 199A is the poster child: IRC 199A itself, Treas. Reg. 1.199A-1 through 1.199A-6, the aggregation rules in 1.199A-4, the safe harbor in Rev. Proc. 2019-38, and income thresholds from Rev. Proc. 2024-40. You can't answer a 199A question from one source.
State-specific variations. Any question involving state conformity multiplies the research scope. "Does my client's state conform?" is never a quick lookup because conformity isn't binary. States adopt federal provisions selectively, with effective dates, carve-outs, and alternative calculations that vary from the federal baseline.
Evolving law. Questions about recently enacted provisions (like the One Big Beautiful Bill Act changes in 2025) require sources that traditional databases may not have indexed yet. If you're searching for guidance on a provision that's less than six months old, the editorial commentary might not exist, and the relevant IRS Notices or Revenue Procedures might be buried in IRB bulletins.
The Recoverable vs. Non-Recoverable Split
Here's the part most practitioners don't think about: not all research time is billable.
If a client comes to you with a specific question, and you spend 45 minutes researching it, you can bill for that time. The research is part of the engagement. Your fee covers it.
But a significant portion of research happens during return preparation, where you discover an issue, research it, and apply the answer without separately billing for the research component. You quoted the client $500 for the return. The 45-minute research question wasn't in that quote. You absorb the cost.
For a solo practitioner preparing 175 returns, the absorbed research time adds up fast. If 40% of your research time is non-recoverable (embedded in fixed-fee returns), you're giving away 34 hours of your season at $300/hour. That's $10,200 in unrecoverable cost.
Faster research doesn't just save time. It shrinks the gap between your quoted fee and your actual cost on fixed-fee engagements.
The Compound Effect Across a Season
Individual questions are expensive. The aggregate is worse.
A solo CPA with 2 complex research questions per day across an 85-day filing season (mid-January through mid-April) handles 170 questions. At an average of 35 minutes per question in a traditional database:
- Total research time: 99 hours
- At $300/hr opportunity cost: $29,750
- Subscription cost (traditional): $3,000 to $8,000
- Combined cost of research: $32,750 to $37,750
The same 170 questions at 5 minutes each with AI-assisted research:
- Total research time: 14 hours
- At $300/hr opportunity cost: $4,250
- Subscription cost (Tax Orator Solo): $948/year
- Combined cost of research: $5,198
The difference is roughly $28,000 to $33,000 per season. That's not a rounding error. That's a staff hire, a vacation, or retained earnings that compound.
When Slower Research Makes Sense
I'm not arguing that every question should go through an AI tool. Some research benefits from the slower, more deliberate process of reading through editorial analysis, tracing legislative history, and sitting with ambiguity before reaching a conclusion.
Tax controversy work is the clearest example. If you're advising a client on an uncertain position that might end up in IRS Appeals or Tax Court, the depth of analysis in a traditional database like CCH or Checkpoint is worth the time investment. You're not trying to answer the question in 5 minutes. You're building a research file that supports your position under examination.
Estate and gift tax planning is another area where speed matters less than thoroughness. The present value of a $15 million estate plan justified spending two hours reading Treas. Reg. 25.2702-3 and related PLRs.
The distinction is between research-as-compliance-step (where speed matters) and research-as-deliverable (where depth matters). Most questions in most practices are the first type.
What This Means for Pricing Your Services
If you know your research cost per question, you can price more accurately.
A return that typically generates 3 complex research questions costs you $675 in research time at traditional speeds (3 x 35 min x $300/hr). That cost should be reflected in your fee. If it isn't, you're undercharging for complexity.
Faster research changes the math. The same 3 questions at 5 minutes each cost $75 in research time. You can either pass that savings to the client (lower fees, more competitive on price) or keep the margin (same fees, higher profit per return). Most practitioners I know choose a mix: slightly lower fees to win clients, with better margins than before.
The Bottom Line
Every research question has a cost: the subscription fee divided across questions, plus your time at your billing rate. For most practitioners, the time component is 5 to 10 times larger than the subscription component. Optimizing for a cheaper subscription while ignoring time-per-question is optimizing the wrong variable.
If you want to run the numbers on your own practice, start with how much tax research actually costs per return. Count your questions, time your process, and multiply. The results will change how you think about research tool selection.
Tax Orator's Solo Practitioner plan is $79/month with 200 queries. For practices where research speed directly affects profitability, the workflow fits solo CPAs and small firms better than traditional databases priced for large firms.