Gilbert + Tobin scales ChatGPT Enterprise and Codex with CEO-led governance

OpenAI published a customer case study on Gilbert + Tobin, an Australian corporate law firm working across capital markets, mergers and acquisitions, disputes and technology. Rather than change how it delivers legal advice, the firm set out to raise the standard of the operational work supporting that advice. ChatGPT Enterprise was introduced first to operations teams to test practical use cases in a controlled setting, then expanded into marketing, business development, recruitment, finance, technology, business transformation and parts of the legal practice. With Codex, the firm moved further, from AI that assists individuals with single tasks to AI that carries out defined steps across larger workflows.

Adoption was driven from the top. CEO Sam Nickless introduced ChatGPT to staff using examples from his own work, framing it with the line "AI is not cheating," and other senior leaders became active users who shared examples with their teams. The business transformation team backed that leadership example with role-specific enablement, joining individual team meetings to demonstrate workflows tailored to each function, plus internal videos, showcases, challenges and use-case stories. The firm did not mandate usage targets; it focused on showing employees what the technology could do while holding people accountable for the result. As of June 2026, 87% of enabled ChatGPT Enterprise seats were active, more than twice the adoption rate Gilbert + Tobin typically sees for other tools, with active users including partners and staff outside the technology function.

Governance sat alongside that rollout. Because the firm handles confidential client and business information subject to varying client requirements, it paired the rollout with guidance on approved tasks, what employees could enter, and how outputs should be reviewed, alongside assessment of contractual protections, role-based access, data-processing requirements and administrative controls. Moving to an OpenAI environment with Australian data residency gave the firm more confidence to widen access while meeting internal requirements and client expectations.

Day-to-day use spans several functions. ChatGPT cut a recruitment research and data-extraction task from about four hours to about 20 minutes, and a separate reference-processing workflow saves an estimated 25 minutes per candidate for a team handling high recruitment volumes. Marketing and business development teams use it to synthesize past pitch materials and tailor new ones across the 400 to 500 pitches the firm produces each year, plus deep research for industry and client reports. Finance teams use it for spreadsheet and data-analysis work, and technology teams apply it to documentation, scripts and training materials. Lawyers who use ChatGPT draw on it only for the operational side of their work; legal-specific workflows run instead through separate approved platforms such as Harvey. Across every use case, people remain responsible for constraining the task, checking outputs, applying professional judgment and approving the final work product. One distinctive use case is a custom GPT built from approved examples of Nickless's writing, priorities, background and feedback, which executives can use to pressure-test ideas before taking up the CEO's own time; the firm describes it as a "digital twin" that does not make decisions or speak on his behalf.

Codex has been applied to defined, multistep operational work under human review. It prepared audit reports covering 300 entities, avoiding a full day of manual work across that workflow, and checked and renamed 1,100 files for upload into another system, replacing work that previously took days. The business transformation team used Codex to turn requirements and an existing specification into a working Python application. The firm also used Codex to help build an AI-enabled workflow for selected conflict, anti-money-laundering, politically-exposed-person and know-your-customer checks, which now produces a report for human review and sign-off; selected checks that previously took up to eight hours can now be completed in minutes. Within the technology function, a DevOps team member used Codex to build a monitoring "watchtower" for the firm's AWS environment that consolidates operational signals, supports diagnosis and remediation, and escalates issues needing human attention.

Gilbert + Tobin intends to extend access to ChatGPT Work and Codex as it puts in place the controls needed for broader use, with a longer-term goal of an interconnected environment where staff can draw on approved organisational context, carry out operational tasks and deliver finished outputs without moving manually between systems.

Key facts

87% of Gilbert + Tobin's enabled ChatGPT Enterprise seats were active as of June 2026, more than twice the adoption rate the firm typically sees for other tools.
ChatGPT cut a recruitment research and data-extraction task from about four hours to about 20 minutes, and saves an estimated 25 minutes per candidate on reference processing.
Codex prepared audit reports covering 300 entities, checked and renamed 1,100 files for a system upload, and built a workflow that cut selected conflict, AML, PEP and KYC checks from up to eight hours to minutes.
Lawyers use ChatGPT only for the operational side of their work; legal-specific workflows run through a separate approved platform, Harvey, not ChatGPT.
The firm set no mandatory usage targets, instead pairing CEO-led example-setting with role-specific enablement sessions run by its business transformation team.

Why it matters

The case study is OpenAI's pitch for how an enterprise should roll out generative AI at scale without loosening accountability: visible CEO example-setting instead of mandates, role-specific enablement instead of generic training, and governance built in from the start rather than bolted on afterward. For a law firm, whose work depends on client trust and the secure handling of sensitive information, the model is deliberately conservative: AI is expanded into operational work while professional legal judgment and client-facing legal advice stay untouched, and every output stays subject to human review and sign-off.

Who it affects

The rollout touches Gilbert + Tobin's operations, marketing, business development, recruitment, finance, technology and business transformation teams, plus parts of its legal practice. Executives named in the piece are CEO Sam Nickless, CIO Mitch Owens, CMO Daniel Quinn and Head of Business Transformation Aviva Leitch. Lawyers are included only for the operational side of their work; their legal-specific tasks are routed to a separate platform, Harvey, rather than ChatGPT or Codex.

How to use it

ChatGPT Enterprise handles research, drafting, spreadsheet and data analysis, documentation and deep research across functions, plus a custom GPT built from approved material that lets executives pressure-test ideas against a model of the CEO's priorities before raising them with him directly. Codex is reserved for defined, multistep operational workflows under human review: audit-report preparation, bulk file processing, building a Python application from a specification, a conflict and AML/PEP/KYC check workflow, and an AWS monitoring tool. The firm plans to extend access to ChatGPT Work and Codex as it puts further controls in place; no pricing, contract value or seat-count figures are given.

How solid is it

This is a customer story published directly by OpenAI, not independent reporting, so its figures come from the vendor and the client rather than a third party. The specific numbers, the 87% active-seat rate, the time-savings figures, the counts of entities and files Codex processed, all carry direct quotes or figures attributed to the firm or its executives, and named executives are quoted by name and title. The piece does not give total headcount or total seat counts (only the 87% figure), no cost or ROI figures, and no date for when the rollout began beyond "when generative AI began reshaping knowledge work."

Risks and caveats

As a vendor-published case study, the piece is promotional and offers no independent corroboration from Gilbert + Tobin itself or any third party. It does not disclose what the program costs, how many total seats or employees are covered, or how mature the AI-enabled AML/KYC and audit workflows are outside the specific examples cited. The firm's own framing keeps legal-specific work off ChatGPT and Codex entirely, and stresses that all outputs from both tools require human review and approval, which limits how far the described gains generalize beyond the operational tasks described.

“AI is not cheating. It gives our people another way to apply their judgment and improve how the firm operates. In a profession built on trust and accountability, that starts with strong governance. OpenAI provides the enterprise-grade foundation that allows us to move forward with confidence.”

— Sam Nickless, Chief Executive Officer, Gilbert + Tobin