Skip to content
Curious Context Labs

case study · Legal and compliance · Agent

Reviewing customer contracts against your own playbook in hours, not days

How written positions on each clause can drive tracked-change redlines and a one-page risk summary for a person to decide on.

business
SaaS
build
Agent
timeline
6 weeks
read
3 min

01 the business

B2B software company

80 staff, no in-house lawyer

Enterprise customers send their own paper. The founders either sign with gaps or wait days for outside counsel on routine terms. The company's preferred positions on liability, payment and termination are not written down anywhere.

02 before

The problem

Customer and supplier paper arrives in every format, and someone has to read all of it line by line. Without in-house counsel, founders or ops leads either sign with gaps or wait days for outside lawyers on routine terms. Your preferred positions on liability, payment and termination are rarely written down in one place.

  • Routine contracts wait days for review.
  • Positions vary by who negotiated.
  • Outside counsel bills for repetitive reading.

03 the build

What the AI does

  1. Step 1: Captures the playbook first: your standard positions, fallbacks and walk-away points for each clause type.
  2. Step 2: Picks up contracts from email, Salesforce opportunities or a shared SharePoint folder.
  3. Step 3: Compares each clause with the playbook and marks it as acceptable, needs a fallback or outside your limits.
  4. Step 4: Drafts tracked-change redlines in Word with a short comment explaining each change.
  5. Step 5: Writes a one-page summary of open risks and the clauses that need a decision from you.

paused: waiting for a person to approve

How it was put together

  • Wrote the playbook first with outside counsel: positions, fallbacks and walk-away points per clause type.
  • Picked up contracts from email and CRM opportunities.
  • Marked each clause as acceptable, needs a fallback or outside limits.
  • Drafted tracked-change redlines with comments and a one-page risk summary.

04 human in the loop

Where a person stays in charge

A named reviewer, either your lawyer or the contract owner, accepts or rejects every redline before anything goes back to the other side. Clauses outside your limits are always escalated to a qualified lawyer.

  • A named reviewer accepts or rejects every redline.
  • Clauses outside limits always go to a qualified lawyer.
  • The agent never sends anything to the other side.

05 rollout

From kickoff to live in 6 weeks

  1. Weeks 1–3Build the playbook with counsel from twenty past negotiations.
  2. Weeks 4–5Review live contracts in parallel with counsel and compare.
  3. Week 6Use as the first pass on all customer paper.

06 after

What changes

Routine contracts reach a first marked-up draft quickly and in your own positions. Your lawyer's time goes on the clauses that carry real risk, not on reading boilerplate.

worked example · your numbers will differ

Assume 15 customer contracts a month, each needing 3 hours of outside counsel at $400 an hour, falling to 1 hour on the escalated points.

15 × 3 × $400 = $18,000 a month today. 15 × 1 × $400 = $6,000. About $12,000 a month saved, and faster deals.

What to measure

  • Days from contract received to first response.
  • Outside counsel hours per contract.
  • Share of redlines accepted by the reviewer.
  • Deals delayed by legal review.

07 evidence

What the research says

  1. The average business loses almost 9% of contract value each year through poor contract management; the best lose about 3%, the worst 15% or more.

    World Commerce & Contracting, Contract management: an overlooked driver of business agility and financial performance (opens in a new tab) · 2025

  2. A US federal judge fined two lawyers and their firm $5,000 for filing a brief that cited cases ChatGPT had invented.

    ABA Journal, Lawyers who defended ChatGPT's fake cases must pay $5K, judge says (opens in a new tab) · 2023

  3. In a field experiment with 758 BCG consultants, those using GPT-4 on tasks within its abilities finished 12.2% more tasks, 25.1% faster and at over 40% higher quality, but did worse on a task outside its abilities.

    Harvard Business School / BCG, Navigating the Jagged Technological Frontier (Dell'Acqua et al.) (opens in a new tab) · 2023

08 lessons

Mistakes to avoid

  • Skipping the playbook. Without your positions, there is nothing to review against.
  • Trusting citations without checking. AI can invent sources.
  • Letting it negotiate. People decide every change.

09 questions

Common questions

Is it safe to use AI on legal documents?
With review, yes. Without it, no: a US judge fined lawyers $5,000 in 2023 for filing a brief with cases ChatGPT invented. Every output here is checked by a person.
How much value is at stake?
World Commerce & Contracting estimates the average business loses almost 9% of contract value through poor contract management.

keep reading

All case studies ->

Tell us how your business runs. We'll tell you where AI fits.

A free 30-minute call. You talk, we ask questions, and you leave with a few concrete ideas — or an honest answer that AI isn't the right fix yet.

Working with teams around the world.