How reviews work
A review that reads like
your best lawyer.
A good contract review isn’t the one that flags the most. It’s the one that flags the right things, grounds every point in the document, and tells you what to do next. Here’s how Clment gets there.
The setup
Four decisions, then it runs.
Launching a review is four quick choices. Most of the time only the last one changes from run to run.
Documents
The contract, plus any attachments — each marked Review or Supporting — and ad-hoc context.
Coverage
The whole document, or just what changed: tracked changes, or a comparison with another version.
Playbook
The standard to judge against — your house positions in plain English, plus optional one-off instructions.
Approach
The review strategy — how assertive the review is and how much lands in your redline.
Review strategy
Tell it where you are with the deal.
A first read of a contract you’ve never seen needs a different review from a final check on a deal you’ve already negotiated twice. The strategy is one plain choice — where are you with this contract? — and it tunes both how assertive the review is and how much lands in your redline by default.
- Four strategies, from a wide first look to a complete compliance record
- Sets a severity floor for what’s pre-included in the redline
- Changes the posture, never what gets checked — that’s the playbook
Explore
First look — understand the contract
Negotiation
You're going to negotiate
High-priority only
You're close to signing
Strict
Every deviation on record
Redline floor applies only to findings you haven’t decided yourself.
Liability cap is set to fees paid in the prior 12 months.
“…the total liability of either party shall not exceed the fees paid in the twelve (12) months preceding the claim.”
Recommendation: widen to 24 months for a deal of this value. A realistic counterparty position; not a dealbreaker.
The posture
A seasoned GC, not an
over-eager junior.
The fair worry about AI review is that it’s overzealous — inventing problems, citing authority it can’t back up, and burying the real issues in noise. Clment reviews are built to behave the opposite way, on every strategy.
- Stay grounded
- Findings cite only what the contract or your playbook actually supports. The AI won’t invoke a statute or a case it can’t tie to the document — if it can’t ground a point, it flags it for human verification instead of asserting it as fact.
- Hold a reasoned line
- It won’t inflate severity because you asked it to “be tough,” and it won’t drop a real issue because the other side pushed back without substance. It changes its assessment only on the merits.
- Be measured, not maximal
- It prefers the least-aggressive position that protects you, and reserves “critical” for genuine dealbreakers — so the findings that matter aren’t buried under trivia.
Your call
You decide. It drafts the redline.
The AI proposes; you dispose. Each finding has four independent controls, so your judgment — not the model’s confidence — decides what reaches the counterparty.
- Verdict — agree, disagree, or partial on the recommendation
- Comment — your reasoning, kept on the record
- Include in redline — what actually gets drafted
- Redline instruction — your wording in place of the AI’s
When you’re done, Clment generates a Word document with real tracked changes — attributed, footnoted with the review reference, ready to send. Switch strategy and the findings you ignored fall in or out automatically; the ones you decided stay exactly as you left them.
…shall not exceed the fees paid in the
twelve (12) twenty-four (24)
months preceding the claim.
Run your first review free
Upload a contract, pick a strategy, and see the difference a grounded review makes. No credit card.