Spellbook Building Background Contract Agents
Diving deeper into
Scott Stevenson, CEO of Spellbook, on building Cursor for contracts
what we'd call artificial employees: agents that are basically always working and surfacing risk for you
Analyzed 4 sources
Reviewing context
This points to Spellbook moving from a tool a lawyer opens on demand to a system that monitors the whole contract base between touches. Once contracts live in one repository with playbooks, workflows, and clause data attached, an agent can keep checking them against new laws, renewal dates, and company rules, then push a lawyer a short list of contracts that need action instead of waiting for manual review.
-
Spellbook already has the building blocks for this. It started as a Word add in, then added Playbooks for automatic first pass review and Associate for multi document tasks like syncing names, schedules, and formatting across deal packages. Artificial employees are the next step, the same logic running continuously in the background.
-
The competitive shift is from drafting help to workflow ownership. Ironclad owns repository, approvals, and post signature workflows across sales, procurement, HR, and finance. Luminance is pushing automated negotiation. A background risk agent only works if the vendor also controls storage, metadata, and the approval trail around each contract.
-
This also matches where legal AI is heading more broadly. Basic legal reasoning is getting cheaper as frontier models improve, so the durable value moves to software that sits inside the daily workflow, knows the company’s preferred fallback positions, and can act across many documents without a human prompting every step.
The next phase of legal AI will look less like chat and more like always on operations software. The winners will be the companies that turn contract data, negotiation rules, and document repositories into systems that catch risk early, route work automatically, and make legal teams feel larger without adding headcount.