An R&D tax credit claim rarely fails loudly. It fails in a letter — a negative or partly negative technical opinion, a request for comment, a line in an audit note — that quietly puts a six-figure cash position back in play. By the time that letter arrives, most of the outcome has already been decided. Not by the facts. By the documentation.
In our first article we argued that, under Austria's evolving 2026 framework, the hard part is no longer claiming the 14% R&D tax credit (Austria's Forschungsprämie) — it is being able to prove entitlement when someone looks closely. This piece is about what "looking closely" actually feels like, and why the difference between a claim that holds and one that collapses is set long before the challenge.
The anatomy of a contested claim
Picture a mid-sized manufacturer or life-sciences company with several years of legitimate, substantial R&D. The premium has been claimed and paid. Then, for one or more years, the technical opinion is not clean — and the amount at stake runs well into six figures.
What follows is a sequence that is draining precisely because it is procedural:
- A request for comment arrives, often with a tight deadline.
- The company must now reconstruct, sometimes years later, what was uncertain in each project and how that uncertainty was resolved.
- Memories have faded, people have moved on, and the evidence that would have made the case obvious was never captured in the first place.
- The defence is built under pressure, in hindsight — the worst possible conditions.
None of this means the underlying work wasn't genuine R&D. It very often is. The claim is at risk not because the science was weak, but because the record of it was.
Why hindsight loses
Reconstructed documentation has a tell. It describes outcomes — what was built, what shipped — rather than the thing the credit actually rewards: the process of resolving technical uncertainty. An auditor reading a results-oriented narrative sees a successful engineering project. An auditor reading a well-kept contemporaneous record sees research. Same work; two very different claims.
| Reconstructed under pressure | Built as the work happens | |
|---|---|---|
| Narrative | "We developed product X" | "We faced uncertainty Y; here is how we investigated it" |
| Evidence | Assembled from memory | Captured at the time, dated, traceable |
| Credibility | Defensive, after the fact | Self-evident |
| Typical outcome | Protracted dispute | Claim stands |
The real cost is not the fee
Companies tend to measure the cost of a contested claim as the advisory time needed to defend it. The larger costs are quieter: the cash that sits frozen while the dispute runs, the management attention pulled into reconstruction, and the chilling effect — teams that, once burned, quietly under-claim in future years to avoid the hassle. A credit designed to encourage risk-taking ends up discouraging it, purely because the paperwork hurt last time.
Built to be defended
The alternative is not more paperwork. It is paperwork built for a different purpose — not to file, but to withstand a challenge that may never come. Project boundaries drawn cleanly. An uncertainty narrative written while the uncertainty is still live. An evidence chain that a reviewer can follow without a guided tour. Done this way, a request for comment stops being a crisis and becomes a formality: the answer already exists.
This is, deliberately, a discipline distinct from the tax filing itself. Your tax advisor owns the return. The technical defensibility of the R&D behind it is a complementary craft — and the companies that sleep well in audit season are the ones who treat it as such, from the first project of the year rather than the last week before filing.
Where to start
You do not need a negative opinion to find out where you stand. The most useful exercise is also the simplest: take one or two of your current R&D projects and ask whether, today, you could hand a sceptical reviewer a clear, dated account of what was uncertain and how you addressed it — without reconstructing anything.
If the answer is "not quite," that gap is the real risk in your claim. And it is far cheaper to close now than to explain later.
Finest Selection Group focuses on the technical defensibility of R&D claims — complementary to your tax advisor. This article is general information, not legal or tax advice; specifics of the 2026 framework remain provisional until the FoPR 2026 is officially published.