What Is a Core R&D Activity?
Whether your work qualifies for the R&D Tax Incentive depends entirely on whether your activities meet the legislative definition of a core R&D activity. This page explains the three requirements, what the statutory exclusions are, and how AusIndustry and the ATO assess activities when they review your claim.
The Legislative Definition
Under section 355-25 of the Income Tax Assessment Act 1997, a core R&D activity is an experimental activity that satisfies two simultaneous requirements:
- Technical uncertainty: The outcome cannot be known or determined in advance on the basis of current knowledge, information or experience anywhere in the world, and can only be determined by applying a systematic progression of work based on principles of established science
- Purpose: The activity is conducted for the purpose of generating new knowledge, including new or improved materials, products, devices, processes or services
Both requirements must be satisfied at the same time. An activity that involves genuine experimentation but is primarily conducted to solve an immediate commercial problem fails the purpose requirement. An activity aimed at generating knowledge but relying on routine, non-experimental methods fails the technical uncertainty requirement.
The definition is deliberately narrow. The R&D Tax Incentive is not a general subsidy for innovation or product development. It applies to activities where the outcome is genuinely unknowable in advance and where knowledge is generated through systematic scientific method. For a broader overview of what this means for your business, see our R&D Tax Incentive eligibility guide.
The Three Requirements in Detail
Technical Uncertainty
The outcome of the activity must be genuinely unknown and cannot be resolved by consulting existing publicly available knowledge anywhere in the world. The uncertainty must be technical or scientific in nature. Market uncertainty (whether customers will want the product) and commercial uncertainty (whether the project will be profitable) do not count as technical uncertainty under the legislation.
The key question is the competent professional test: could a skilled expert in the relevant field determine the outcome of this activity without conducting experiments, using current publicly available knowledge from anywhere in the world? If a competent professional could determine the outcome, there is no technical uncertainty. If they genuinely could not predict the outcome, the technical uncertainty requirement may be met.
The competent professional is assessed by reference to global public knowledge, not just what your team knew at the time. A problem that has already been solved elsewhere, even if your team was unaware of the published solution, is not technically uncertain under this test.
Systematic Progression of Work
The activity must follow a systematic progression of work based on principles of established science. This means proceeding from hypothesis to experiment, observation, evaluation, and logical conclusion. Undirected trial-and-error without a structured experimental methodology does not meet this requirement.
The progression must be documented contemporaneously. In practice, this means that when each experiment is designed, the hypothesis being tested should be recorded. When results are observed, those observations should be captured at the time. When conclusions are drawn, the reasoning should be written down during the R&D period, not reconstructed months later from memory or commercial project records.
Routine debugging, standard quality assurance testing, and user acceptance testing are generally not systematic experimentation under this definition. These activities verify known specifications rather than test hypotheses under genuine uncertainty.
New Knowledge Outcome
The activity must be conducted for the purpose of generating new knowledge. New knowledge under the legislation includes knowledge about new or improved materials, products, devices, processes, or services. The purpose test focuses on intent: the dominant reason the activity was conducted must be knowledge generation, not simply delivering a commercial project or meeting a deadline.
A development team can incidentally generate new knowledge while completing commercial work without that knowledge generation being the activity's purpose. That distinction is important. AusIndustry and the ATO assess whether knowledge generation was the actual dominant reason the specific experimental steps were taken, not whether useful knowledge happened to be produced as a byproduct.
The knowledge generated does not need to be academically novel. It must not have been publicly available before your activities. However, knowledge that a competent professional could have obtained by consulting existing documentation, prior art, or standard professional practice does not meet the new knowledge requirement.
What Qualifies and What Does Not
The following examples span multiple industries and illustrate where the line falls. For expenditure rules once your activities are confirmed as eligible, see our R&D Tax Incentive guide.
✅ LIKELY QUALIFIES:
- Systematically testing multiple novel machine learning architectures to determine which can achieve required accuracy on a sparse, domain-specific dataset where published models underperform and no established approach resolves the problem
- Developing and testing a new composite material formulation where the structural properties under repeated load cycles are genuinely unknown and cannot be determined from published data or prior experiments
- Iterating through experimental catalyst configurations for an industrial chemical process where the reaction yield under specific conditions cannot be predicted from current scientific literature
- Designing and testing firmware for a new sensor interface where timing constraints create edge cases no existing protocol addresses and the viable approach is genuinely uncertain
- Experimentally determining whether a novel bioactive compound achieves target receptor binding, where computational models give inconclusive predictions and laboratory testing is the only means of resolution
- Testing multiple structural design approaches for a load-bearing component where material behaviour under a specific combination of temperature, stress, and vibration has not been documented or studied
❌ DOES NOT QUALIFY:
- Building a standard web application using established frameworks following documented patterns, even if the application is novel to your business
- Integrating a third-party API or commercial software library where the integration approach is described in vendor documentation and involves no genuine uncertainty
- Routine user acceptance testing or quality assurance testing to verify a feature meets a known specification, where there is no hypothesis and no uncertainty about the outcome
- Manufacturing a product using an existing, commercially proven process, even if the specific product is new to your company
- Market testing a new service offering to determine customer demand or pricing sensitivity
- Adapting an existing algorithm for a new use case where the adaptation method is documented and routinely applied across the industry
Statutory Excluded Activities
Section 355-25(2) of the ITAA 1997 lists activities that can never be core R&D activities, regardless of how technically complex or innovative they appear. These exclusions apply by statute, even if an activity otherwise seems to involve experimentation or knowledge generation.
Market Research and Testing
Market research, market testing, market development, or sales promotion, including consumer surveys. The exclusion covers research into what customers want, how they respond to products, or how to develop a market. The sophistication of the methodology does not matter: if the purpose is market understanding rather than technical knowledge generation, it is excluded.
Internal Administration Software
Developing, modifying, or customising software used for the dominant purpose of administering your own business. This covers developing or customising accounting systems, CRM platforms, ERP software, HR systems, and similar internal tools, regardless of the technical complexity involved in building them.
Routine Testing and Quality Control
Routine data collection, testing, or analysis of materials, components, products, processes, soils, atmospheres, and other things. Testing to verify that a product or process meets a known specification is excluded. Eligible testing must involve genuine uncertainty about whether the approach will work at all, not verification against a predetermined standard.
Social Sciences, Arts, and Humanities
Activities in social sciences, arts, or humanities are excluded entirely. The RDTI applies only to experimental activities based on principles of natural or applied sciences. Research into human behaviour, cultural phenomena, or artistic practice does not qualify regardless of the experimental rigour applied.
Regulatory Compliance Work
Activities associated with complying with statutory requirements or standards, where the work is essentially procedural rather than genuinely experimental. Routine regulatory submissions, compliance audits, and mandatory reporting do not qualify. Note that therapeutic goods trials may be treated differently depending on their experimental character.
Commercial and Legal Aspects of IP
Commercial, legal, and administrative aspects of patenting, licensing, or other activities related to intellectual property protection or transfer. The legal and commercial work surrounding a patent does not become R&D by association with the underlying invention. Only the experimental work that generated the invention itself is potentially eligible.
How AusIndustry and the ATO Assess Core Activities
Registration with AusIndustry is a self-assessment process, not an approval. Submitting a registration and receiving a registration number does not mean AusIndustry has confirmed your activities qualify. Both AusIndustry and the ATO can review your claim at any point up to four years after lodgement, and the ATO has independent authority to deny claims without needing an AusIndustry Finding.
AusIndustry: Activity-Level Assessment
AusIndustry reviews each registered activity separately, not the project as a whole. During a compliance review, they assess:
- Whether a genuine technical hypothesis existed at the start of the activity, not just a commercial goal
- Whether the state of publicly available knowledge at the time meant a competent professional could not resolve the uncertainty without experimentation
- Whether the experimental methodology followed a systematic progression from hypothesis to evaluation and conclusion
- Whether contemporaneous records document the hypothesis, experiments, observations, and conclusions made during the R&D period
- Whether knowledge generation was the primary purpose of the activity, not incidental to commercial delivery
ATO: Independent Eligibility Determination
The ATO has authority to determine the eligibility of R&D activities independently, without needing an AusIndustry review outcome. At the tax return stage, the ATO commonly challenges:
- Claims where the technical uncertainty could have been resolved using existing knowledge or standard professional practice
- Activities described in commercial or outcome-focused terms rather than technical terms framed around hypotheses and uncertainty
- Records created retrospectively rather than contemporaneously during the R&D period
- Entire projects claimed as single core R&D activities when only specific experimental phases would meet the criteria
- High-value claims in software, AI, and technology sectors where the line between routine development and genuine experimentation is contested
Documentation: What Both Bodies Expect
Contemporaneous records are the strongest evidence in any review. Records created after the fact carry significantly less weight. For each core R&D activity, records should capture:
- The specific technical problem and the knowledge gap that existed when the work began, including background research showing what was already known
- The hypothesis being tested and why a competent professional could not resolve it without conducting experiments
- The experimental method used, the scientific principles applied, and the iterations or variations conducted
- Observations and results captured at the time, including failed experiments and unexpected outcomes
- Conclusions drawn and the new knowledge generated as a result of the experimental work
Acceptable forms include lab notebooks, Git commit notes with technical explanations, design iteration records, test logs, prototype documentation, and progress reports written during the R&D period. Engineering change orders, technical design documents, and expert consultant notes can also support a claim.
Frequently Asked Questions
Does a core R&D activity need to succeed?
No. A core R&D activity does not need to produce a successful outcome. The legislation requires that the activity be conducted for the purpose of generating new knowledge, with genuine technical uncertainty at the outset and a systematic experimental approach. An experiment that proves a hypothesis wrong still qualifies, provided it met those criteria when conducted. Failed experiments are, in many ways, the strongest evidence of genuine technical uncertainty, since failure confirms the outcome was not known in advance.
What is the competent professional test?
The competent professional test asks whether a skilled expert in the relevant field could determine the outcome of the activity without conducting experiments, using knowledge publicly available anywhere in the world at the time the work was done. If yes, there is no technical uncertainty and the activity fails the core R&D definition. The test applies to global public knowledge, not just what your team knew. If a solution already existed in published literature or established industry practice, it cannot be technically uncertain regardless of whether your team was aware of it. See our eligibility guide for how this test applies across different industries.
Can only part of a project be registered as core R&D?
Yes, and this is the norm rather than the exception. Most commercial projects combine genuinely experimental phases with routine development, integration, or testing work. Only the phases that involve genuine technical uncertainty, systematic experimentation, and a primary purpose of generating new knowledge can be registered as core R&D activities. Registering an entire project as a single core activity when only specific phases qualify is one of the most common compliance risks and is frequently challenged by both AusIndustry and the ATO. Separating eligible experimental work from routine development within a project is a key part of preparing a defensible claim.
How does the ATO assess whether my activities genuinely qualify?
The ATO assesses claims against contemporaneous records, AusIndustry registration descriptions, and the technical plausibility of the claimed uncertainty. They may request internal documentation, engage technical specialists to assess whether the stated uncertainty was genuine, and compare described activities against publicly available knowledge in the relevant field. Claims that describe R&D activities in commercial or outcome-focused terms rather than technical, hypothesis-driven terms are significantly more likely to attract scrutiny and result in reductions. Our R&D consultants work with companies to frame activities in technically accurate terms from the outset of each registration.
More questions? See our R&D Tax Incentive FAQ →
Not Sure Whether Your Activities Qualify?
The line between a genuine core R&D activity and routine development is not always clear. We assess your specific activities against the legislative tests and give you an honest answer about what will hold up under AusIndustry and ATO review.
Free Eligibility Assessment