Contract review that keeps pace with a growing deal flow
A review agent that checks contracts against the firm’s playbooks, redlines standard deviations and summarizes risk before a lawyer ever opens the file.
- Client
- Atlas & Vale
- Industry
- Legal
- Duration
- 8 weeks
- Services
- AI Strategy, AI Agent Development, Workflow Design

- Less first-pass review time
- 75%
- NDA turnaround, from four
- 1 day
- Contracts reviewed each quarter
- 2,600
- Clause detection accuracy
- 99%
Where the work was stuck.
Junior lawyers at Atlas & Vale spent hours on first-pass review of NDAs, supplier agreements and engagement letters. As client deal flow grew, the backlog started delaying signatures.
Partners wanted faster turnaround without compromising the firm’s standards or its liability position.
- Average first-pass review
- 3 hrs
- NDA turnaround
- 4 days
- Contracts each quarter
- 2,600
What Cortex built.
We encoded the firm’s negotiation playbooks into a review agent that extracts clauses, compares them with approved positions and drafts redlines for standard deviations.
Each contract arrives on a lawyer’s desk with a risk summary and suggested edits, routed by practice area and urgency.
How the system works, step by step.
- 01Trigger
Contract received
Client portal or email
- 02AI agent
Clause extraction
Every term identified
- 03System
Playbook comparison
Approved firm positions
- 04AI agent
Redline and summary
Risk ranked, edits drafted
- 05Human
Lawyer review
Judgment and sign-off
- 06Action
Return to client
Tracked to signature
First-pass review time dropped by three quarters, NDAs now go back the same day, and lawyers spend their hours on negotiation and advice instead of comparison.
Our lawyers trust the first draft because it follows our own playbook. Clients simply see that we are faster, and that matters.

Halden Freight
Automating freight invoice reconciliation across 40,000 shipments a month

Facing a challenge like Atlas? Let’s talk.
Book a 45-minute consultation. We will map your highest-value automation opportunities and show you what a first system could look like, at no cost.