Gilbert + Tobin, one of Australia's leading corporate law firms, reports 87% active usage of ChatGPT among enabled users, more than twice the rate it typically sees for other tools. The figures come from an interview with Partner and Chief Executive Officer Sam Nickless, published by OpenAI as a customer story. The firm, which works across capital markets, mergers and acquisitions, disputes and technology, began by letting staff experiment with ChatGPT under a strict rule that no client data or confidential information could be entered. It then introduced ChatGPT Enterprise in a controlled environment, citing Australian data residency, contractual protections, role-based access and administrative controls as the basis for scaling up. It later added ChatGPT Work and Codex to move from individual help toward defined, multistep workflows, with people responsible for review and approval. Nickless said the firm set no usage targets. Instead, leaders were expected to use the tools themselves, the business transformation team demonstrated workflows for different functions, and videos, challenges and an early 'AI bounty' surfaced ideas across the firm. He said the main value so far has been freeing staff for client meetings, team coaching and mentoring rather than reducing headcount. Nickless built a custom GPT trained on his writing and feedback, which he shared with his direct reports so they can test their thinking against his working style before meetings. He also said ChatGPT Work completed a confidential annual review involving 120 partners, updating files from a spreadsheet and producing bundled PDFs in minutes, at a usage cost of about $62 according to the firm's dashboard. The article is a vendor-published case study, so these figures are self-reported and not independently verified, and it does not disclose licence costs or measured productivity gains. For Australian businesses and professionals, the case is a practical template rather than a change in the rules. The Privacy Act 1988 applies to many organisations handling personal information, so before entering client or staff data into any AI tool, check where data is stored, what contractual protections apply and who can access it. The firm's sequence of a no-confidential-data rule, Australian data residency, role-based access and named human sign-off is a reasonable model to copy. The $62 task cost is small, but enterprise licence fees are not disclosed, so firms should cost a limited pilot before scaling.