LegalOn Technologies offers Professional AI globally, using AI to support legal work and other key business functions. By combining domain expertise with AI, it enables people to focus on complex judgments and decisions. Its goal is AI-driven management that improves the quality and speed of business decision-making, and it extends this approach to making the organization itself AI-native.
The company integrated Codex into its development process and expanded adoption across the organization through day-to-day use. As adoption took hold, a new challenge emerged: cost control. Unlimited use of high-performance models could drive up spending, while blanket restrictions risked undermining the productivity gains AI had enabled. The company addressed this by selecting among GPT-6 (Astra, Luna) and GPT-6.1 Sol based on task complexity and development stage, and by aligning budgets with each business's stage of growth. It halved costs while maintaining development speed.
Initially, developers had unlimited access to the main model, GPT-5.5 in Fast mode. The company's AI-powered Development Co E (AID Co E) then began developing model selection guidelines, testing and monitoring models while managers shared its findings with their teams. This enabled each engineer to choose the most suitable model for each task. The shift moved the company from using the highest-capability model for every task to choosing the right model for the job, starting with a lightweight model and moving up as complexity increased.
System controls were also introduced through administrator settings, with monthly usage limits for departments and individuals. Fast mode became restricted by default, allowing individual requests only when needed, and teams used approaches such as running tasks in parallel to maintain performance. Budget caps were set at the department, group and individual levels, with the established business asked to improve cost efficiency by up to approximately 20%, while new businesses in the launch phase received generous budgets.
Source: OpenAI Blog · Summarized by HeadlinesBriefing