International law firm Cooley has introduced GO Public, a proprietary AI workflow designed to support the preparation phase of initial public offerings. The system is built with ChatGPT Work and is intended to help legal teams organize and analyze large volumes of transaction material before lawyers review the output.
See also: official website
From general AI assistants to transaction-specific systems
The important shift is not simply the use of a language model inside a law firm. GO Public combines client-provided information, relevant public material and curated precedents inside a controlled workflow built around capital-markets work. Automated steps handle part of the initial analysis, while lawyers remain responsible for review, validation and professional judgment.
That architecture reflects a broader enterprise pattern: organizations are moving beyond standalone AI assistants and embedding models into defined business processes with explicit checkpoints. In regulated or high-stakes work, the value proposition increasingly depends on how well the workflow constrains, verifies and routes model output rather than on model capability alone.
IPO preparation becomes an AI workflow test case
IPO preparation involves extensive disclosure, document review and coordination across legal and management teams. Cooley says the new system is designed to surface relevant issues earlier and reduce the amount of time spent on first-pass information sorting. Those efficiency benefits are company-reported and have not been independently quantified in the announcement.
The firm also says it advised on 180 transactions globally in 2025, representing more than $51.5 billion in deal volume. That scale gives the project significance beyond a small internal pilot: it places agentic AI inside a professional-services workflow where errors, governance and accountability carry direct commercial consequences.
The competitive asset may be the workflow, not only the model
For professional-services firms, the emerging differentiator may be the proprietary layer built around general-purpose AI: domain rules, curated knowledge, review gates and institutional process design. As access to capable models becomes more widespread, firms that can turn expertise into repeatable, auditable workflows may gain more durable operational advantages than those relying on generic copilots alone.
