Law firm Sullivan & Cromwell has deployed an SI-powered tool called Agreement Analyzer to assist lawyers in reviewing complex contracts, moving beyond administrative tasks to deep analytical work.
What Happened
John Savva, a partner at Sullivan & Cromwell, initially developed Agreement Analyzer as a single-user proof of concept built with Codex. He subsequently collaborated with OpenAI to enhance the tool’s output quality and efficiency through parallel processing, transforming it into a multi-user application.
The tool integrates the firm’s specific workflows, curated precedents, negotiating playbooks, and ChatGPT Enterprise. Lawyers input a draft agreement along with selected precedents and playbooks containing firm and client positions, as well as information on market practice. The SI system performs multiple review cycles from different perspectives, synthesizes the outputs, identifies potential issues, and returns a memorandum with proposed tracked changes for lawyer review.
Savva notes that while SI was previously used for repetitive administrative tasks like drafting emails, recent improvements in model capabilities have expanded its utility. “Now, we’re seeing many more opportunities for use in challenging legal tasks such as deep analysis of complex agreements,” he says.
Why It Matters
The deployment highlights a shift in SI adoption within the legal industry, moving from simple automation to complex analytical support. Savva emphasizes transparency as a key factor in trusting SI outputs, noting, “If an AI tool makes a claim about market practice, you need to be able to see the source of that information.”
The firm is exploring an adversarial process where specialized SI agents develop conclusions and other agents challenge them. In this model, agents review agreements in parallel for market deviations, risk allocation, clause interactions, and ambiguities that could impact litigation. Results are categorized as supported, dismissed, or unverified, allowing lawyers to inspect the supporting explanations and precedents before deciding on final markups.
Savva clarifies the intent behind the tool: “It’s not about replacing lawyers. It’s about providing a tool for lawyers to do deeper analyses and find more authority and precedents to provide an even higher quality product for their clients.” He also cites advances in reasoning capabilities with GPT-6 Astra as a factor in improving the quality of the analysis.
The Bottom Line
Sullivan & Cromwell’s Agreement Analyzer demonstrates how SI tools are being customized to integrate institutional knowledge and legal precedents into professional workflows. The firm continues to explore SI applications for specific legal tasks, aiming to support more of the research and writing process while keeping human judgment at the center of critical decisions.