When Documentation Fails: What Two Active Tax Court Cases Tell CFOs About R&D Credit Substantiation

For CFOs and finance leaders who rely on the IRC §41 Research and Development Tax Credit to generate meaningful working capital, the message coming out of the U.S. Tax Court in 2025 and 2026 is worth your attention. The IRS has raised its expectations around substantiation, and two active cases make clear that how you document your qualified research activities matters just as much as whether those activities occurred.

Understanding what the IRS is now looking for, and where taxpayers are falling short, is the first step toward building an R&D credit claim that can withstand examination.

A Shift in How the IRS Values Substantiation

Historically, many R&D tax credit studies leaned heavily on subject matter expert (SME) testimony. After a tax year closed, companies would bring in outside advisors to interview engineers, scientists, and technical staff, then compile a study report estimating how employees spent their time on qualified research activities.

That approach is now under heightened scrutiny. According to two pending cases before the U.S. Tax Court Kyocera v. Commissioner and George v. Commissioner the IRS is placing significantly less weight on SME interviews conducted after the fact and is instead expecting taxpayers to demonstrate qualified research through contemporaneous documentation: records created while the research was actually being performed.

This is not a minor procedural preference. In both cases, the absence of contemporaneous records proved costly.

Case One: Kyocera v. Commissioner

Kyocera Corporation is a Japan-based multinational manufacturer of industrial ceramics and electronics. Here is what happened:

Kyocera originally claimed an R&D tax credit of approximately $400,000 on its 2018 U.S. tax return.
More than 16 months after the tax year ended, its accounting firm conducted a retrospective R&D tax credit study, interviewing 36 employees identified as subject matter experts.
The study concluded that over 1,200 Kyocera employees had engaged in qualified research, and the company amended its return, increasing the claimed credit to approximately $1.7 million.
The problem: Kyocera employees had kept no timecards or contemporaneous records of any kind. The entire credit rested on retrospective estimates drawn from SME interviews.
Upon audit, the IRS denied the entire $1.7 million claim, stating in a court filing that because the study was based on mere estimates, the Tax Court should decline to recognize any portion of the credit.
The case remains pending before the U.S. Tax Court.

For companies that manage large employee populations across multiple functions, this case is a direct warning against relying on post-hoc allocation of research time without real-time substantiation.

Case Two: George v. Commissioner

Georges of Missouri, Inc. (GOMI) is a large U.S. poultry processing company. Here is how its case unfolded:

Between 2012 and 2014, GOMI conducted seven research trials related to broiler chicken production.
A consulting firm later identified nearly $63 million in qualifying supply expenses over the three-year period, resulting in an R&D credit claim exceeding $4.5 million. No credit was claimed for wages or contract research only supplies.
The IRS disallowed the credits in full and assessed accuracy-related penalties.
The U.S. Tax Court partially sided with the taxpayer, allowing credits that could be substantiated through a combination of contemporaneous records and supporting documentation from the credit study. Accuracy-related penalties were waived.
The court described GOMI’s claim as a “mixed basket of eggs” some credits supported by contemporaneous records, others that could not be substantiated despite GOMI being a highly data-driven business with extensive operational documentation.
In several instances, the court was unable to find sufficient evidence that specific research trials occurred, or that costs connected to those trials satisfied the four-part statutory test under IRC §41.

The lesson here is not that supply costs cannot qualify the court confirmed that they can. The lesson is that documentation must connect specific expenses to specific qualified research activities, and that connection must be established through records created at the time, not reconstructed afterward.

Side-by-Side: What These Two Cases Teach CFOs

Factor	Kyocera v. Commissioner	George v. Commissioner
Industry	Electronics / industrial ceramics	Poultry processing / agriculture
Original credit	~$400,000	Credits for 2012, 2014, and 2016 tax years
Amended / claimed credit	~$1.7 million (amended after retrospective study)	~$4.5 million in supply QREs
Contemporaneous records	None. Zero timecards or real-time documentation of any kind.	Partial. Some trials well-documented; others had gaps or contradictions.
SME study approach	36 employees interviewed 16+ months after year-end; 1,200+ employees identified as qualified researchers	Extensive on-site visits and interviews by experienced consulting firm with documented methodology
IRS position	Denied entire $1.7M entire claim based on estimates	Denied all credits; sought accuracy-related penalties
Court outcome	Pending IRS argues no portion should survive	Partial win: 4 of 7 trials allowed; penalties waived for good-faith reliance
Core lesson	No contemporaneous records = no defensible credit, regardless of study quality or size	Records created at the time of research are the deciding factor not the sophistication of the credit study

Three Things Finance Leaders Should Take Away

Contemporaneous documentation is no longer optional: SME interviews remain useful to contextualize primary records and fill narrow gaps. But as both cases illustrate, they cannot serve as the primary foundation for an R&D credit claim. The IRS and Tax Court will look for records created while the research was conducted on technical logs, project plans, testing data, timecards, and experiment documentation. Studies assembled 12 to 18 months after the fact are inherently limited by what employees can accurately recall.
Proactive planning changes the outcome: In both Kyocera and George, the companies completed their research activities before giving serious thought to credit documentation. Studies conducted 12 to 18 months after the fact are inherently limited by what employees can accurately recall. Beginning the credit study in the same year, the research is conducted produces more reliable SME input and demonstrates that the company approached the credit as a compliance exercise, not a retrospective estimate.
Supply cost claims require traceability: George reinforces that supply expenses can qualify, but only when the taxpayer can trace each cost to a specific, qualified research activity. Without that traceability, the IRS will challenge the connection, and the Tax Court will not fill the gap with inference alone.

Documentation Standard	What It Looks Like in Practice	What Happens Without It
Contemporaneous records as the primary foundation	Technical logs, project plans, timecards, testing data created during the research period, not assembled afterward	Kyocera: IRS argues entire $1.7M claim is based on estimates and should receive zero credit
Expense-to-activity traceability for supply costs	Each supply cost linked to a specific qualified research activity through records created at the time	George: three trials disallowed because supply costs could not be traced to specific documented research activities
Internal consistency between records and claims	Documentation accurately reflects what the research actually was including dosages, protocols, and experimental variables	George: one trial disqualified because GOMI’s own feed records showed the tested dosage never changed
Good-faith reliance on qualified professional advice	Credit study conducted by experienced professionals with documented methodology and on-site verification	George: penalties waived because GOMI relied on professional advice. Kyocera’s reliance on retrospective estimates without primary records provides no equivalent protection.

How TaxDrone.AI Supports the Documentation Standard the IRS Now Expects

The pattern that emerges from Kyocera and George is that the companies best positioned to sustain their R&D credits are those that capture qualifying information in real time, systematically, and in a format that connects financial data to technical activity.

TaxDrone.AI is designed precisely for that challenge. By integrating with existing project management, engineering, and accounting systems, the platform enables companies to capture qualified research data closer to when it is generated supporting the type of contemporaneous record that both the IRS and the Tax Court have now made clear they expect to see. Automated data capture, structured project tracking, and statistical sampling tools help eliminate the reliance on retrospective estimation that made the Kyocera claim indefensible.

Software alone, however, does not interpret tax law or prepare a company for examination. National Tax Group (NTG) provides the advisory expertise that sits behind TaxDrone.AI from methodology design and project qualification under §41, to documentation review and audit defense. The combination of structured technology and experienced professional judgment is what converts a well-intentioned credit claim into one that is genuinely audit-ready.

The Takeaway for R&D-Intensive Businesses.

The IRS has not abandoned the R&D tax credit. It has raised the bar for how that credit must be supported. The value of the credit remains substantial, but companies that treat documentation as an afterthought are increasingly exposed to full disallowance, extended examinations, and the kind of outcomes playing right now in federal Tax Court.

Organizations that invest in contemporaneous record-keeping, disciplined cost-allocation methodology, and proactive professional guidance are the ones that will continue to realize those benefits reliably.

Secure your R&D credits today. Book a discovery call to see how TaxDrone.AI makes audit-ready documentation effortless.