Most Australian software projects do not qualify for the R&D Tax Incentive, even when they are technically demanding or expensive. Eligibility under section 355-25 of the ITAA 1997 turns on three tests: a technical uncertainty that cannot be readily resolved, a systematic process of experimentation, and the generation of new knowledge. Standard development, configuration, integration, UI/UX work, performance tuning, and replicating known solutions sit outside the program. Only the experimental portion of a software project is claimable.
The R&D Tax Incentive is one of the most valuable programs available to Australian businesses building software. It is also one of the most misunderstood, and a large share of software claims are reduced or rejected on review.
A common assumption is that any sufficiently complex or expensive software project will qualify. In practice, claims fail not because the work was easy but because it does not meet the specific criteria the program was designed to reward. The legislation tests whether the activity resolved a genuine technical uncertainty through experimentation, not whether the project was hard, late, or over budget.
According to the ATO's R&D Tax Incentive Transparency Report 2022-23, professional, scientific and technical services together with information media and telecommunications accounted for over 6,200 R&DTI claimants, or roughly 48% of all 12,956 claims that year. Those two industries are where most software work is captured. AusIndustry has published a dedicated software development sector guide precisely because so many software claims are flagged as ineligible on review.
Source: ATO, R&D Tax Incentive Transparency Report 2022-23, Table 2 (published September 2025).
Eligible software R&D under section 355-25 of the ITAA 1997 must resolve a technical uncertainty that cannot be answered using existing knowledge, follow a systematic progression of work (hypothesis, experiment, observation, evaluation), and generate new knowledge. Difficulty, expense, and project duration are not part of the test. Routine engineering and configuration are excluded.
AusIndustry's software development sector guide sets out how to self-assess software activities against the legislative tests. The guide makes clear that the eligibility line is drawn at the activity level, not the project level. A single software project can contain a small set of eligible experiments alongside a much larger volume of ineligible implementation work.
Claims are most often reduced because companies treat the entire software project as R&D, conflate commercial novelty with technical uncertainty, and lack contemporaneous records of hypotheses, experiments, observations, and evaluations. AusIndustry and the ATO jointly assess registrations and tax claims, and both agencies look for evidence at the activity level, not just the project level.
Companies that build their R&DTI program alongside their broader business advisory and forecasting tend to capture eligible activity correctly the first time. Those that file at year-end without contemporaneous records often see their claim narrowed during review. The pattern is consistent: eligibility is decided by the records, not by the technology stack.
Across the activities our team reviews, the following categories account for the largest share of software work that is incorrectly included in R&DTI claims. Where they form part of a wider project that also contains genuine experimentation, only the experimental portion can be claimed.
Most software projects are implementation-driven rather than experimentation-driven. Examples include building web applications on established frameworks (React, Node, Django, Rails), developing mobile apps with conventional architectures, and creating dashboards, CRMs, or internal tools. The work can be demanding, but it relies on known methods and predictable solutions, which is why it does not satisfy the s.355-25 test. For wider context on how R&DTI fits with broader software accounting, see our SaaS and IT accounting guide.
A significant portion of what is labelled "software development" is in fact configuring SaaS platforms, customising existing systems, integrating APIs or third-party services, and migrating or transforming data. The execution may be difficult, but the outcome is broadly predictable, which sits outside the scope of R&D under business.gov.au guidance.
Design and usability improvements are often mistaken for R&D. Refining user journeys, redesigning interfaces, running A/B tests, and improving usability are driven by design judgement and user feedback rather than technical uncertainty. They are commercially valuable but they do not meet the eligibility threshold.
Post-launch work is almost always ineligible. This includes bug fixing, refactoring, improving load times, and scaling infrastructure. These activities apply known techniques to known problems and are treated as routine software maintenance rather than experimentation. Eligibility is the exception, not the rule, and only when a specific performance outcome is genuinely unpredictable in advance.
If a product is built on well-established concepts (marketplaces, SaaS platforms with familiar feature sets, common system architectures), it generally will not qualify. The product may be new to your business, but the program asks whether it is new to the broader field of knowledge, not just new to you. This is the most frequent reason early-stage startup R&DTI claims are reduced.
If a competent developer could solve the problem by reading documentation or applying existing libraries, frameworks, or design patterns, the work is unlikely to qualify. The knowledge already exists, so there is no genuine technical uncertainty to resolve. AusIndustry's competent professional test is consistent on this point.
Pricing engines, workflow automation, and rules-based decision systems can be intricate, but they are deterministic implementations of business rules. Complexity is not the same as technical uncertainty, and these systems sit firmly on the implementation side of the line. For SaaS businesses combining R&DTI scoping with revenue and contract design, our SaaS revenue recognition guide covers the related compliance.
Writing user documentation, managing project scope, drafting commercial requirements, running stand-ups, and resolving stakeholder requests are not core or supporting R&D activities. They can sometimes form part of supporting activities under s.355-30 if their dominant purpose is to support a specific eligible core experiment, but that link must be evidenced. Time spent on general business operations is not claimable.
Eligible software R&D typically involves new algorithms or computational methods, performance under constraints that cannot be modelled in advance, novel system architectures, or experiments where multiple approaches need to be tested before one works. The work generates new technical knowledge rather than new features. This activity is far less common than standard development, which is precisely why the program exists to support it.
For software businesses, R&DTI scoping should sit alongside tech and SaaS accounting and broader tax planning. Treating R&DTI as a year-end exercise consistently produces narrower, lower-quality claims than treating it as an ongoing engineering and finance discipline.
Our team can review your software project, identify which activities meet the s.355-25 tests, and quantify the experimental portion before you commit to an AusIndustry registration.
Contact Us Today →To qualify under the R&D Tax Incentive, software development must satisfy all three of the following conditions, set out in section 355-25 of the ITAA 1997 and explained in the AusIndustry how to conduct eligible activities guidance:
| Test | What it means in software |
|---|---|
| 1. Technical uncertainty | A competent professional could not have determined the outcome in advance using existing knowledge, documentation, or design patterns. |
| 2. Systematic experimentation | A documented hypothesis, planned experiments, observed results, and evaluated conclusions, captured contemporaneously rather than reconstructed at year-end. |
| 3. New knowledge | The work produces information not previously available in the public domain, not just a feature that is new to your business. |
Both the ATO and AusIndustry enforce this framework jointly. The underlying distinction is straightforward: eligible R&D is driven by uncertainty and experimentation, ineligible work is driven by execution of known solutions.
If you are trying to assess a software project quickly, the most useful question to ask is:
Were we solving a problem that did not already have a known solution, and did we genuinely need to experiment to get there?
If the honest answer is no, the work almost certainly is not R&D, no matter how challenging it felt. If the answer is yes for some specific activities within a wider project, those activities (and only those activities) form the basis of an eligible claim.
For broader context on how software activity sits within the wider claim landscape, see our companion post on the eight most common areas for R&DTI claims in Australia.
Most software development is not R&D, and the program is not designed to subsidise it. The work that does qualify is the genuinely uncertain experimental subset of a project, evidenced by contemporaneous records and tested against the s.355-25 framework. Companies that scope R&DTI carefully against the eight categories above, claim only the experimental work, and document it as they go, consistently achieve more durable claims than those that file the entire engineering ledger and hope for the best.
42 Advisory provides end-to-end R&D Tax Incentive advisory, from eligibility assessment and AusIndustry registration through to ATO schedule lodgement and audit support. We also support tech businesses with business tax services and CPA accounting to keep R&D scoping aligned with broader compliance.
Speak with a CPA who has scoped, registered, and defended R&DTI claims for SaaS, fintech, and software businesses across Australia.
Schedule a meeting →Disclaimer: The information in this article is general in nature, current as at May 2026, and does not constitute specific tax, legal, or financial advice. R&D Tax Incentive eligibility is fact-specific and subject to ongoing AusIndustry and ATO interpretation. We recommend seeking professional advice tailored to your individual circumstances. 42 Advisory is a CPA firm and Registered Tax Agent (TPB No. 26303651).
No. Standard application or product development that uses established frameworks and known design patterns does not satisfy the s.355-25 ITAA 1997 test. Only the specific activities that resolve a genuine technical uncertainty through experimentation are eligible.
Technical uncertainty exists where a competent software professional could not determine the outcome of the work in advance using existing knowledge, documentation, libraries, frameworks, or design patterns. Commercial uncertainty, design preference, and project complexity are not technical uncertainty.
Generally no. Post-launch maintenance work, including bug fixes, refactoring, load-time improvements, and infrastructure scaling, applies known techniques to known problems. It is treated as routine software maintenance rather than experimentation, and is excluded from R&DTI eligibility.
No. Configuring SaaS platforms, customising existing systems, integrating APIs, and migrating data is implementation work. The execution may be technically demanding, but the outcome is broadly predictable using existing tools, which puts it outside the R&DTI definition of experimental activity.
AusIndustry expects contemporaneous records that show the hypothesis being tested, the experiments conducted, the results observed, and the conclusions evaluated, all linked to specific eligible activities. Reconstructed narratives written at year-end without supporting records are unlikely to satisfy review.
Companies with aggregated turnover under $20 million receive a refundable offset at the corporate tax rate plus 18.5% (typically 43.5%). Companies at or above $20 million receive a non-refundable offset based on R&D intensity, set at 8.5% above the corporate rate where intensity is 2% or less, and 16.5% above where intensity exceeds 2%.