When the R&D is Real but Your Records are Not: Defending a Genuine Claim with Weak Documentation

You know your business conducted genuine research and development. You remember the problems you were trying to solve. You saw the experiments, the failures, the iterations, the breakthroughs. The innovation was real.

But when the ATO or AusIndustry come knocking, your documentation doesn’t tell that story.

This is the uncomfortable position many businesses find themselves in: technically eligible for the R&D tax incentive, but unable to prove it cleanly. And here’s the thing that keeps advisors awake at night: being right on the technical merits isn’t enough if you can’t demonstrate what actually happened.

Let’s talk about what that means in practice, what you can realistically do about it, and when weak documentation becomes a dispute you need to take seriously.

The uncomfortable reality: genuine R&D, imperfect records

This situation is more common than anyone wants to admit.

You’re a software business iterating through sprints, documenting everything in Jira but never formally writing up hypotheses or experiment plans. You’re a manufacturer running trial after trial on a new process line, with engineers scribbling results on whiteboards that got wiped at week’s end. You’re a founder-driven business where the technical founder holds most of the knowledge in their head, and “documentation” means a few emails and some rough notes.

The innovation happened. The experiments were real. But the file doesn’t show it.

This matters because the R&D tax incentive doesn’t operate on good faith. It operates on evidence.

The ATO administers the tax offset and wants to see that your expenditure is legitimate. AusIndustry (through the Industry Innovation and Science Australia decision-making body) determines whether your activities are actually eligible R&D. Both agencies care deeply about contemporaneous documentation because it’s the most reliable way to test whether what you claimed is what you actually did.

When your documentation is weak, you face a double problem. First, you can’t easily prove the activities were genuinely experimental rather than routine business operations. Second, you can’t demonstrate how costs were genuinely incurred on R&D rather than general business activities.

And here’s what makes this particularly uncomfortable: the regulators don’t need to prove you’re lying or fraudulent. They just need to conclude that you haven’t discharged your burden of proof. That’s a much lower bar.

If you’re reading this and recognising your own situation, you’re not alone. But you do need to understand what you’re facing.

How the ATO and AusIndustry test your claim

Understanding who does what makes a material difference to how you respond.

AusIndustry (IISA) determines eligibility of activities. They assess whether what you did meets the legal definition of core R&D activities (experimental activities for the purpose of generating new knowledge) or supporting R&D activities (directly related to core R&D). This is a technical, science-based assessment. They issue findings on whether registered activities are eligible.

The ATO administers the tax side. They audit your expenditure, test whether costs were genuinely incurred on the activities you claimed, and determine whether you’re entitled to the tax offset. They also impose penalties if claims are found to be incorrect.

In practice, these two processes often run together. The ATO may refer questions about eligibility to AusIndustry. AusIndustry’s adverse finding on eligibility will typically flow through to the ATO’s assessment of your tax position.

When a review begins, both agencies are looking for the same core things: evidence of a structured approach to experimentation, records showing what hypotheses you tested, documentation of results, and a clear decision trail showing how experimental outcomes influenced your next steps.

They want to see the scientific method in action, even if your business doesn’t think in those terms.

What they don’t want is generic descriptions, retrospective justifications, or broad assertions that “we were innovating”. They’ve seen too many claims where the activity was just normal product development or troubleshooting dressed up as R&D after the fact.

This is where weak documentation becomes a real problem. If your registration describes activities in vague terms, if your project records don’t show clear hypotheses and testing, if your cost allocations rely on rough estimates rather than timesheets or project codes, you will struggle to satisfy either regulator.

The review process itself typically starts with a letter. You’ll be asked to provide documentation supporting your claim: project plans, technical reports, meeting notes, lab notebooks, test results, correspondence showing technical problem-solving. You’ll usually have 28 days to respond, though extensions are often granted.

If your initial response doesn’t satisfy the agency, you may face a formal review or audit. AusIndustry can conduct a formal compliance review and issue an adverse eligibility finding. The ATO can amend your assessment and disallow the offset.

At that point, you’re into dispute territory. You can object, seek internal review, and ultimately appeal to the Administrative Review Tribunal or Federal Court. But every step requires evidence. Weak documentation at the start doesn’t magically improve as the dispute escalates.

When weak documentation becomes a dispute issue

Not every documentation gap triggers a dispute. Some do.

The most common flashpoints are predictable because they reflect the areas where the agencies focus their scrutiny.

Vague or generic activity descriptions in your registration are a red flag. If your registration says “developing new software solutions” or “improving manufacturing processes” without explaining what specific technical uncertainties you were addressing or what experiments you conducted, you’re inviting questions. AusIndustry needs to understand what made the activity experimental rather than just iterative development or troubleshooting.

Missing or superficial project records raise immediate concerns. If you can’t produce experiment plans, test protocols, or documented results, the regulator will question whether systematic experimentation actually occurred. They’ve seen too many businesses where R&D claims are based on normal project work that happened to involve some problem-solving.

Inconsistent cost allocation is another common trigger. If your claimed expenditure doesn’t align with project timesheets, if you’re using rough percentage estimates, or if costs include activities that look like general business operations, the ATO will push back. They want to see that the dollars you claimed genuinely relate to eligible R&D activities.

Business-as-usual activities dressed as R&D get picked up quickly. If what you’re describing sounds like standard product development, routine testing, or adapting existing technology to a new use without genuine experimentation, you’re in trouble. This is where the line between innovation (commercial) and R&D (experimental) matters most.

Projects that lack a clear technical narrative face scrutiny. Regulators want to see a story: what you didn’t know at the start, what hypotheses you formed, how you tested them, what you learned, and how that informed your next steps. If your documentation jumps from “we had a problem” to “we built a solution” without showing the experimental work in between, you haven’t proven R&D occurred.

When these issues arise, a routine review can escalate into a full dispute surprisingly fast.

You might start with an AusIndustry compliance review that results in an adverse finding: activities not eligible. That finding flows through to the ATO, who then amend your assessment to disallow the R&D offset. You’re now facing an amended assessment, a tax bill, potentially interest, and the question of penalties.

At this point, you need to make a strategic call: do you have enough evidence, reconstructed or otherwise, to mount a credible defence? Or are you better off accepting the adverse outcome, paying what’s owed, and resetting your approach for future claims?

That decision turns almost entirely on the quality of evidence you can now assemble. And this is where many businesses discover, too late, that weak contemporaneous documentation is very hard to overcome.

Rebuilding the evidence when your file is light

If your documentation is weak but you genuinely conducted eligible R&D, you need to rebuild the evidentiary picture. This is possible. It’s not easy, and it’s never as strong as proper contemporaneous records, but it can be done.

The goal is to construct a coherent, credible narrative that demonstrates what you did, why it was experimental, and how costs relate to those activities. You’re working backwards from the outcome to prove the process.

Start with whatever project materials you do have. Even incomplete records can provide anchors. Look for emails discussing technical challenges, design iterations, or problem-solving. Dig through code repositories for commit histories and comments that show experimentation. Pull Jira tickets, Trello boards, or other project management tools that capture what was being worked on and when. Find meeting notes, Slack threads, or internal memos that reference testing or technical decisions.

These materials won’t be perfect. They weren’t created with R&D compliance in mind. But they can show that genuine experimental work occurred, and they provide a timeline.

Next, reconstruct project timelines and experimental activities. Map out what you were trying to achieve, what uncertainties you faced, how you approached testing, and what you learned. This often requires sitting down with the technical team and walking through the project step by step. What hypotheses did you form, even if you never wrote them down formally? What tests did you run? What failed? What worked? How did results influence the next iteration?

Document this reconstruction carefully. It’s not contemporaneous evidence, but it is evidence. If it’s consistent with the limited contemporaneous materials you have, it adds weight.

Staff interviews and witness statements can be powerful, but they must be done properly. A vague statement that “we did R&D” is worthless. A detailed witness statement from a lead developer or engineer explaining specific technical problems, the approaches tested, the results observed, and the decisions made carries real weight, particularly if it aligns with emails, code commits, or other records.

Technical reports and expert evidence become critical when contemporaneous documentation is thin. An independent expert who reviews your activities, examines whatever records exist, and provides an opinion on whether the work constitutes R&D under the legislative tests can significantly strengthen your position. This is not the same as getting an expert to simply agree with you. The expert must critically assess the work and explain why it meets the criteria.

Physical evidence matters too. Prototypes, test samples, failed iterations, equipment logs, lab results. Anything tangible that proves experimentation occurred. If you’ve still got the physical artefacts of the R&D process, they can corroborate your reconstructed narrative.

Cost reconstruction is harder but necessary. If you didn’t track time to projects properly, you need to work backwards. Who worked on the R&D activities? What percentage of their time was genuinely experimental work versus routine tasks? Can you support those estimates with emails, calendar entries, or project milestones? The ATO will challenge rough estimates, so your allocations need to be as defensible as possible given the gaps.

All of this requires rigour and honesty. You cannot manufacture evidence. You cannot exaggerate. Regulators and tribunals are experienced at spotting reconstructed evidence that’s been tailored to fit the claim. What you’re aiming for is a credible, evidence-based account that withstands scrutiny.

But here’s the reality: reconstructed evidence is always weaker than contemporaneous records. A tribunal or reviewing officer will give more weight to a lab notebook completed during the experiment than to a witness statement written two years later recalling what happened. You can close the gap, but you can’t eliminate it.

This means your decision on whether to fight comes down to how strong your reconstructed case is. If the gaps are too wide, if the story doesn’t hold together, if the costs can’t be defended, you may be better off conceding and moving on.

How decision-makers view genuine but under-documented R&D

You can be technically eligible and still lose.

That’s the uncomfortable truth. Tribunals and regulators apply a principle that’s simple but unforgiving: the burden of proof is on you. You must demonstrate, with evidence, that your activities meet the legislative tests. Assertions, even honest ones, are not enough.

The law requires core R&D activities to be experimental activities conducted for the purpose of generating new knowledge. You need to show what knowledge you were seeking, why existing knowledge was insufficient, and how you systematically tested hypotheses to fill that gap. Supporting activities must be directly related to core R&D.

If your evidence is too vague, too general, or too thin, decision-makers cannot conclude that these tests are met, even if they suspect the activities were probably eligible.

A recent tribunal case illustrates this. The business had registered activities and genuinely believed they were conducting R&D. But the evidence provided was insufficient to demonstrate that the activities met the statutory tests. The registration described activities in broad terms. The supporting materials lacked detail on hypotheses, experimentation, and outcomes. The tribunal found the activities were not eligible, not because the business was dishonest, but because the evidence didn’t prove what needed to be proven.

This is the risk you face with weak documentation. Even where a tribunal accepts that innovation occurred, that doesn’t automatically mean R&D occurred in the legislative sense. The innovation might have been incremental adaptation, skilled problem-solving, or iterative development, none of which qualify.

Decision-makers look for specificity. General statements like “we experimented with different approaches” don’t help. Specific statements like “we tested three different algorithms to reduce processing time, documented the performance benchmarks for each, and selected the approach that met our latency requirements after eight iterations” do help.

They look for contemporaneity. Evidence created at the time carries more weight than evidence created later. If your only project records are a report written months after the work finished, prepared specifically for the R&D claim, it will be viewed skeptically.

They look for consistency. Do your witness statements align with the emails? Do the timesheets match the project descriptions? Do the claimed costs correspond to the documented activities? Inconsistencies undermine your entire case.

And they apply the law strictly. It’s not enough to show that you were innovative, entrepreneurial, or solving hard problems. You must show that you were conducting systematic experimental activities for the purpose of generating new knowledge. That’s a technical legal test, and weak evidence makes it very hard to satisfy.

This doesn’t mean reconstructed evidence never succeeds. It can. But you need to be realistic about the standard you’re being held to and whether your evidence meets it.

If you’re facing an adverse finding or considering an objection or appeal, this is the lens you need to apply: will a tribunal or reviewing officer, looking at the evidence we can now assemble, be satisfied on the balance of probabilities that eligible R&D activities occurred? If the answer is marginal or unclear, the risks of proceeding may outweigh the benefits.

Risk, consequences, and when to fight

The stakes in an R&D dispute are not trivial.

If your claim is disallowed, the ATO will amend your assessment to remove the R&D tax offset. You’ll owe the tax you didn’t pay (or lose the refund you received). That’s the starting point. But it rarely ends there.

Interest accrues on unpaid tax from the original due date. Depending on how long the matter has taken to resolve, interest can be substantial.

Penalties are a real risk. The ATO can impose penalties for making a false or misleading statement, even if you didn’t intend to mislead. The penalty percentage depends on your behaviour: did you take reasonable care? The base penalty for failure to take reasonable care is 25% of the tax shortfall. That percentage can be reduced through voluntary disclosure or increased if the shortfall is significant. Penalty remission is possible, but it requires demonstrating reasonable care and cooperation. Weak documentation makes that harder to argue.

There’s also the question of future audit risk. If the ATO or AusIndustry has disallowed claims for one year, they will likely scrutinise your other years. If you’ve claimed R&D offsets for multiple years with similarly weak documentation, you could be facing a multi-year unwind. The financial exposure multiplies quickly.

Reputational consequences also matter. If you’re in a sector where R&D claims are common, being known as a business that had claims knocked back can affect relationships with advisors, auditors, and even investors. It signals a lack of rigour.

So when do you fight, and when do you concede?

This is a strategic decision, not an emotional one. It turns on the strength of your reconstructed evidence, the size of the financial exposure, and the likely cost of dispute proceedings.

If your reconstructed evidence is strong, if you have credible witness statements supported by contemporaneous materials, if an independent expert can corroborate your position, and if the disputed amount is material, fighting may be the right call. Objection to the ATO, followed by tribunal proceedings if necessary, gives you a formal process to present your case.

If your evidence is weak, if the gaps are significant, if the story doesn’t hold together, or if the cost and time of disputing outweigh the potential recovery, conceding is often the smarter commercial decision. You pay what’s owed, you avoid penalty risk escalating, and you reset for future claims with better processes.

There’s a middle path too: partial concession. You might concede that some activities or costs were not defensible but maintain that others were. This can reduce your exposure and demonstrate cooperation, which helps with penalty remission.

The key is to make this decision based on a cold, hard look at the evidence, not wishful thinking. Many businesses pursue disputes because they’re convinced they were right, without honestly assessing whether they can prove it. That path leads to wasted legal costs and worse outcomes.

If you’re going to fight, you need to be confident that your evidence, as reconstructed, will satisfy a tribunal or court on the balance of probabilities. If you can’t reach that level of confidence, think very carefully before proceeding.

What advisors and businesses can do differently on future claims

If you’ve been through a dispute, or if you’re advising a client with weak documentation, the lesson is clear: fix the process for next time.

This doesn’t require over-engineering or creating compliance bureaucracy that stifles innovation. It requires embedding simple, sustainable documentation habits into the R&D workflow.

Start with activity descriptions that are specific and technical. When registering R&D activities with AusIndustry, describe what technical uncertainty you’re addressing, what approach you’re testing, and what new knowledge you’re seeking. Avoid generic language. Make it clear why the activity is experimental, not just innovative.

Implement lightweight project documentation. This doesn’t mean formal lab notebooks for every business. But it does mean capturing, at the time, what hypothesis you’re testing, what you tried, what the result was, and what you learned. An email summarising a sprint review that discusses what worked and what didn’t is contemporaneous evidence. A Confluence page documenting a technical decision and the experiments that informed it is evidence. A Git commit message explaining why a particular approach was tested and then abandoned is evidence.

The goal is not perfection. It’s capturing enough, in real time, to tell the story later.

Track time and costs properly. If you’re claiming R&D expenditure, you need to be able to show that the costs genuinely relate to R&D activities. Use project codes, timesheets, or task tracking to separate R&D work from business-as-usual work. If your team works on both R&D and non-R&D projects, allocate their time accurately. Rough estimates will not survive scrutiny.

Get technical input before lodging. Before you register activities or lodge an R&D claim, have someone with technical expertise review the activity descriptions and supporting documentation. Can they clearly identify the experimental work? Can they explain the technical uncertainties and how you addressed them? If not, your documentation isn’t ready.

Document assumptions and judgements. If you’ve made judgement calls about whether an activity qualifies, or how to allocate costs, document those decisions and the reasoning behind them. If you’re later challenged, you can demonstrate that you took reasonable care. This also protects advisors. If you’re relying on client-provided information, document what you were told, what you reviewed, and what assumptions you made.

Review high-risk areas proactively. Some activities are inherently harder to defend: agile software development with minimal formal planning, iterative engineering work, activities that blend R&D with product development. If you’re claiming in these areas, documentation needs to be tighter, not looser. Anticipate the questions a regulator will ask and make sure you can answer them.

Consider periodic compliance reviews. If R&D offsets are material to your business, an annual or biannual internal review of your R&D claims can catch documentation gaps before the ATO or AusIndustry do. This might be done by your advisor, or by an external R&D specialist. The cost is modest compared to the cost of a dispute.

None of this is radical. It’s basic good practice. But it’s surprising how many businesses lodge R&D claims year after year without ever stress-testing whether they could actually defend those claims if challenged.

Talking to clients about documentation risk without killing innovation

One of the hardest conversations for advisors is telling a client their R&D claim might not survive scrutiny because the documentation is weak.

Clients hear “you’re doing it wrong” or “you’re at risk” and they panic. Or they dig in. Or they decide R&D claiming is too hard and stop altogether.

None of those outcomes serves anyone.

The better framing is this: documentation is not red tape. It’s making your story provable. You’ve done the hard work. You’ve solved genuine technical problems. You’ve experimented, failed, learned, iterated. That’s all real. But if you can’t show that process to someone who wasn’t in the room, you can’t access the incentive designed to reward it.

This isn’t about creating a compliance burden that slows down your team. It’s about capturing just enough, in the moment, so that later you can prove what happened.

Most clients respond well to this framing because it validates the innovation while acknowledging the practical reality of needing evidence.

You can also point to the commercial risk. The R&D offset is often material. Losing it in a dispute, plus interest and penalties, can be a significant financial hit. Simple documentation processes are cheap insurance against that risk.

For clients who resist, it’s worth being direct: if you’re not willing to document your R&D in a way that makes the claim defensible, you probably shouldn’t be claiming. The downside risk is too high.

But for most clients, once they understand that documentation doesn’t mean bureaucracy, they’re willing to adjust. A few simple changes to workflow, a bit more discipline around recording decisions and results, and their position improves dramatically.

The goal is not to scare clients away from claiming R&D. It’s to make sure that when they do claim, they can defend it. That’s in everyone’s interest.

The reality: clarity before you claim

Litigation isn’t about hoping for the best. It’s about knowing where you stand and making decisions based on evidence, not optimism.

If your R&D is genuine but your documentation is weak, you’re in a difficult position. You can rebuild some of the evidentiary picture. You can assemble witness statements, expert reports, and whatever contemporaneous materials exist. But you need to be honest with yourself about whether that reconstructed evidence will satisfy a tribunal or reviewing officer applying strict legal tests.

The better path is not to end up there in the first place.

For future claims, build documentation into your R&D process. Capture enough, in real time, to prove the story later. Track costs properly. Get technical input before lodging. Stress-test your position before the regulator does.

And if you’re already facing a dispute, make the strategic call based on the evidence you can realistically assemble, not the evidence you wish you had.

The right advisor won’t just prepare your R&D claim. They’ll help you understand whether it’s defensible. And if it’s not, they’ll tell you what needs to change.

Disclaimer: This article is general information only. It is not legal advice. R&D tax incentive disputes involve complex factual and legal issues. Outcomes depend on the specific circumstances of each case. If you are facing an ATO or AusIndustry review, or considering an R&D claim with incomplete documentation, seek legal advice specific to your situation before taking action.