A two-lawyer Gold Coast firm has scored a potentially significant land-tax victory against the Queensland Revenue Office after an in-house artificial intelligence system identified an argument that several senior barristers reportedly considered unwinnable.
The ruling is notable not just for its potential tax implications but for what it suggests about the changing economics of legal practice.
Ezylegal has only two lawyers, but its AI-supported workflow gave it the capacity to test an interpretation that more conventional legal advisers had apparently discounted. Drummond says the firm’s systems are largely built around Anthropic’s Claude Code and are designed to reduce the cost and time of legal work.
AI will not eliminate the need for legal judgment, advocacy or accountability. But it can allow smaller practices to conduct deeper statutory research, identify unconventional arguments, draft first versions of documents and manage litigation workflows at a cost that could once have required a much larger legal team.
Ezylegal, a small Burleigh Heads practice led by lawyer Michael Drummond, (pictured) used a homemade AI case-management tool known as “Rachel” to challenge the revenue office’s interpretation of Queensland’s land-tax legislation.
The Queensland Civil and Administrative Tribunal accepted the central argument: a property owner may still “use” land as their principal place of residence even while not physically living on it during a rebuild.
That distinction could materially affect land-tax assessments for homeowners whose rebuilding projects extend across more than one tax year.
The $350,000 land-tax dispute
The case involved Ezylegal client who owned an oceanfront property at Mermaid Beach on the Gold Coast.
The home had been exempt from land tax as the principal place of residence for the client, but after he and his wife demolished the house in 2020 for a full rebuild and lived elsewhere while construction continued, the Queensland Revenue Office assessed more than $350,000 in land duties in 2024.
The dispute turned on an exemption available to homeowners who move out to demolish or rebuild their residence.
The revenue office had taken the position that the exemption was effectively limited to 12 months because the owner needed to have physically occupied the property during the relevant tax year. On that view, homeowners whose projects took longer than a year could lose the benefit of the exemption.
That approach has obvious practical difficulties. Major rebuilds can be delayed by approvals, financing, weather, builder availability, supply-chain issues and contractual extensions. A strict 12-month limit may bear little resemblance to the realities of residential construction.
AI found a different reading
Ezylegal’s AI system challenged the assumption that “use” of a principal residence means actual physical occupation. The decision could have wider consequences for homeowners who vacate their principal residence while undertaking a demolition and rebuild, particularly where construction runs beyond 12 months.
The argument was that an owner can continue to use land as their principal place of residence despite temporarily living elsewhere, depending on the surrounding evidence. Relevant factors may include:
- Maintaining electoral enrolment at the original address
- Not establishing a permanent alternative residence
- Keeping possessions connected with the property or construction site
- Demonstrating an ongoing intention to return once the rebuild is completed
The tribunal accepted that broader, more practical interpretation.
Drummond said the firm’s AI system drafted the litigation material, while barrister was retained to present the case at hearing. The case is a useful reminder, however, that AI output alone does not litigate a dispute as lawyers must still assess the legal risk, gather admissible evidence, make strategic decisions, instruct counsel and remain professionally responsible for the case being run.
A signal for smaller AI-first firms
For clients, particularly in expensive practice areas such as tax, this could mean more viable challenges to official decisions that might previously have been abandoned as commercially uneconomic.
For law firms, it is another example of AI shifting competition away from headcount alone. A well-equipped specialist practice may increasingly be able to take on work traditionally reserved for larger firms with deeper research and drafting resources.
The case also provides a sharp illustration of where legal AI may have its greatest immediate impact – not replacing lawyers, but helping capable smaller firms find viable arguments, run leaner cases and challenge entrenched assumptions.
