Standardized Contracts Drive Legal AI
Scott Stevenson, CEO of Spellbook, on building Cursor for contracts
Standardization is why contracts became one of legal AI's first real beachheads. A contract is usually judged by whether it matches approved language, flags risky clauses, and moves a deal forward faster, not by whether it sounds novel. That makes contract review look much more like bug finding in code or issue spotting in a checklist than litigation writing, where the value often sits in a new argument tailored to one messy fact pattern.
-
Spellbook is built around this repeatability. Teams load playbooks with preferred fallback language, then the product reviews a Word document, marks deviations with track changes, and proposes edits against company standards. That is a clean fit for AI because the target output is a narrower range of acceptable language, not open ended persuasion.
-
The split shows up in adjacent categories. eDiscovery has worked well because it is a large scale sorting problem across thousands of documents. Contract tools like Spellbook and Luminance extend that same pattern into redlining and clause analysis. Litigation drafting is harder because judges and partners care about case specific framing, tone, and strategic originality.
-
This is also why contract AI is fragmenting into workflow tools, while broader chat shaped legal products face more pressure from frontier models. Ironclad owns process around drafting, approvals, and repositories at roughly $200M estimated revenue, while faster growing specialists like Spellbook and Luminance focus on the review moment where standardized language and high document volume create immediate ROI.
The category is heading toward software that handles the whole contract lane, intake, first pass review, redlines, approvals, and ongoing monitoring. As models get better, the advantage shifts away from raw legal reasoning and toward products wired into Word, email, Slack, repositories, and company playbooks, where the work is repetitive enough for AI to take over larger parts of the flow.