OpenAI Lawyer Builds Tools to Automate Legal Work

Business Insider reports that Nicole Diaz, an associate general counsel at OpenAI, learned to build internal tools using ChatGPT and Codex after joining the company. Per Business Insider, Diaz, who had not previously written code, built agents and small apps to turn dense law-firm memos into plain-English policies, triage employee email, draft replies, and track results. Business Insider quotes Diaz: "My sense of what's possible has rapidly expanded in the last six months, even three months." The article states Diaz uses these tools to absorb repetitive tasks while preserving her professional judgment.

Industry-pattern observations: code-generation and low-code agents such as Codex lower the technical barrier for domain experts to build automation, enabling non-developers to assemble workflows that combine extraction, summarization, and templated drafting. For practitioners, this typically shifts effort from writing production-grade software to validating prompts, designing prompt-to-code handoffs, and adding guardrails for accuracy and auditability.

Industry-pattern observations: legal and compliance functions are common early adopters for automation because they produce high volumes of repetitive text work-summaries, policies, and templated responses-where time savings compound across many users. Business Insider frames Diaz's experience as an example of an in-house user converting model outputs into operational processes while retaining human review for judgment-sensitive steps.

For practitioners: monitor how teams add verification, versioning, and audit trails around model-generated legal text; watch for workflow patterns that separate claim-generation from legal sign-off. For teams evaluating similar approaches, key indicators are error rates on summary tasks, time-to-review improvements, and controls for sensitive-data handling when using external APIs.