The R&D Tax Incentive for fintech companies

Introduction

Financial technology companies in Australia often operate at the frontier of software engineering, data science, and regulatory architecture. Whether you are building algorithms for real-time fraud detection, designing a tokenised settlement layer, or creating new encryption methods for open banking, your team regularly tackles problems that require genuine experimentation. When that work meets the definition of research and development under Australian law, your fintech may be eligible for the R&D Tax Incentive, a self‑assessment program jointly administered by AusIndustry (for registration) and the ATO (for the tax offset).

This article walks through the process step by step, with a particular focus on how fintech businesses can distinguish eligible technical experimentation from the work of compliance, integration, and routine software development. The guidance is general information only, not tax or legal advice. Every claim involves specific facts, and the rules can change between income years. Always confirm your position with a registered tax agent before lodging.

Before we begin: GrantsMAX does not lodge claims or provide tax advice. GrantsMAX prepares an evidence‑backed application pack from your own accounting data and hands it to your registered accountant or tax agent, who reviews, refines, and lodges. The business owns the claim.

Prerequisites

Before you work through the steps, you need a clear picture of your business’s R&D‑active projects and some structural basics in place.

• A registered tax agent. Under the Tax Agent Services Act, only a registered agent can charge for preparing and lodging an R&D tax incentive claim. If you do not already have one, find an agent experienced in R&D claims for software‑intensive companies.
• A working understanding of your core and supporting R&D activities. You will need to describe the hypothesis, the technical uncertainty, the systematic progression of work, and the outcome for each activity. A loose goal is not enough.
• Access to your accounting records. The ATO requires a direct link between claimed expenditure and the registered activities. Cloud accounting data from Xero, MYOB, or QuickBooks is the foundation. (GrantsMAX’s browser connector reads that data read‑only, so you can surface the raw numbers quickly.)
• A process for contemporaneous record keeping. The ATO expects you to be able to demonstrate the R&D narrative and expense allocation without relying on hindsight. If you have not yet set up a system, now is the time.

Pro Tip: Start documenting before you register. The moment you frame a hypothesis and begin a systematic experiment, record the technical challenge, the approach, the test results, and the time and resources committed. A simple Word document or cloud note is better than nothing, but tools that link activity logs to financial data simplify substantiation at claim time.

Step 1: Understand R&D Tax Incentive Eligibility for Fintechs

Eligibility is defined by reference to “R&D activities” as set out in the Income Tax Assessment Act 1997 and administered by AusIndustry. The program uses two categories: core R&D activities and supporting R&D activities. The Department of Industry, Science and Resources publishes detailed guidance on business.gov.au that is worth reading in full.

Core R&D activities in fintech

A core R&D activity is an experimental activity whose outcome cannot be known or determined in advance on the basis of current knowledge, information or experience, and that is conducted for the purpose of generating new knowledge (including about the creation of new or improved materials, products, devices, processes, or services).

In a fintech context, core activities might include:

• Designing a novel consensus mechanism for a distributed ledger that must satisfy contradictory requirements for speed, energy consumption, and finality.
• Developing a machine‑learning model to detect synthetic identity fraud, where the model architecture and training approach involve fundamental uncertainty.
• Creating new cryptographic protocols to enable secure, low‑latency data sharing under the Consumer Data Right, when existing standards cannot deliver the required performance without compromising privacy.
• Building a real‑time liquidity engine that must solve an optimisation problem for which no known algorithm guarantees a solution within the required latency budget.

Notice that the common thread is a technical unknown that cannot be resolved simply by applying standard engineering practices or consulting an API manual. The work must involve a genuine scientific or technological gap.

Supporting R&D activities

Supporting R&D activities are activities directly related to a core R&D activity, but that do not themselves constitute experimentation. Examples relevant to fintechs include:

• Setting up a sandbox environment to test a new model.
• Running structured trial transactions to generate data for training or validation.
• Preparing technical documentation that captures the systematic progression of the core activity.
• Cleaning and labelling data when that work is directly required for the experiment and is non‑trivial.

Activities that are not R&D

AusIndustry and the ATO are clear that certain activities, however complex, fall outside the definition. The plain‑English guide on the GrantsMAX blog covers many of these, but for fintech specifically, the exclusions that trip companies up include:

• Compliance work: mapping regulatory rules to software requirements, testing against a known standard, or building a standard AML/KYC workflow are not R&D, because the outcome is determined by an external specification.
• Integration with existing systems: connecting to a core banking platform, a card scheme, or an identity provider via a documented API is implementation, not experimentation.
• Routine debugging, performance tuning within known parameters, and user‑acceptance fixes.
• Market research, customer surveys, and commercial scoping.

Warning: Many fintechs classify their entire platform build as R&D. The fact that a product is new to the market does not make its development R&D. You must be able to point to the specific technical uncertainty that you attempted to resolve through systematic work. If the pathway was known from the start, it is unlikely to be eligible, no matter how hard the coding.

Step 2: Identify and Distinguish Eligible R&D from Compliance and Integration

For a fintech, this step is make‑or‑break. The ATO and AusIndustry have publicly stated that mischaracterising ordinary software development is a major focus area. When you are building a product that must operate inside regulated environments, the lines can feel blurred, but the test is always a technical one.

Ask yourself five questions for any project:

1. What was the specific technical uncertainty? (e.g. “We did not know whether a gradient‑boosted tree model could detect transaction fraud with fewer than 5 false positives per 100,000 transactions on a mobile payments dataset.”)
2. Why could you not resolve that uncertainty using publicly available information or professional knowledge? (e.g. “Published benchmarks assumed full‑fat features; our memory budget was 300 MB.”)
3. What hypothesis did you form to address the uncertainty?
4. How did you systematically test that hypothesis? Describe the experimental design, the data, the iterations.
5. What was the outcome? What did you learn, even if the experiment failed?

If you can give crisp answers to these five questions, you likely have a core R&D activity. If you find yourself answering with “We built a module that was new to us” or “The project was complex” or “We had to comply with AUSTRAC rules,” then you are still in implementation territory.

The compliance trap

Compliance with Australian financial services regulation is a business necessity, not a research activity. Building a responsible lending assessment engine that implements prescribed formulas is not R&D, even if the work is difficult. However, if you are trying to create a new, probabilistic methodology that goes beyond the prescribed standard because existing methods cannot meet a novel consumer need, that may qualify, provided you document the technical uncertainty and the experiment.

Pro Tip: Separate the compliance and R&D streams in your project plans and timesheets. When you can show that a specific engineer spent 60% of their time on the experimental kernel and 40% on regulatory adjustments, the claim will be much easier to substantiate.

Integration with existing systems

Writing code that calls external endpoints, maps JSON structures, or handles retries is generally not R&D. The knowledge is known: the API documentation tells you what to do. Even if the integration is messy and time‑consuming, it does not involve an experimental outcome.

The bar changes if you need to invent a novel integration pattern because no standard exists. For example, designing a way to atomically exchange value and data across two independent blockchains without a trusted intermediary might be R&D, because the technical problem is unresolved in the field. But piping Open Banking data into your own database via an accredited gateway is not.

Many global tax authorities draw the same distinction. The US IRS, for instance, requires that a qualified research expense involve the elimination of uncertainty concerning the development or improvement of a product, while the UK’s HMRC publishes detailed guidance on software projects and R&D relief. The common thread is that the work must tackle something technically uncertain, not something merely unfamiliar to your team.

Step 3: Set Up Evidence‑Backed Record Keeping

Good records do more than satisfy the ATO; they make the difference between a claim that sails through review and one that requires months of follow‑up. The ATO expects you to keep records that explain the R&D activities, link the expenditure to those activities, and demonstrate that you conducted the activities as described.

Start with three layers of evidence:

1. Project descriptions and experiment logs. For each core activity, keep a living document that records the hypothesis, the plan, the tests, the results, and any pivot decisions. It helps to date each entry and link it to a version of the code or dataset.
2. Expenditure records. Claimable expenses typically include salary and on‑costs for staff directly engaged in R&D, consumables, and a proportion of overheads. The ATO requires a reasonable apportionment method. If you use cloud accounting software, the raw data is already there; GrantsMAX’s browser connector can read that data read‑only and map it to the activities you have described.
3. Corroborating documents. Screenshots of test outputs, internal presentations, emails that record technical decisions, and version‑control commits that show the experimental branch evolving all help prove that the work was systematic and genuinely exploratory.

Consider using a tool or platform that ties the narrative to the financial data. For instance, GrantsMAX’s discovery and matching engine identifies potential R&D activities by reading your accounting ledger; from there you can draft narratives and link expenses directly, building a pack your accountant can review.

Pro Tip: The ATO can review R&D claims up to four years after lodgment, so keep records at least until the end of that review period. Digital storage is fine, provided the documents are legible and you can retrieve them on request.

Step 4: Register Your Activities with AusIndustry

Registration is a prerequisite for claiming the offset. You must register your R&D activities with AusIndustry within 10 months after the end of the income year in which you conducted them. The form is submitted through the business.gov.au portal, and you need to describe each core and supporting activity in clear, plain language.

Broad, generic descriptions such as “software development” will be rejected or will invite scrutiny. Instead, write a short paragraph for each core activity that covers:

• The precise technical objective.
• The technical uncertainty that existed at the outset.
• The methodology (e.g. iterative prototyping, A/B testing of model architectures, formal verification).
• The outcome, even if the hypothesis was disproved.

The same registration description will later underpin your tax claim, so invest the effort upfront. The GrantsMAX for technology companies page explains how specific, data‑backed narratives help reduce back‑and‑forth with reviewers.

Step 5: Calculate the R&D Tax Offset and Prepare Your Claim

Once the activities are registered, the next step is calculating the offset and assembling the claim ready for your registered tax agent.

Refundable vs. non‑refundable offset

The structure of the offset depends on your aggregated turnover. For entities with aggregated turnover of less than $20 million, the offset is refundable, meaning you can receive cash back if the offset exceeds your tax liability. For entities with aggregated turnover of $20 million or more, the offset is non‑refundable and carried forward to reduce future tax. The ATO publishes the current rates on its website (ato.gov.au), and you should confirm the applicable rate for your income year because both the rate and the turnover threshold can change. Note: the government has announced a proposal to raise the refundable threshold to $50 million with effect from income years starting on or after 1 July 2025, but at the time of writing this announcement is not yet enacted. Speak to your tax agent about which law applies to your claim.

GrantsMAX prepares a pack for your accountant

Manually building a claim pack that ties timesheet data to AusIndustry activity descriptions and statutory schedules is labour‑intensive. GrantsMAX addresses that by reading your accounting data read‑only, identifying activities that may be eligible, and drafting the narrative, cost allocation, and supporting schedules. The outcome is a complete, defensible pack that your registered accountant reviews, refines, and lodges.

For first‑time claimants who have never navigated the program before, the feeling of staring at a blank AusIndustry form can be daunting. The GrantsMAX for first‑time claimants page walks through how the platform reduces that friction. For R&D‑active startups that spend a significant portion of their early burn on engineering, GrantsMAX for R&D‑active startups shows how the tool surfaces eligible activities from raw Xero transactions.

Step 6: Review and Lodge with a Registered Tax Agent

Your registered tax agent will:

• Review the AusIndustry registration to ensure the descriptions align with the law.
• Verify that the expenditure included in the claim is eligible and correctly apportioned.
• Check that the offset calculation matches your tax position.
• Lodge the company tax return, which includes the R&D tax incentive schedule.

The agent’s professional opinion matters. If the ATO reviews the claim, the agent represents you. Do not attempt to self‑lodge unless you are an exempt entity; the Tax Practitioners Board strictly limits who can charge a fee for preparing an R&D claim.

Warning: The ATO has sophisticated data‑matching capabilities and is paying close attention to R&D claims from software companies. A claim that is not grounded in contemporaneous evidence and a solid technical narrative is likely to be adjusted. Never lodge a claim that you are not confident your agent would defend under audit.

Conclusion and Key Takeaways

Fintech companies in Australia are building genuinely new technology, and the R&D Tax Incentive exists to support that work. But accessing the incentive requires more than a hunch that your product is innovative. It demands a methodical process: identifying the true technical uncertainties, systematically testing hypotheses, keeping records that connect the experimentation to the costs, and registering with AusIndustry on time.

The dividing line between eligible R&D and quotidian software work is especially sharp for fintechs because so much of the day‑to‑day involves regulation, compliance, and integration. Treat those activities as necessary operational work, not as R&D. Reserve the R&D label for the projects where your team could not know the outcome in advance and had to experiment to find a solution.

• Eligibility turns on technical uncertainty and systematic experimentation, not on commercial novelty.
• Compliance, integration, and routine development are not R&D, even when they are difficult.
• Contemporaneous records are your strongest asset when the ATO asks questions.
• Registration with AusIndustry is mandatory and time‑sensitive.
• The offset rate and refundable threshold can change; always confirm the figures with the ATO and your tax agent for the relevant income year.
• A registered tax agent must review and lodge the claim; the business owns the outcome.

GrantsMAX helps Australian fintechs take the grunt work out of the process by preparing an evidence‑backed application pack from your own accounting data, for your registered accountant to review and lodge. To see how it works, book a 30‑minute walkthrough or join the waitlist. We would be glad to show you how your fintech’s real experimentation can surface the opportunities it may deserve.