Cooley's partnership with OpenAI to build GO Public is not a minor product release. It is a deliberate play to own the S-1 automation space at scale. The tool ingests company documents, financial statements and structured questionnaires, then produces workable first drafts of major registration statement sections. For pre-IPO companies, this cuts weeks from the regulatory filing cycle. For Cooley, it creates a moat: clients who trust the AI with their S-1 are more likely to stay for ongoing capital markets work. The implication for UK regulated firms—especially mid-market law practices, insurance firms handling product disclosure, and accountancy practices supporting fundraising clients—is stark. The work that currently requires senior associate time to research, structure and draft is now automatable. If your firm cannot offer this capability, you become a cost centre, not a strategic adviser.

This is the third wave of legal AI, and it separates the movers from the stuck. First came contract intelligence tools like Luminance and Harvey—good at spotting risks and classifying documents, but passive. Second came ChatGPT Enterprise and generic LLM wrappers—powerful but ungrounded, prone to hallucination, and legally indefensible without human oversight. Now comes the integrated play: domain-specific AI agents that work within known document templates, validated business rules, and regulatory frameworks. Cooley's GO Public sits squarely in this third wave. It does not try to make law; it automates the reliable, repeatable parts of drafting at scale. Every major law firm with capital markets, regulatory or transactional capacity is building or licensing something equivalent. UK firms are watching. Some are experimenting quietly. Most are still deciding. The decision window is open now. By 2027, it will be closed.

Here is Trovix's honest view: GO Public works because it operates within a known constraint. S-1 registration statements follow a strict template set by the SEC. The document structure, required sections, and regulatory boundaries are fixed. Feed the AI clean data, and it produces a workable first draft. This is fundamentally different from asking an AI to 'automate legal work'—which remains dangerous. The risk Cooley has managed, and which many UK firms will mismanage, is the gap between generation and validation. A generated S-1 section still requires senior review. A generated insurance product disclosure document still requires FCA-aligned sign-off. A generated tax return still requires qualified scrutiny. The firms that succeed with AI in regulated environments are those that treat AI as a draft engine, not a replacement engine. Trovix Sift exists precisely because document extraction and intelligent drafting must be paired with explainability and audit trail. You need to know why the AI suggested what it suggested, and you need to prove to the FCA (via FCA Consumer Duty PS22/9), the SRA Code, or the ICO (via UK GDPR), that a human with accountability reviewed and approved the output before it left your firm.

What should a mid-market law firm, insurer or accountancy practice do right now? First, inventory your high-volume, templated document work. Identify the pieces that could be reliably automated: S-1 sections, insurance policy schedules, tax schedules, compliance memos. Second, run a pilot with a tool that combines document intelligence with compliance governance. Third, build the review process before you build the automation—define who signs off, what they check, and how you log it. Fourth, do not wait for a perfect in-house solution. Cooley partnered with OpenAI because speed mattered more than building alone. Your firm should evaluate whether to build, partner, or integrate existing platforms that already handle the governance layer. The firms that act in the next six months will own the next three years of efficiency gains. The firms that wait for certainty will become expensive by accident.

Source: Forbes