AIToday
Large Language ModelsAI Business & IndustryOpenAI BlogPublished: Sep 2, 2026, 13:00 JST3 min read

Gilbert + Tobin scales AI with OpenAI, 87% seat activation

Gilbert + Tobin scales AI with OpenAI, 87% seat activation

Key takeaway

  • Gilbert + Tobin, a top Australian law firm, reports 87% of its ChatGPT Enterprise seats were active as of June 2026.

  • That is more than double its typical adoption for other tools.

  • The firm scaled AI with leadership example and strict governance.

3 Key Points

  1. What happened

    Australian law firm Gilbert + Tobin reports that as of June 2026, 87% of its enabled seats for ChatGPT Enterprise were active—more than double the adoption rate it typically sees for other tools. The firm has also used OpenAI's Codex to automate tasks like preparing audit reports for 300 entities and renaming 1,100 files.

  2. Why it matters

    The firm paired visible leadership support—CEO Sam Nickless told employees 'AI is not cheating'—with governance and Australian data residency to expand AI use across marketing, finance, recruitment, and parts of legal work. These measures helped build trust while meeting client expectations for handling confidential information.

  3. What to watch

    The firm plans to extend access to ChatGPT Work and Codex as it puts in place controls for broader use. Its longer-term goal is an interconnected environment where employees use ChatGPT to access approved context and complete workflows without manual system navigation.

Ask the AI about this article →

Summaries like this, in your inbox every morning.

Context & Analysis

Gilbert + Tobin's AI rollout stands out because it deliberately avoided mandating usage targets. Instead, the firm focused on demonstrating what ChatGPT could do in role-specific workflows, with leaders like CEO Sam Nickless using the tool themselves and telling staff that 'AI is not cheating.' This approach, combined with a controlled introduction that started in operations teams, helped adoption spread naturally across marketing, finance, recruitment, and parts of legal work—reaching 87% active seats on enabled accounts as of June 2026, more than double the firm's typical tool adoption rate.

The firm's governance framework—clear guidance on approved tasks, role-based access, data-processing requirements, and Australian data residency—was key to scaling access while satisfying client expectations for confidential information handling. ChatGPT has delivered concrete time savings, such as cutting a recruitment research task from four hours to 20 minutes, and Codex has automated irregular tasks like preparing audit reports for 300 entities and renaming 1,100 files. These use cases show the firm progressing from AI as an assistant for discrete tasks to AI that executes defined steps in operational workflows, always under human review.

The ultimate test for Gilbert + Tobin will be whether it can extend this controlled expansion to broader access across ChatGPT Work and Codex without eroding the governance and accountability it has built. The firm's ambition to create an interconnected environment—where employees use ChatGPT to access approved context and complete workflows without manual system navigation—hinges on maintaining that same balance of enterprise-grade controls and professional judgment. As more legal and professional services firms explore similar AI adoption, Gilbert + Tobin's model of combining visible leadership, strict governance, and measurable operational gains could serve as a reference point, though its specific results depend on its unique culture and client requirements.

FAQ

What specific time savings did Gilbert + Tobin see from using ChatGPT?
It reduced a recruitment research task from about four hours to about 20 minutes. A separate reference-processing workflow saves an estimated 25 minutes per candidate.
How is Codex used at Gilbert + Tobin?
Codex has prepared audit reports covering 300 entities and checked/renamed 1,100 files. It also helps complete selected conflict, anti-money laundering, and know-your-customer checks in minutes instead of up to eight hours.
What is the 'digital twin' of the CEO?
It is a custom GPT built from approved examples of CEO Sam Nickless's writing, priorities, and feedback. Employees can test ideas with it before going to Nickless, but it does not make decisions or speak for him.

Get the latest Large Language Models news every morning

For example, today's edition would include:

  • OpenAI says GPT-6 Astra 'low' beats GPT-5.6 Sol 'high'ITmedia AI+ · 3h ago
  • OpenAI reveals AI agents accelerating research at 3.1× human paceITmedia AI+ · 6h ago
  • OpenAI agents hack German site, incident undisclosedSemafor Tech · 6h ago

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleAI agents outpace human security teams, forcing endpoint defenses