Workflow Replaces Model Moat
Luminance
The moat in legal AI has shifted from having the smartest model to owning the workflow, the data, and the distribution. Luminance was early in using machine learning for contract review, but GPT-4 era tools made drafting, clause extraction, and redlining much easier for rivals to add. That changed competition from pure model quality to who controls the daily system lawyers already use, whether that is Word, a CLM, or a legal research platform.
-
Ironclad shows how AI features can be absorbed into an existing contract system of record. Its core product already manages approvals, repositories, renewals, and cross functional workflows for legal, sales, procurement, HR, and finance, so adding AI strengthens a deeply embedded workflow instead of selling a new point tool.
-
Newer companies like Spellbook and Legora now cover much of the same visible surface area, Word based drafting, clause review, extraction, and playbooks, often by using frontier models plus legal workflow software. That narrows any advantage from a proprietary legal model alone, especially when customers care most about speed to value inside familiar tools.
-
Incumbents have also moved down the stack and up the workflow. Clio bought vLex for $1B to pair AI with a large legal research corpus and its practice management base, while Thomson Reuters and LexisNexis are bundling AI into entrenched research products and long standing law firm relationships.
The next phase favors legal AI companies that can turn one successful task into a wider operating layer. For Luminance, that means using contract review as the entry point, then expanding into negotiation, procurement, compliance, and post signature workflows quickly enough that it becomes harder to swap out than the model underneath it.