Building a Research Playbook for Your First Staff Accountant

A written research playbook gets a new hire to competent tax research output in weeks instead of months. The playbook has five components: how to frame the question, where to look first, how long to spend before escalating, what a complete answer looks like, and how to document the conclusion. Most firms skip all five and hope that whoever picks up the file will figure it out through repetition. That approach produces inconsistent work product, wasted hours, and research that has to be redone before it can be used. Documenting your process once saves you from re-explaining it on every question for the rest of the firm's life.

Why a Playbook Exists

Every research question has a time cost. At a $300/hour billing rate, a question that takes 45 minutes costs $225 in practitioner time. If that question gets answered twice because the first attempt was incomplete, you've spent $450 on a $225 problem. The cost per return analysis makes this visible at the season level, but the waste happens at the individual question level.

The quality argument matters more. Treas. Reg. 1.6694-1 defines the due diligence standard for paid preparers. That standard applies to everyone who touches the return, regardless of tenure. A new team member who signs off on a position without adequate research exposes the firm to the same penalties as anyone else would. The playbook isn't about controlling people. It's about giving every practitioner a clear standard for what "adequate research" looks like before a position goes on a return.

Treas. Reg. 1.6695-2 reinforces this for returns claiming credits like the EITC, CTC, AOTC, and head of household status. The due diligence checklist on IRS Form 8867 is a practical example of what structured documentation looks like for specific claim types. Your playbook should produce that level of rigor for every research question, not just the ones requiring a form.

Component 1: The Question Intake Template

Research quality starts before anyone opens a database. The most common reason research takes too long is that the question was never defined clearly. "Can the client deduct this?" is not a research question. "Can a single-member LLC owner deduct home office expenses under IRC 280A when the LLC is disregarded for federal purposes and the taxpayer also has W-2 income from an unrelated employer?" is a research question.

Your intake template should capture six things every time:

1. Client facts. Entity type, filing status, state of residence, relevant amounts, any unusual circumstances.
2. The specific tax issue. Which IRC section, credit, deduction, or filing requirement is in question? "Can they deduct this" becomes "Does this qualify under Section X?"
3. The tax year. Rules change. Thresholds adjust. A 2025 question answered with 2024 authority is a wrong answer.
4. Prior research. Has anyone at the firm looked at this issue before? Checking existing work product before starting fresh saves the most time of any single step.
5. Deadline. A return due in two days needs a different research depth than a planning question with a six-month runway.
6. Client communication. Has the client been told we're researching this, or does the answer need to come back as part of the return without a separate conversation?

Make this template the required starting point for every research request. The five minutes it takes to fill out saves 20 minutes of false starts.

Component 2: The Source Hierarchy

Treas. Reg. 1.6662-4(d)(3) provides the actual legal authority hierarchy that determines whether a tax position has "substantial authority." Your playbook should mirror this hierarchy because it is the standard the IRS uses to evaluate positions under the substantial understatement penalty.

The order matters:

1. Internal Revenue Code. The statute itself. Always start here. If the IRC answers the question directly, you're done.
2. Treasury Regulations (final and temporary). Regulations interpret the Code. Both are primary authority, though final regulations carry more weight than temporary ones.
3. Revenue Rulings and Revenue Procedures. Published IRS guidance applying the Code and Regulations to specific fact patterns.
4. Tax Court and other federal court opinions. Court decisions interpreting the Code. Tax Court is especially relevant because most individual and small business disputes start there.
5. Private Letter Rulings, Technical Advice Memoranda, and Chief Counsel Advice. These apply only to the taxpayer who requested them, but they show how the IRS applies the law to specific facts.
6. IRS Publications, Instructions, and other informal guidance. Plain language, but not binding authority. A starting point, not an endpoint.

Every practitioner on your team should know this order. When someone says "I found the answer in a Publication," the follow-up is always: "What does the Code say? What do the Regulations say?"

A common mistake is stopping at the first source that answers the question. If someone finds a Revenue Ruling supporting the client's position, the research feels complete. But a later-enacted regulation or court opinion that narrows the ruling changes the answer. The playbook should state: confirm the primary authority (Code and Regulations) before relying on any secondary source.

Component 3: Time-Boxing and Escalation

Research without a time limit is a budget without a cap. Without guidance, every question expands to fill the time available. Your playbook should set three boundaries:

15 minutes for a simple lookup. Standard deduction amounts, filing thresholds, contribution limits, form instructions. If the answer isn't clear within 15 minutes, the question is more complex than it appeared. Reclassify it.

45 minutes for a moderate question. Single-authority interpretive questions: does this activity qualify, does this deduction apply, what's the correct treatment. If you haven't found a defensible answer in 45 minutes, it's time to escalate.

90 minutes maximum before escalation. Multi-authority, multi-jurisdiction, or novel questions. After 90 minutes, write up what you've found and bring it to whoever can make the judgment call. The billable-hour economics of a 90-minute question at $300/hour are $450. Past that point, you need a decision about whether to continue, change approach, or get an outside opinion.

Escalation is not a failure. Escalation with a summary of what you've already checked is efficient. Escalation with "I couldn't find anything" is expensive because whoever picks it up next starts from zero.

The escalation memo should include: the question as framed, sources checked, what you found, what's unresolved, and your preliminary conclusion. This becomes part of the research file whether the question gets answered in-house or goes to an outside specialist.

Component 4: The Quality Check

Before any research conclusion reaches a client, it passes through a quality check. This isn't a review by a second person, though that's ideal. It's a self-check that whoever completed the research performs before marking the question as answered.

1. Does the answer cite primary authority? It needs a Code section, a regulation cite, or a court opinion. An answer based solely on an IRS Publication is not complete.
2. Is the authority current? Check effective dates. Check whether the regulation has been amended or the Revenue Ruling superseded. Tax law from 2022 might not apply to a 2025 return.
3. Does the answer address the client's specific facts? "Section 199A allows a QBI deduction" is a true statement. "Client's rental activity qualifies as a trade or business under Treas. Reg. 1.199A-1(b)(14) based on these facts" is a research conclusion. Only the second one is useful.
4. Would this hold up under examination? If the IRS questions this position, can you point to the authority and explain why it applies? If not, the research is not done.
5. Is the conclusion documented? The research file should contain the question, the facts, the authorities consulted, the conclusion, and the date.

These five questions take two minutes. They prevent the most common failures: unsupported positions, outdated authority, and conclusions that don't connect to the client's situation.

Component 5: Tool Setup

The best playbook is useless if your team's tools don't match the process.

For simple lookups (15-minute questions): AI-assisted research tools handle these fastest. Type the question, get a cited answer, verify the citation. The source hierarchy still applies, but the tool does the initial search in seconds instead of minutes.

For moderate questions (45-minute questions): Start with AI-assisted research to identify relevant Code sections, regulations, and published guidance. Then read the primary authority directly. For CPA firms running a mix of individual and business returns, this is where 70% of research time goes.

For complex questions (90-minute questions): Use every tool available. AI research for the initial framework, traditional databases for editorial analysis and legislative history, court opinion databases for precedent.

Document which tools your firm has, how to access them, and what each is best at. New team members shouldn't have to guess whether the firm has a Checkpoint subscription or find login credentials on their first research question.

Making It Stick

A playbook that lives in a shared drive folder and never gets referenced is a documentation exercise, not a process improvement. Three things make it stick.

Use it during onboarding. Walk through real research questions from past engagements using the playbook structure. Show the intake template filled out, the source hierarchy in action, the time-boxing in practice, and the quality check at the end. Real examples beat abstract instructions.

Reference it in review. When reviewing work product, point to specific playbook steps. "The quality check asks whether the authority is current. This ruling was superseded in 2024." This connects the playbook to actual work outcomes.

Update it annually. Tax law changes. Your client base changes. Your tools change. Review the playbook at the start of each filing season and remove references to tools you no longer use. A playbook from 2024 that references a discontinued subscription erodes trust in the entire document.

The Return on Documentation

If your firm handles 600 research questions per year and the playbook saves 10 minutes per question by eliminating false starts and incomplete first attempts, that's 100 hours. At $300/hour, that's $30,000 in recovered practitioner time. The playbook takes half a day to write.

That 10-minute savings compounds when you think about how to price CAS engagements that include research. If your research process is predictable, your pricing can be too.

We built Tax Orator to fit into this kind of structured process, handling source identification and citation in seconds so practitioner time goes to analysis and judgment. See the plans and pricing and decide whether it fits your playbook.