AT&T is working with OpenAI to change how its in-house legal team handles everyday requests. According to OpenAI, the telecom giant’s LegalEdge team is using AI to rethink legal intake and to “scale legal judgment” across the company. The goal, as OpenAI describes it, is to “make legal knowledge easier to access” for the employees who rely on the legal department every day.
The announcement is short on hard numbers. Still, it’s worth paying attention to, because it shows where enterprise AI is moving: out of pilot projects and into the core workflows of corporate departments.
⚡ The Quick Version
- Who: AT&T’s LegalEdge team, working with OpenAI.
- What: Using AI to rebuild legal intake, meaning how employees bring questions and requests to the legal department.
- Why it matters: Legal teams are a classic bottleneck in big companies. If AI can answer routine questions and send the complex ones to the right lawyer, the whole business moves faster.
- What to watch: Whether other Fortune 500 legal departments copy the model, and how much of the work stays with human lawyers.
🏢 Why Legal Intake Is the Right Starting Point
In-house legal teams at large companies field a constant stream of requests. Can we sign this NDA? Does this marketing claim need review? Which contract template do I use? A lot of these questions repeat, and many have answers that already exist somewhere in policy documents, past decisions, or a senior lawyer’s head.
That’s the core problem AT&T is going after. Intake is where requests pile up, get misrouted, or sit waiting for someone to triage them. It’s also low-risk compared to, say, letting AI draft litigation strategy. Starting there makes sense.
What stands out here is the phrase “scaling legal judgment.” That’s a bigger ambition than a chatbot that answers FAQs. It suggests AT&T wants to capture how its lawyers think through common issues and make that reasoning available to the whole organization, without a lawyer touching every single request.
📈 How This Fits the Bigger Picture
Legal has become one of the most active areas for generative AI. Here’s the context:
- Legal AI startups are booming. Companies like Harvey, which builds on OpenAI’s models, have raised large rounds selling to law firms and corporate legal teams.
- Established players moved in. Thomson Reuters and LexisNexis have both rolled out AI assistants for legal research and drafting.
- In-house teams are the next wave. Law firms got most of the early attention. Corporate legal departments, which care more about speed and cost than billable hours, are a natural fit.
The status quo before tools like this was simple: email the legal team, wait, and hope your request lands with the right person. Some companies built ticketing systems or self-serve knowledge bases, but those rarely kept up with how people actually ask questions. Large language models change that because they can read a messy, plain-English request and figure out what’s being asked.
OpenAI also benefits from stories like this. It’s been pushing hard into enterprise sales, and a recognizable brand like AT&T using its tools in a high-stakes function like legal is a strong signal to other big buyers.
⚖️ The Open Questions
The source doesn’t go into deep technical detail, so a few questions remain:
- Where’s the line between AI and lawyers? Legal advice carries real liability. Expect AT&T to keep humans in charge of anything beyond routine guidance.
- How is accuracy checked? AI models can still produce confident wrong answers. In legal work, that’s a serious risk.
- What about confidentiality? Legal data is some of the most sensitive information a company holds. Enterprise deployments usually come with strict data controls, and legal teams will want proof those controls hold up.
🧭 What This Means for You
If you work in or alongside a corporate legal team, this is a preview of where things are going:
- Audit your intake process. Figure out which requests repeat most often. Those are the first candidates for AI help.
- Document your judgment. AI can only scale knowledge that’s written down. Playbooks, policies, and past decisions become far more valuable.
- Expect role shifts. Lawyers will likely spend less time on triage and routine answers, and more on complex, high-value work.
AT&T’s move suggests the next phase of enterprise AI isn’t about flashy demos. It’s about fixing the slow, unglamorous workflows that quietly drag on big companies. Legal is one of the first places that’s happening at scale. You can read more about how LegalEdge is approaching it in OpenAI’s original write-up.