Skip to content
CloseWeave

Legal & risk

Contracts reviewed in hours, with lawyers on the judgement calls.

Legal teams are asked to approve AI and to use it. We build both: review and drafting that follow your playbook, and the rules, records and controls the rest of the business needs.

At a glance

A customer strikes out the liability cap in a services agreement.

Deployed AI

Two to three days in the queue

Lawyers review every clause

Operational AI

Hours, not days

AI handles standard terms; lawyers take the judgement calls

See all three ways

Illustration

What matters

What legal and risk teams are judged on.

  • Turnaround

    Contracts back to the business in hours, not days.

  • Consistent positions

    Every reviewer applies the same playbook.

  • Visible risk

    Obligations, renewals and exposures tracked, not buried.

  • Safe AI across the business

    Clear rules and records for every AI system in use.

Walkthrough

The same request, three ways.

Switch between deployed, specialised and operational AI to see who does the work, how long it takes and what keeps it safe.
closeweave.com/ai/legal-and-risk/walkthroughIllustration

The request

ContractSales, deal desk

Customer redline: master services agreement

The customer removed the cap on liability for data breaches.

Annual value £180,000; signature wanted on Friday.

Show it handled by

Autonomy level 1

Deployed AI

A chat assistant. Someone asks, it answers, people do the rest.

  1. A person types

    Is this limitation of liability clause acceptable? [clause pasted]

  2. AI replies

    The clause caps liability at fees paid in the last 12 months but excludes data breaches from the cap. This may expose you to significant risk. You may want to consult a lawyer.

  3. Person

    Lawyer. Compares it with the playbook, drafts fallback wording and emails sales.

3 steps · one to a person · two to three days in the queue

How long
Two to three days in the queue
Who does the work
Lawyers review every clause
What keeps it safe
None: no link to your playbook
What goes to people
Everything

Autonomy level 2

Specialised AI

It works from your data and tools and prepares the work. A person approves.

  1. Trigger

    The redline arrives in the legal intake queue.

  2. Tools
    playbook.positions("limitation of liability") · contracts.compare("MSA v3", redline)

    Returned: Standard: a cap at 12 months' fees. Fallback: data breaches capped at three times annual fees. Walk away: unlimited liability.

  3. AI replies

    Three changes. Two are within the playbook. Uncapped data breach liability is a walk-away position, so the fallback is drafted with a plain-English note to sales.

  4. Person

    Lawyer. Reviews the redline and approves it.

    Approved
  5. Action

    Redline returned to sales.

5 steps · one to a person · about 30 minutes of a lawyer's time

How long
About 30 minutes of a lawyer's time
Who does the work
AI reviews against the playbook; lawyers approve
What keeps it safe
Every position traced to the playbook
What goes to people
Every contract

Autonomy level 3

Operational AI

It runs the process within set limits, logs every step and hands people the exceptions.

  1. Trigger

    Every contract and every amendment.

  2. Rule

    Accept automatically only terms inside the playbook, and only below the agreed deal value.

    The legal floor always stands: liability for death or personal injury caused by negligence, and for fraud, cannot be excluded.

  3. Action

    Low-risk contracts approved the same day; obligations and renewal dates logged.

  4. Person

    Exception: goes to a person

    Lawyer. This redline is outside the playbook and above the deal value, so a lawyer sends it back with the fallback.

    Sent back
  5. Record

    [16:30]Every position, approval and version kept.

  6. Report

    Monthly: turnaround, the positions most often negotiated, and obligations coming due.

6 steps · one to a person · hours, not days

How long
Hours, not days
Who does the work
AI handles standard terms; lawyers take the judgement calls
What keeps it safe
Playbook limits, a deal value limit, an approval log
What goes to people
Anything outside the playbook or above the value

Controls

Owner
General counsel
Measure
Turnaround and share within the playbook
Limit
Playbook terms under the agreed value
Log
Every position and version
Fallback
Contracts queue for lawyers
Stop
Send everything to lawyers in one switch

Illustrative: names, figures and timings are invented.

Illustration: A customer strikes out the liability cap in a services agreement. Shown three ways: Deployed AI, Specialised AI, Operational AI.

Where to start

Processes that pay back first.

Each at the autonomy level we recommend. Starting points, not case studies: we confirm each one with your numbers.
  • 2Automate with approval

    Contract review against your playbook

    Lawyers approve any departure.

  • 3Run with exceptions

    NDAs and standard terms

    Approved within the playbook.

  • 3Run with exceptions

    Obligation and renewal tracking

    Alerts before dates pass.

  • 2Automate with approval

    Regulatory and policy monitoring

    Changes summarised for review.

  • 2Automate with approval

    AI register and risk assessments

    For every AI system in the business.

  • 1Assist

    Disputes and new advice

    Lawyers lead; AI researches.

What we build and run

Built in your systems, run with your team.

  • Contract review that follows your playbook
  • Obligation and renewal tracking
  • An AI register, policies and risk assessments for the business
  • Controls designed for UK GDPR, and the EU AI Act where it applies

What we measure

Numbers you already track.

  • Contract turnaround
  • Share approved within the playbook
  • Obligations missed
  • AI systems registered and assessed
  • Lawyer hours on routine work

The Academy

Learn to run it yourselves: the Legal & risk track.

Review and draft with AI, and set the rules for everyone else.

Being built

Start with one legal and risk process.

We map the work, pick the process that pays back first, and run it with your team.