OpenAI Lawyer Builds Tools to Automate Legal Work
Business Insider reports that Nicole Diaz, an associate general counsel at OpenAI, learned to build small apps using ChatGPT and Codex after joining the company. Diaz told Business Insider she had never written code before OpenAI; over the past year she used the tools to convert law-firm memos into plain-English policies, triage employee email, draft replies, and track outcomes. She is quoted saying, "My sense of what's possible has rapidly expanded in the last six months, even three months." Business Insider describes these tools as automating repetitive legal tasks while Diaz retains legal judgment.
What happened
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.
Editorial analysis - technical context
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 context
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.
What to watch
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.
Scoring Rationale
This is a practical, user-focused example showing how code-generation and agent tools enable domain experts to automate tasks. It is useful for practitioners but not a frontier-model or platform announcement, so its impact is modest.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems