In-house vs adviser vs software for R&D documentation

01What are the three ways to document R&D?

Companies typically keep their R&D Tax Incentive (RDTI) records in one of three ways: in-house, using the spreadsheets, wikis and ticketing tools they already run; through an adviser who assembles the evidence at claim time; or with software built to capture R&D evidence as the work happens. Many use a mix of the three.

• In-house. The team records R&D in its own tools — a spreadsheet of projects and hours, a wiki page per experiment, tickets in the issue tracker — and hands it over at year-end.
• Adviser assembly at claim time. An R&D adviser interviews the technical team after the income year, gathers what exists, and writes the activity descriptions and expenditure schedules.
• Contemporaneous documentation software. A dedicated tool prompts the team to record hypotheses, experiments, results and time as the work happens, in a structure an adviser can review. Dossio is one example of this category.

None of these changes the rules. The RDTI is a self-assessment program, and the ATO is clear that "the responsibility for compliance rests with the taxpayer" (ATO — Helping you get R&D claims right). The approaches differ in how well the records they produce match what the regulators ask for.

02When is the evidence created in each approach?

Timing is the biggest difference. Business.gov.au says records should be created at the time the activity is conducted, and the ATO says documents created after the fact will generally not be adequate on their own. In-house and software approaches can capture evidence as work happens; adviser assembly at claim time usually starts after it has finished.

The guidance is consistent across both regulators:

• Records "should be created at the time the activity is conducted, to ensure the most accurate and comprehensive information is included" (business.gov.au — Record keeping for the R&D Tax Incentive).
• Business.gov.au recommends setting up a record-keeping system before starting R&D activities, so records are kept "as activities and expenditure occur".
• The ATO says it is a requirement that companies maintain contemporaneous records, and that documents created after the fact "will generally not be adequate on their own without some contemporaneous records" (ATO — Helping you get R&D claims right).
• The ATO also says records that are backdated or contain non-specific details of the work are not appropriate records (ATO — Keeping records and calculating your notional deductions).

Adviser assembly does not have to be retrospective — an adviser can be engaged at the start of the year and check in as work progresses. But when the first conversation happens after year-end, the adviser can only work with what the team kept at the time. For more on the timing question, see what contemporaneous documentation means.

03How do the three approaches compare?

In short: in-house is cheapest to start but depends on discipline; adviser assembly at claim time brings expertise but often starts late; software builds structure into capture but adds a tool the team must adopt. The table compares them on the factors reviewers and finance teams tend to care about.

04Who structures the evidence, and who makes the eligibility call?

Structure and judgement are separate jobs. Structure is about recording the right facts in a consistent form; judgement is deciding whether an activity meets the legislative test. Any of the three approaches can supply structure. The judgement belongs to a qualified, registered adviser in all of them — software does not replace that.

Business.gov.au lists what records should show, whichever tool holds them (Record keeping for the R&D Tax Incentive):

• background research showing the outcome could not be known in advance
• what was done, how it was done, when it happened and who did it
• the resources used, and the related expenditure — what it was on, how much, who incurred it, when, and which R&D activity it relates to
• documents supporting any apportionment methods and adjustments.

In-house records often hold much of this, but scattered across tools and not linked to each other. An adviser assembling at claim time imposes structure afterwards. Documentation software applies a structure at the point of entry. Each is a way of getting the same facts into the same shape.

Eligibility is different. The Tax Practitioners Board (TPB) lists "assisting clients with tax concessions for expenditure incurred on research and development activities where the service involves the application of taxation laws" as a tax agent service, and a person must not charge a fee for a tax agent service unless registered (TPB — Tax agent services). A spreadsheet, a wiki or a software tool does not make that judgement. A registered adviser does.

05How do the cost models differ?

In-house records cost staff time. Adviser engagements may be priced as a fixed fee, by time, or as a share of the benefit. Documentation software is usually priced as a subscription or fixed fee. The pricing model matters because it shapes incentives, and the ATO has publicly named commissions linked to the refundable offset as a concern.

Advisers are used widely. The Government's 2026 review of R&D support, Ambitious Australia, cites DISR analysis that 86% of businesses surveyed rely on R&D tax consultants, and reports that most firms the panel spoke with pass more than 10% of their claims to consultants as success-based fees — see our analysis of where adviser fees go for the sources and method.

Two points of public record are worth knowing when comparing models:

• Among its "concerning practices", the ATO lists "charging excessive commissions that are a large percent of the refundable R&D tax offset" (ATO — Helping you get R&D claims right).
• The same ATO page notes that registered tax agents who advise companies to make incorrect claims may be referred to the TPB, and that promoter penalty laws under Division 290 of Schedule 1 to the Taxation Administration Act 1953 may apply to promoters of schemes to access the RDTI for ineligible activities.

Neither point means a fee linked to the benefit is improper in itself. They are reasons to understand what you are paying for, and on what basis. If you want to put your own numbers side by side, the Quote Reality Check sets an adviser fee next to the internal hours spent gathering evidence.

06How much effort does each approach take?

Effort is not removed by any approach, only moved. In-house shifts it onto the team throughout the year. Adviser assembly concentrates it into interviews and document hunts after year-end. Software spreads it into short, prompted entries during the work. The total depends on how many projects run and how well records are kept.

We have not found a primary source that measures hours per approach, so we do not quote one. What the ATO does say is that time is one of the items that has to be shown: companies "must substantiate the time spent on eligible R&D projects" by the researchers and technicians who carry them out (ATO — Keeping records and calculating your notional deductions). The ATO also prefers that record keeping allows tracking of expenditure to R&D activities on a "real time" basis, with apportionment used only in limited situations.

Time is usually the hardest thing to rebuild after the fact. A year-end interview can recover what an experiment was for; it rarely recovers who spent how many hours on it in March.

07What do reviewers ask for?

Whoever kept the records, the questions under review are the same. The ATO says business records must verify the amount spent on R&D, the nature of the activities, and the relationship between the two. When registering, companies also declare they have maintained appropriate records while conducting their activities.

The ATO's three tests are in Helping you get R&D claims right: records must be sufficient to verify the amount of expenditure on R&D activities, the nature of the activities, and the relationship of the expenditure to the activities. If the ATO reviews a claim, it "may request a full description of your R&D activities" (ATO — Keeping records).

The Department of Industry, Science and Resources can separately review the registration. When applying, a company must declare that it has "maintained appropriate records while conducting your activities" (business.gov.au — Record keeping). Records must be kept for 5 years after the claim.

08What are the trade-offs of documentation software?

Software is not the right answer for every company. It adds a subscription and another tool, needs the technical team to use it consistently, and cannot make up for thin entries. It does not decide eligibility. For a company with one small, well-documented project, disciplined in-house records reviewed by an adviser may be enough.

Fair trade-offs to weigh:

• Adoption. A tool only helps if engineers and scientists use it during the work. If entries stop in April, the gap is the same as with any other method.
• Quality of entries. Structured fields can still be filled with vague text. The ATO's point about non-specific records applies to software as much as to a spreadsheet.
• Cost and overlap. A subscription is another line item, and some teams already have good records in their ticketing or lab systems.
• Fit and export. Check that records can be exported in an open format and kept for the 5-year retention period, whether or not you stay with the tool.
• No eligibility call. Software can structure evidence and flag missing items. A registered adviser still has to make the judgement and sign off.

09How do you choose an R&D documentation approach?

Start with the facts, not the tool. Count how many R&D activities you run, how much of the evidence already exists in dated form, and when your adviser first sees it. The more projects and people involved, the more a structured, contemporaneous method tends to matter. Keep a qualified adviser for the eligibility judgement whatever you choose.

Questions to ask:

1. When is our evidence created? If the honest answer is "at year-end", fix the timing first.
2. Can we show time against each activity? If not, that is usually the first gap a reviewer will find.
3. When does our adviser first see the records? Earlier involvement means fewer gaps to fill later.
4. How is each option priced, and on what basis? Compare fixed, time-based and benefit-linked pricing on the same claim.
5. Will the team actually use it? The best method is the one that gets used during the work.

For a plain-English overview of the rules themselves, see how the R&D Tax Incentive works. If you have a specific question about the rules, you can ask the R&D Tax Incentive chatbot — it explains the rules; your adviser makes the eligibility call. Software teams may also find our guide to software development and the RDTI useful.

This article is general information, not tax advice. Speak to a registered tax agent about your circumstances.

10Sources

• ATO — Helping you get R&D claims right
• ATO — Keeping records and calculating your notional deductions
• business.gov.au — Record keeping for the R&D Tax Incentive
• Tax Practitioners Board — Tax agent services
• Ambitious Australia: Strategic Examination of Research and Development (March 2026), as summarised in How much of the R&D Tax Incentive goes to advisers?