The new knowledge test in the R&D Tax Incentive, explained

Understanding whether your work qualifies for the R&D Tax Incentive often comes down to one central question: was the outcome of your experiment or development project new knowledge that could not be determined in advance by a competent professional working in the field? This is the new knowledge test, and it sits at the heart of Australia's R&D Tax Incentive. In this guide, we unpack what the test means in plain English, walk through a step-by-step method you can use to assess your own activities, and explain how to build the documentation that a registered tax agent needs to prepare a claim. This article provides general information only; it is not tax, financial, or legal advice. You should always confirm your position with a qualified professional before lodging any claim.

Many Australian businesses, from software startups and manufacturing firms to agricultural producers and clean energy developers, rely on the R&D Tax Incentive to support innovation. Yet the most common point of confusion, and the area where claims most often fail, is the new knowledge test. Getting it right is not about ticking a box; it is about demonstrating, with evidence, that you tackled a genuine technical problem whose solution was not already known in your industry. When you connect your own accounting data through GrantsMAX's read-only integration, the platform helps draft an evidence-backed pack that keeps this test front and centre. But the underlying reasoning must be yours, and a registered tax agent must review and lodge the claim.

Prerequisites for applying the new knowledge test

Before you start mapping your activities against the new knowledge test, a handful of building blocks need to be in place. First, you must have a clear understanding of what constitutes a core R&D activity under the Australian regime. The Department of Industry, Science and Resources defines core R&D activities as experimental activities whose outcome cannot be known or determined in advance on the basis of current knowledge, information or experience, and that are conducted for the purpose of generating new knowledge. Supporting R&D activities, those that directly support a core activity, also matter, but the new knowledge test is applied to the core activity. So step one is to separate what is core from what is supporting. In its guidance, AusIndustry stresses that simply applying existing techniques to a known problem does not meet the threshold.

Second, you need a well-defined hypothesis or technical gap. Vague statements like "we improved our manufacturing process" will not satisfy the ATO or AusIndustry. You must be able to articulate, in engineering or scientific terms, what unknown you set out to resolve. This is where many first-time claimants struggle. A business might have spent months building a new software module, but if the team cannot describe the specific technical uncertainty they faced, say, a novel algorithm whose performance on a given dataset was uncertain, then the activity will struggle to pass the test. GrantsMAX's eligibility assessment can flag where an activity may fall short, but it is the business's job to articulate the knowledge gap.

Third, you need a system of contemporaneous records. The new knowledge test is not proven by a post-hoc justification; it is evidenced by lab notebooks, technical meeting minutes, emails, version-control logs, and design review documents that show the decision-making at the time. GrantsMAX's evidence trail feature can index these and tie them to cost lines, but you need the raw records. If you do not have them, it may be difficult to support a claim. As the ATO regularly points out, reconstructing records after the fact is not a substitute for contemporaneous evidence.

Finally, work with a registered tax agent who understands the R&D Tax Incentive. The rules are technical, and the consequences of getting the new knowledge test wrong can be severe. An accountant experienced in the space can help you decide which activities are likely to pass and how to document them. Platforms like GrantsMAX prepare a draft pack that your accountant reviews, refines, and lodges, so you get the benefit of AI-assisted drafting without removing the professional oversight that keeps a claim defensible.

What is the new knowledge test?

The new knowledge test is not a recent invention; it has been a core eligibility requirement since the current R&D Tax Incentive took effect. However, the language can be confusing because it is often discussed alongside the separate concept of a "new or improved" product, process, or service. Under Australian law, a core R&D activity must be an experiment whose outcome is unknown and cannot be determined in advance. That is, the activity must be undertaken to generate new knowledge. This new knowledge might be a discovery about a material's properties, a novel algorithm's behaviour, or a biological strain's yield characteristics. It does not require that the entire project succeeds; a failed experiment can still generate new knowledge and may be eligible. What matters is that, at the outset, the result was not deducible by a competent professional in the field.

AusIndustry and the ATO use this test to draw a line between genuine experimentation and routine problem-solving. For example, if a software company writes code using standard libraries to build a feature that any competent developer could specify, the outcome is considered deducible, even if the company had never built it before. On the other hand, if the same company sets out to determine whether a new machine-learning architecture can reduce inference latency below a threshold that no current approach can meet, and it cannot find a clear answer in the literature, then the outcome is not deducible. That is where the new knowledge test comes into play.

Internationally, similar tests exist. The IRS Research Credit in the United States uses a four-part test that includes the elimination of uncertainty, while New Zealand's RDTI explicitly requires an activity to seek to resolve scientific or technological uncertainty. The UK's SME R&D relief also hinges on whether an advance in science or technology is being sought. These parallels can help you think about the Australian test, but you must always apply the domestic rules as described by AusIndustry and the ATO. For broader context on why governments design incentives this way, the OECD's Innovation and Taxation work shows that well-targeted R&D tax measures encourage genuine investment in knowledge creation rather than subsidising business-as-usual activity.

Who is a competent professional?

The test asks whether the outcome could be determined in advance by a competent professional in the relevant field. This is not a generic person; it is someone with the skills, knowledge, and experience that would be expected of a practitioner working in that technical area. AusIndustry guidance says you should consider what a qualified engineer, scientist, software developer, or other specialist would have known at the time the R&D activity began. The competent professional is not required to be a world-leading researcher; they simply need to represent the ordinary state of knowledge in the field.

For a manufacturer exploring a new composite material, the competent professional might be a materials engineer with experience in similar composites. For a biotech startup testing a novel gene-editing approach, the competent professional would be a molecular biologist familiar with the relevant techniques. The question is: would that professional, given the existing knowledge and tools, have been able to predict the result with confidence? If the answer is yes, the activity does not pass. If the answer is no, if the professional would have had to run the experiment to find out, then the activity may meet the test (subject to other requirements).

One common mistake is to assume that because your organisation lacked the knowledge, it must be new. The test is objective; it looks at the state of knowledge in the field, not in your shop. A software company that has never built a mobile app might find the exercise challenging, but if the app uses standard frameworks and known design patterns, a competent mobile developer would have been able to predict the outcomes. That activity would likely fail the new knowledge test. Conversely, if the company is experimenting with a novel cryptographic protocol whose security properties are unproven, even a specialist could not predict the result without rigorous testing. That is the sort of uncertainty the test is designed to capture.

Step-by-step: applying the new knowledge test to your R&D activities

You can use the following five steps to assess your own core R&D activities against the new knowledge test. These steps are general guidance only and do not replace professional advice. Always consult a registered tax agent before lodging a claim.

Step 1: Identify your core R&D activities

Start by listing each experimental activity you believe is core. For a technology company, this might include developing a new data-processing algorithm, testing a hardware-software interface, or experimenting with a novel user authentication scheme. For a food manufacturer, it might involve trialling a new fermentation process or a preservation technique that differs from established methods.

For each candidate activity, write a one-sentence description that captures the technical objective. For example: "We conducted experiments to determine whether a platinum-free catalyst could achieve the same conversion efficiency as our current platinum-based catalyst." Or: "We built and tested a prototype of a real-time image segmentation model to see if it could run on a resource-constrained embedded device without exceeding latency requirements." The activity must be experimental; designing a website or performing routine testing does not count.

Pro tip: Many businesses mistakenly list the entire product development as one core activity. The ATO expects you to break the work into discrete experiments. If your project involved four separate technical unknowns, you should document four core activities. This is where GrantsMAX's draft pack can help structure the narratives so your accountant can review each one on its own merits.

Step 2: Define the hypothesis or knowledge gap

For each activity identified, articulate what you did not know at the outset, and why existing knowledge was insufficient. This is the heart of the new knowledge test. A strong hypothesis statement reads something like: "We hypothesised that by varying the voltage and pulse frequency of the electroplating process, we could achieve a coating thickness uniformity of ±2 microns on a curved surface, but the existing literature contained no data on this substrate-and-geometry combination." The key is to show that competent professionals in the field would not have been able to consult a textbook, a standard, or a public database to get the answer.

If you cannot write this statement, the activity may not pass. For exporters working on product adaptation, be careful: simply modifying a product to meet overseas regulations is often considered routine because a competent engineer can look up the relevant standards and predict the changes needed. However, if the adaptation requires you to resolve a technical uncertainty, for instance, whether a novel coating can withstand a unique climate-and-handling combination that no standard exists for, there may be a knowledge gap.

Warning: Avoid claiming that the outcome was uncertain simply because you had not done it before. The test is not about your inexperience; it is about the state of knowledge in the wider field. If the solution was available in a journal article, a patent, or a commercial product, the activity is unlikely to pass.

Step 3: Determine if the outcome was deducible by a competent professional

Now step outside your project and ask: would a competent professional, given the same information you had at the start, have been able to predict the result? This requires an honest assessment. If your team had to run an experiment because no one could confidently forecast the outcome, you have a strong case. If, however, the experiment simply confirmed something that the literature already suggested, it might be routine development.

One way to test this is to conduct a literature search at the beginning of the project and document what you found, or did not find. If a search of academic databases, industry standards, or patent filings yields no definitive answer, that lends weight to the claim that the outcome was not deducible. The ATO and AusIndustry commonly ask for such evidence during reviews. GrantsMAX's eligibility assessment prompts you to consider these factors and flags areas that a reviewer might question.

Step 4: Document contemporaneous evidence

Even the most genuine knowledge gap will not pass muster without records. The evidence must show that you were, in real time, grappling with uncertainty. Examples include:

• Technical meeting minutes that record the team discussing why the outcome was unknown
• Lab books that show failed experiments and the reasoning for changing parameters
• Emails where engineers debate whether a certain approach will work
• Design review documents that list assumptions and the tests needed to validate them

Records should be dated and, where possible, tied to specific activities. The audit-ready evidence trail built by GrantsMAX can index these sources and connect them to the claim narratives, but it cannot create records you never kept. Start building the habit of recording technical decisions now, even before you make a claim.

Pro tip: When you record a result, also record what you expected to happen. If your expectation was wrong, that is strong evidence that the outcome was not deducible.

Step 5: Prepare for review by your accountant

Once you have identified the activities, framed the hypotheses, and gathered evidence, you are ready to produce a pack for your registered tax agent. This is where the GrantsMAX accountant review workflow becomes valuable. The platform drafts R&D activity descriptions and cost structures from your Xero, MYOB, or QuickBooks data, attaches the evidence index, and creates a shared workspace. Your accountant can then review, request changes, and ultimately lodge the claim with the ATO, with the business owning the claim throughout.

Do not underestimate the importance of this step. A registered tax agent brings professional judgement to whether the new knowledge test is satisfied and how to present it in a way that aligns with ATO expectations. They also take responsibility for the lodgement under the Tax Agent Services Act, which adds a layer of quality control that an AI tool on its own cannot provide.

Documenting the knowledge gap: what the ATO and AusIndustry want

Both the ATO and AusIndustry have published resources that help you understand what constitutes acceptable evidence for the new knowledge test. AusIndustry's eligibility page is a good starting point. In practice, evidence tends to fall into two buckets: proof that the knowledge gap existed and proof that you followed a systematic process to fill it.

For the first bucket, you need a record of what the competent professional would have known. A literature review, a patent search, or a technical benchmarking report can serve this purpose. For the second bucket, you need to show that your experiments followed a recognised methodology. For a software project, that might mean an agile sprint with documented hypothesis-and-test cycles. For a manufacturing trial, it might mean a design-of-experiments matrix. The key is that the process was not haphazard; it was deliberate and aimed at resolving the specific uncertainty.

The ATO also looks for contemporaneous records because they are harder to fabricate. A notebook entry dated months before the claim was even contemplated is more credible than a retrospective memo. If you are already using cloud accounting, GrantsMAX can read that data to help allocate costs, but the technical records need to come from your team.

Common challenges and how to avoid them

Many businesses stumble on the new knowledge test because they misunderstand its scope. Here are some frequent pitfalls and practical ways to steer clear.

Challenge 1: Confusing complexity with uncertainty. A task can be technically difficult but still deducible. For example, building a complex database migration script might require deep SQL knowledge, but a competent database administrator could plan the steps and foresee the outcome. That is not R&D. To avoid this, ask: would a competent professional need to run an experiment, or could they design a solution from known principles?

Warning: If the main challenge is effort, not uncertainty, the activity probably does not involve new knowledge. R&D is about discovering what you do not know, not about doing something you understand but at scale.

Challenge 2: Failing to document the knowledge gap at the start. Many claims fail because the business cannot show that it assessed the state of knowledge before the project kicked off. You can avoid this by making a "knowledge gap statement" a mandatory part of every project initiation document. Even a paragraph emailed to the team can be powerful evidence a year later.

Challenge 3: Treating the whole project as one activity. We touched on this earlier, but it is worth repeating. A multi-year project may contain several distinct experiments, each with its own knowledge gap. If you lump them together, you risk an AusIndustry finding that some parts did not involve new knowledge, which can taint the entire claim.

Challenge 4: Overlooking supporting activities. While the new knowledge test applies to core activities, supporting activities (like building a test rig or writing code to run the experiment) can still be eligible if they directly support a core activity that passes the test. Just be careful: assembly-line upgrades that are routine in the industry are not supporting R&D just because they help a lab project.

How GrantsMAX supports the new knowledge test

GrantsMAX is not a tax agent, and it does not lodge claims. What it does is use your own accounting and operational data to draft an evidence-backed application pack that keeps the new knowledge test at its centre. Because it connects read-only to Xero, MYOB, QuickBooks, Microsoft 365, and Google Workspace, it can surface costs tied to specific activities and help you structure the narrative around the technical uncertainties you faced. This makes the accountant's review faster and more precise.

The opportunity discovery feature continuously scans relevant programs, including the R&D Tax Incentive and state innovation grants, so you can see what you may be eligible for without manually trawling directories. The eligibility assessment then flags any areas where your data suggests a gap, such as insufficient evidence for a particular activity. From there, the AI application pack drafting generates a complete pack, which flows into the accountant review workflow. Throughout the process, the business remains in control, and the registered tax agent is responsible for final lodgement.

For founders and CFOs who are used to paying legacy consultancies tens of thousands of dollars to prepare a claim, this approach can significantly reduce the friction of applying while still requiring that a human professional signs off. For growing companies approaching the refundable-offset turnover threshold, which is proposed to rise from $20 million to $50 million, staying on top of the new knowledge test across a portfolio of activities becomes even more important. And for accountants and bookkeepers, the platform offers a way to extend services to clients without building an in-house R&D tax team.

Conclusion: key takeaways

1. The new knowledge test is whether a core R&D activity's outcome was not deducible in advance by a competent professional. It is not about your organisation's lack of knowledge but about the state of knowledge in the field.
2. To apply the test, identify discrete core activities, articulate the specific hypothesis or knowledge gap, assess deducibility from the perspective of a competent professional, and gather contemporaneous evidence.
3. Documentation is critical. The ATO and AusIndustry expect to see records that show you systematically tackled an uncertainty, not that you simply built something complex.
4. Common pitfalls include confusing effort with uncertainty and failing to record the knowledge gap at the project's start. Avoiding these mistakes can strengthen a claim significantly.
5. While tools like GrantsMAX can draft an evidence-backed pack, a registered tax agent must review and lodge the claim. This division of responsibility is essential for a defensible application.

If you are an Australian business with R&D activities and want a clearer picture of what you may be eligible for, consider joining the GrantsMAX waitlist. The platform reads your accounting data, helps you structure the evidence, and prepares a pack that your accountant can review, so you can focus on the innovation, not the paperwork.