Gilbert + Tobin says 87% of staff enabled to use ChatGPT are active users, after the Australian law firm expanded access to OpenAI tools across its business. Its account offers a concrete look at how a confidentiality-sensitive profession is putting generative AI to work, with lawyers still responsible for checking and signing off their work.
OpenAI Watch analysis
What happened
The firm says its rollout includes ChatGPT Enterprise, ChatGPT Work and Codex. It began with staff experimenting on the consumer version of ChatGPT under a rule barring client and confidential information, then moved to a controlled Enterprise environment with Australian data residency, contractual protections, role-based access and administrative controls.
According to Gilbert + Tobin chief executive Sam Nickless, adoption progressed from individual productivity to organisational workflows across operations, marketing, recruitment, finance and parts of the legal practice. The firm says Codex is being used for multi-step tasks under human review. Nickless also described using ChatGPT Work to update and bundle materials for an annual review involving 120 partners, at a reported cost of about $62.
Why it matters
The 87% figure is a striking adoption claim from a law firm, where confidentiality and professional accountability are not footnotes. Gilbert + Tobin says it did not set usage quotas; instead, it used function-specific workflows, internal demonstrations and leadership participation to encourage use.
The firm’s examples point to a model built around governed access and human sign-off, rather than handing legal judgement to a chatbot. Its reported figures and outcomes are the firm’s account, but they give other professional-services organisations useful specifics to compare against their own deployments.
Our read
This is more than a tool trial: Gilbert + Tobin describes broad access, operational safeguards and agentic workflows already being tested in real work. The useful test is whether those workflows reliably give staff more time for complex work without making confidentiality and checking someone else’s problem. A digital twin of the chief executive is certainly a memorable use case; keeping it out of the decision-making chair is the more important detail.
What to watch
- Whether the reported 87% active-use rate holds as access broadens.
- Which workflows the firm expands beyond its initial teams.
- Whether the firm publishes more detail on time saved, output quality or client outcomes.
- How it maintains human review as Codex workflows take on more steps. Activity teaser Gilbert + Tobin says 87% of staff enabled to use ChatGPT are active users, with ChatGPT Enterprise, ChatGPT Work and Codex in use across the Australian law firm. Its account describes safeguards including Australian data residency and role-based access, while lawyers remain responsible for checking outputs. The reported 87% is a firm-reported figure, but the deployment offers a concrete case study in bringing AI into confidential professional work. Can broad AI adoption in law be made routine without weakening confidentiality or professional accountability?
Discussion spark: Should law firms make AI use a default part of everyday work when lawyers retain sign-off, or should confidential work face stricter limits even inside a controlled system?
Sources and evidence
- Australian Law Firm Gilbert + Tobin Rolls Out OpenAI Tools, Logs 87% Active Usage – BigGo Finance (12 November 2026, 00:35 UTC)
OpenAI Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by OpenAI.