If you are claiming the R&D Tax Incentive, you have three broad options: do it yourself, engage a consultant, or use software with expert review. They cost very differently, and the headline rate is not the whole story. As someone who reviews claims for a living, here is an honest comparison, including where each option genuinely wins.
A disclaimer worth stating up front
The R&DTI is a self-assessment program and no provider can guarantee an outcome. This is general information, not financial or tax advice. The right choice depends on your specific circumstances, and you should seek independent advice for them.
What are your options for running a claim?
Three: do it yourself, engage a traditional consultant, or use software with built-in expert review. They differ on cost, effort, and the quality of evidence they produce.
Do it yourself. You self-assess, register with AusIndustry, and claim through your tax return. Cost is your time and the risk of getting it wrong. Workable for very simple claims, painful and risky as soon as the activities or financials get involved. See the software eligibility guide to gauge whether your claim is simple enough.
Traditional consultant. A firm prepares the claim for you, usually charging a percentage of your R&D benefit, commonly in the 15 to 20 percent range, and often a retainer up front. You get expertise and offload the work. You pay a meaningful slice of the benefit, the engagement is typically once a year, and evidence gathering is usually manual.
Software with expert review. A platform drafts the claim from the work you already do and an R&D expert reviews it. Pricing is lower and structured differently. Evidence is captured continuously rather than reconstructed at year end.
How do the costs compare?
This is the comparison Rand publishes on its own pricing page, generalised to the market:
| Typical consultant | Software with expert review (Rand) | |
|---|---|---|
| Pricing model | 15-20% of R&D benefit | 5-10% marginal on eligible expenditure |
| Upfront cost | $2k-$5k retainer common | None |
| When you pay | Through the engagement | Only after the benefit is issued |
| Evidence prep | Mostly manual | Captured continuously from your tools |
| Turnaround | Weeks to months, annual | Continuous, always current |
| Expert review | Often outsourced | Built in |
On a claim of any size, the difference between roughly 15 to 20 percent and a marginal 5 to 10 percent is material. On a benefit of $200,000, that gap is tens of thousands of dollars. For an early-stage company, the timing matters too, paying nothing up front and nothing until the benefit lands is very different from a retainer plus a percentage.
Watch the fee base
"Percentage of benefit" and "marginal percentage of expenditure" are different bases. Always model the actual dollar fee on your numbers, not just the headline rate. A marginal, bracketed structure (lower rates on higher spend, like tax brackets) usually works out far cheaper than a flat percentage of the whole benefit.
When is a consultant the right choice?
When the claim needs bespoke judgement that software cannot standardise. A traditional consultant is the right call when:
- Your claim is unusually complex, multiple entities, overseas activities, or contested eligibility that needs bespoke legal judgement.
- You want a single named adviser to own the engagement end to end and represent you directly.
- You have no interest in any involvement and will pay a premium to hand it off entirely.
The honest trade-off: you pay materially more, the work happens once a year, and the evidence is only as good as what can be reconstructed at claim time.
When does software win?
For most software-company claims: it costs less, captures stronger evidence, and keeps the claim current all year.
- Cost, for the large majority of software-company claims.
- Evidence quality. Capturing the experimental trail continuously, from your pull requests, issues, and time records, produces stronger, contemporaneous evidence than a year-end reconstruction. That matters more every year, and especially under the 2028 reforms, which narrow the program toward a well-evidenced core. See documenting a defensible claim.
- It stays current. The claim is always up to date, not a once-a-year sprint.
Rand's model is deliberately a blend: AI drafts the claim, you review it, and an Expert Reviewer (in our case, me) approves it before submission. You get expert judgement without the percentage-of-benefit fee, and an evidence trail built as you work rather than rebuilt at the end.
The bottom line
For a complex, contested, or multi-entity claim, a specialist consultant earns the premium. For a typical Australian software company, software with built-in expert review is usually cheaper, produces better evidence, and keeps the claim current year-round. Model the actual fee on your numbers, look hard at the quality of the evidence each option produces, and weigh the 2028 changes when you decide.
The R&D Tax Incentive is a self-assessment scheme. Rand helps you prepare and document a defensible claim but cannot submit on your behalf and does not guarantee any outcome. This is general information, not tax, legal, or financial advice.