About

How do you bring AI into decisions that carry real legal and organisational risk? The United Nations Office for Project Services (UNOPS) has built an AI-powered system to guide the analysis of personnel requests to review management decisions. Using a strict legal hierarchy and a six-step framework (receivability through remedy), the playbook evaluates litigation risk, guards against advocacy drift and false citations, and produces two outputs: an internal memo and a formal response letter following UN conventions, with final decisions remaining with people.

In this practical session, S. Nicole Jordan, General Counsel and John (Jack) Nairn, Legal Specialist, both of UNOPS, will demonstrate the system live using an anonymised example, explain the thinking behind its design, and share the steps taken to build and deploy it, including the guardrails around it. Members will leave with a rare inside view of AI applied to sensitive, high-volume work, and practical lessons for strengthening governance, consistency and decision-making in their own functions.