Prompt of the day

Clause-by-clause contract review against YOUR playbook — deviation table + redlines

Opposing counsel sent a draft agreement. Run a first-pass review against your standard positions — deviations, missing protections, and replacement language — before partner time touches it.

You are a transactional attorney performing a first-pass ISSUE-SPOTTING review for a licensed lawyer — a drafting aid, not legal advice, and not a substitute for attorney review. I will paste an ANONYMIZED contract draft (party names replaced with [PARTY A] / [PARTY B], pricing and deal identifiers removed) plus my standard positions.

Produce:

A) DEVIATION TABLE — clause by clause: clause reference, what the draft says (quote the exact language), my standard position, deviation severity (deal-breaker / negotiate / acceptable), and which party the current language favors.

B) MISSING-CLAUSE CHECK — standard protections absent from this draft for this contract type: indemnification caps, limitation of liability, termination rights, IP assignment, confidentiality, dispute resolution, force majeure. List only what is genuinely absent.

C) REDLINES — for each deal-breaker or negotiate item: proposed replacement language in plain contract prose, plus a one-line fallback position.

D) NEGOTIATION SUMMARY — the 3 points worth spending negotiation capital on, and why.

Inputs: [PASTE ANONYMIZED CONTRACT] · [CONTRACT TYPE + GOVERNING LAW] · [MY SIDE + STANDARD POSITIONS OR RISK TOLERANCE]

Rules: Anonymize first — strip party names, signature blocks, pricing, and confidential deal terms before pasting into a consumer LLM, and keep privileged negotiation notes out entirely (ABA Formal Op. 512 confidentiality duties apply). Quote only language that actually appears in the draft — do not invent clauses, defined terms, or "market standard" claims you cannot ground in my inputs. Verify every proposed redline against the agreement's defined terms and cross-references before anything goes back to the other side.

Make AI interview you before it answers

90% of bad AI output comes from a thin prompt. Flipping it — the model gathers context first — turns a generic draft into something that sounds like you, on the first try.

Before asking AI to write or plan anything, tell it: 'Ask me 5 questions that will make your answer 10x better before you start.' Answer them, then let it produce the work.

1. Paste your rough goal ('write a launch email', 'plan this project').
2. Add: 'Ask me 5 clarifying questions first, then wait for my answers.'
3. Answer briefly, then say 'now produce it.'