At Neo.Tax, we talk to scores of controllers and Heads of Tax every month. One thing is clear: the old way of filing for R&D Tax Credits has become a burdensome pain. You’d think that so many businesses moving their project management and payroll online, via software like Jira, Rippling, and Workday would simplify the filing process. But, in some ways, it’s made it harder.

The payroll and project management output is a rich dataset, but it’s also, well, voluminous. If that data is not inputted perfectly—and let’s be honest, it never is—that means it’s up to a controller or Head of Tax to sift through the mess to create the filing. Engineers need to be trained (and retrained) on the correct project tagging process, and then cajoled to sort out any inputting errors. It’s a waste of time for your filer and your engineering teams.

Which is why Neo.Tax created our AI system. 

No one starts a company because they’re excited about accounting and taxes. Well, no one except the Neo.Tax founders :)

‍

But for those other 99% of founders, here’s a cheat sheet and guide to ASC 350-40, an accounting guideline that pertains to Internal-Use Software Development Costs. Maybe you’ve never heard of ASC 350-40, but you should get to know the rule if you ever plan to go public.

In July, the government moved for a summary judgment ruling that Kyocera AVX, the multinational ceramics and electronics manufacturer, is not entitled to the $1.3 million amended Section 41 R&D Credit they recently claimed. The company had hired ​​PricewaterhouseCoopers (PwC) to make the filing, and the accounting firm did interviews with subject matter experts to compute the eligible amount of Section 41 credits they could claim for the tax years 2017-2020. PwC found $1.3 million worth of unclaimed credits, which they filed in the amended Section 41 study. But, ultimately, the IRS claims that the R&D Study and documentation provided were not sufficient to support the claim, and when the IRS requested additional documentation to support the claim neither Kyocera nor PwC could produce it.

Two of the government’s objections in its case against Kyocera are especially relevant to any company that files for an R&D tax credit: 1) “The PwC study exclusively relied on interviews to determine employees' time spent on projects; it did not use documentation” and 2) “Kyocera does not have a centralized system for tracking employee time, and generally does not track employee time on projects.”

The IRS has been emphasizing the necessity of contemporaneous documentation for decades, as far back as Eustace v. Commissioner in 2002. But this recent motion for summary judgment against Kyocera could mark a potential shift towards a stricter application of their long-standing policy.

The direction seems clear—reliance on after-the-fact interviews without supporting documents is increasingly untenable.

Why AI-Enhanced R&D Tax Credit Calculation is More Relevant Than Ever: Lessons from the Kyocera Case

Kyocera's reliance on a PwC-conducted study, which was ultimately deemed insufficient by the IRS due to lack of proper documentation and over-reliance on estimates, highlights a systemic risk in the current practices of R&D credit calculation.

The IRS mandates strict substantiation requirements for claiming R&D tax credits under Section 41 of the IRC. Specifically, Treasury Regulation § 1.41-4(d) emphasizes the necessity for contemporaneous documentation to support the expenditures claimed. This documentation must detail the nature of the qualifying activities and their direct connection to the claimed credits—a challenging task for businesses relying on traditional methods, particularly for those that do not have employee time tracking systems in use.

In the case of Kyocera, the IRS disputed the R&D tax credit claims due to inadequate substantiation of the research activities claimed under Section 41. The agency focused on the lack of contemporaneous documentation, relying instead on retrospective estimates from interviews conducted by PwC long after the fact. This method was seen as inadequate in the eyes of the court; they demanded a more rigorous substantiation to prove that the activities qualified as R&D.

Neo.Tax’s AI-Enhanced Solution

This is where Neo.Tax separates itself in the R&D tax space. Unlike traditional methods that often rely on after-the-fact recollections and manual calculations, Neo.Tax leverages advanced AI to analyze real-time data from integrated systems such as Jira and GitHub. This not only ensures greater accuracy but also provides a contemporaneous, auditable trail of documentation that meets IRS requirements. Better yet, our AI not only reviews these systems to identify the research activities, but also uses the meta data to help allocate the time. And because the meta-data is recorded "at the same time" as the tickets in Jira, it is — by definition — "contemporaneous documentation."

Real-Time Data is the Gold Standard for Compliance

By accessing data from systems such as Jira and GitHub at the time of tax preparation, Neo.Tax’s platform can automatically extract, classify, and substantiate qualifying R&D activities. This method aligns with the IRS's emphasis on contemporaneous records. The agency prefers these types of records because they are more reliable than documentation created far after when the activity occurs — certainly, an engineer creating a timesheet of his work at the end of the month is likely more accurate than asking them to recollect their work on a project completed a year, or multiple years, in the past.

Unlike traditional R&D studies conducted by large accounting firms, which rely on retrospective interviews and questionnaires, Neo.Tax’s solution utilizes real-time data to support R&D credit claims. This direct link to project-specific activities minimizes reliance on employee recollection and significantly enhances the accuracy and defensibility of claims during IRS audits.

We often speak of the way our AI solution saves time for controllers, heads of tax, and engineers by eliminating the need for time-consuming retrospective interviews. But the Kyocera case highlights another, far more stark and costly risk of relying on memory and estimations for look-back claims. The IRS is making it clear that they’re emphasizing the need for contemporaneous data. Companies who rely on the R&D credit should take note.

This Should Be A Wake-Up Call for Innovators

As businesses increasingly adopt project-management tools that capture their developmental activities in real-time, a new approach to R&D tax credit documentation is possible. Neo.Tax’s AI-powered platform provides a reliable, efficient, and IRS-compliant method to claim R&D tax credits. By leveraging data directly from the systems so-often used by innovative companies, Neo.Tax ensures that your spending on R&D is rigorously documented in accordance with current tax laws. 

We can get you the money you’re owed. Just as importantly, we can deliver the data-rich study that the IRS prefers and expects.