For general counsel · In-house legal teams
Your department's constraint isn't drafting. It's the front door.
Most legal AI pitches to departments start with contract drafting, because that's where the visible volume sits. In department after department, the actual bottleneck sits further upstream - at intake, before any lawyer has even opened the request.
The economics run opposite to a law firm's
Three things true of every department we've worked with.
Demand is unbounded. Headcount is not.
A firm sells hours and must generate demand. A department absorbs demand it cannot decline, against a budget it does not control. Efficiency in a firm reduces revenue. In a department it is the only lever that exists at all.
You cannot select your own tooling
Security review, data-residency assessment and procurement gate anything new, often measured in months rather than weeks. A full AI-native rebuild is rarely honestly available, and pretending otherwise wastes a quarter before the first tool even clears review.
Nobody can currently answer 'what would two more headcount buy'
Most departments have worse throughput data than an equivalently sized firm, because there was never a billing model forcing anyone to record it. That makes the case for more resourcing impossible to make in numbers a CFO will act on.
A firm automates and must decide what to do about the revenue it just deleted. A department automates and immediately serves a queue it was already failing to serve. The second case is much easier to justify and much harder to argue against.
The six primitives, translated for a department
Same doctrine. Different words for the same six pieces.
| Primitive | In a legal department |
|---|---|
| Matter | The request or contract - often a Teams message with no container at all. |
| Actions | The named steps of a request type: NDA review, vendor contract, employment query. |
| Time | Cycle time and capacity, not billable duration - how long a request waits, and how much throughput exists. |
| Team | Doer, reviewer, decider - plus the business-side approver, who is frequently the real bottleneck. |
| Evidence | Every assertion anchored to source, with added weight since departmental advice often ships with no external review. |
| Research | A compounding stock, plus a self-service layer so the business answers routine questions without opening a request. |
Departments like yours, already transformed
Two in-house transformations, both AI Adoption.
Case 13 · India · 6-lawyer department · 8 weeks
In-house legal department, mid-sized manufacturing group
A six-lawyer department inside a manufacturing group serving eleven business units, absorbing legal requests through whatever channel a business stakeholder happened to reach for. Eight weeks later the department runs one front door, a classified request taxonomy, a deflection layer answering routine questions automatically, and a contract-review agent inside the company's existing procurement-approved tooling - all installed within a corporate architecture the department itself had no authority to change.
Case 07 · Singapore · 9-lawyer department · 8 weeks
In-house legal department, technology company
A nine-lawyer department absorbing every legal question in a fast-growing company, through five uncontrolled channels. We closed the front door to one, classified the requests, and deflected the routine third before it ever reached a lawyer.
Timelines run to your procurement
We tell you honestly at intake what your security review and procurement cycle mean for the timeline. The naming and systematising work needs no procurement at all, so it can run in parallel with your approval track from week one.
Read our security posture →Start with the front door, not the drafting tool