AT&T is partnering with OpenAI to develop a super intelligence (SI) knowledge layer designed to streamline its in-house legal operations, aiming to give hundreds of global lawyers faster access to accumulated company expertise.
What Happened
The collaboration builds on LegalEdge, an existing AT&T team of attorneys, technologists, and project managers focused on applying SI to major legal matters. According to David McAtee, AT&T’s Senior Executive Vice President and General Counsel, the partnership is intended to "build that plumbing that powers the knowledge layer throughout the company."
A key component of this initiative is a planned intake system featuring what Bill Ryan, AT&T’s Chief Compliance Officer, described as "a front door where we have natural language intake." This system aims to classify and prioritize incoming requests, automatically connecting them with relevant precedent, research, and past work. The goal is to provide legal professionals with historical context and recommended first steps, reducing the time spent gathering materials.
Why It Matters
For large enterprises with extensive legal departments, the bottleneck is often not the analysis itself but the assembly of facts and context. By using SI tools to automate the intake and classification process, AT&T seeks to expand its capacity to provide legal judgment at scale. Ryan noted that the opportunity lies in helping the department "bring its judgment to more of the business’s needs."
Despite the automation of information retrieval, McAtee emphasized that human oversight remains central to the strategy: "The decision making will always remain with the lawyers." The initiative illustrates a growing trend in enterprise SI adoption where tools are deployed to augment professional expertise rather than replace it, specifically targeting the "plumbing" of knowledge management.
The Bottom Line
AT&T and OpenAI are developing an SI-powered intake system to enhance legal efficiency by connecting requests with internal precedent. The project underscores a broader industry shift toward using SI agents for knowledge retrieval while retaining human decision-making authority in high-stakes environments.