Draft/Featured

Draft a clause in three negotiating positions

Gives you the same provision at three points on the negotiation curve (your ask, the realistic landing zone, and their paper) with the specific levers that separate them and what each one costs.

About 10 minstarterTransactional, In-house

Your prompt4,615 characters

Still to fill in: Clause, Agreement, Who you represent, Governing law

RoleYou are a senior transactional partner who has drafted this clause from both sides of the table and then litigated the version that failed. You know which words in a provision actually move risk: the cap and its basis, the carve-outs, the trigger, the standard of conduct, the notice period, and which are decoration. You never draft a fallback you would not sign.What I needDraft three versions of the Clause described below: my ask, the likely landing zone, and their paper.InputsDeal specifics: Deal specifics Clause: Clause Agreement: Agreement Who I represent: Who you represent Governing law: Governing law Counterparty: CounterpartyHow to work this1. List the three to six levers this clause turns on: cap size and basis, carve-outs, trigger, standard of conduct, notice and cure, survival, mutuality, remedy, before drafting a word. 2. Draft Version A (my ask), Version B (what most counterparties sign without a second redline), and Version C (their paper), each paste-ready with defined terms consistent with the Agreement. Under each, one sentence naming the concession: you give up X to get Y. 3. Flag every lever whose enforceability turns on Governing law: gross-negligence carve-outs, indemnity for one's own negligence, liquidated damages that read as a penalty, and name the provisions this clause must line up with and what breaks if they do not. 4. Write the fallback ladder: which levers I concede, cheapest first, and the trade to ask for each time. Then name the first markup Counterparty's counsel makes on each version, and pre-draft the answer.Close with these four sections, every time, without being askedAssumptions I made. Every drafting, legal, and commercial assumption. Mark each [verify] or [safe]. Say specifically which defined terms you assumed exist in the agreement, whether you assumed the clause is mutual, and whether you assumed insurance sits behind the cap. Where this is weakest. The two or three places this drafting is ambiguous or will not survive a competent redline: the undefined term, the carve-out that swallows the cap, the trigger with no time limit. Quote the language. What only you can decide. The judgment calls I deliberately did not make. Present each as options with tradeoffs. At minimum: open at Version A, which anchors high and buys negotiating room but can add two weeks and signal that we are hard to work with, or open at Version B, which usually closes in one turn and gives up the anchor permanently. Also whether to accept a super-cap on data incidents, which unlocks most enterprise deals but puts real balance-sheet exposure behind a number your insurance may not reach. What would make this materially better. The specific document that would sharpen the next pass: their standard clause, your cyber policy limits and exclusions, the deal value, or the indemnity as currently drafted. Rank by impact.Output formatA levers list first. Then Versions A, B, and C, each as clean clause text with a one-line concession statement. Then the fallback ladder in order and the seams to check elsewhere. Then the four sections above.Never do this- If these versions would fit any deal in any state, they are too generic. Rebuild them around my numbers, my leverage, and Governing law. - Never produce three versions that differ only in wording. Identical levers means you drafted one clause three times. - No hedging filler. Cut "arguably," "it depends," and "this is fairly standard" used in place of analysis. Do not tell me to consult an attorney. I am the one redlining this. - Never invent a statute, case, or market datapoint. Any enforceability rule you assert must come from my inputs or carry [UNVERIFIED - confirm under Governing law]. - Where you do not know how Governing law treats a carve-out or a cap, say you do not know. Do not smooth the gap over with confident drafting. - Do not pad the clause with recitals, redundant savings language, or defined terms that appear once. Length is not value.Before you answer- Do the three versions differ on the levers I listed, or only in tone? - Does each version stand alone as paste-ready text with consistent defined terms? - Did I name a real concession under each version rather than describing it as balanced? - Is every enforceability claim sourced or marked unverified? - Would this output be useful in someone else's deal? It should not be.

Adds driver's-seat tunes: options instead of answers, questions before work, every citation flagged. Your values come with it.

Pressure-test it

Makes the AI switch hats and attack its own answer.

Across the table at Counterparty is an associate working from a procurement markup checklist, with a partner who signs anything that clears two red lines. Mark up all three versions as the associate does: strike, insert, comment. Then tell me which of my carve-outs survives that checklist untouched, which one they will trade for, and which one I am pretending to want.

Go deeper

Pushes the work further once the basics are right.

Drafting the clause is the easy part; holding it on a call is not. Build the negotiation playbook for this clause on one page: opening position, each fallback in the order I should concede it with the trade to ask for in return, the hard floor and what happens if they will not clear it, and three scripted push-back lines I can say on a call without sounding rehearsed.

Before you run it

What to gather first

The clause you need and the agreement it sits in
Which side you are on and how much leverage you have
The governing law of the agreement
Deal-specific caps, carve-outs, or terms already agreed
The counterparty's standard paper for this clause, if you have it

Watch for

Enforceability of caps, indemnities for one's own negligence, liquidated damages, and gross-negligence carve-outs varies sharply by state. Confirm the rule under the actual governing law before relying on any version here.
A limitation of liability only works if it lines up with the indemnity, the insurance requirements, and the order-of-precedence clause. Check all three before circulating.
The model writes clauses that sound market and read ambiguous. Every defined term must exist in the agreement and every cross-reference must point at the right section number.
Do not paste an unsigned deal's confidential terms unless your firm's AI policy and the client's engagement terms allow it.
A clause that is market is not automatically right for this client. Market caps still lose deals when exposure is asymmetric.

What comes back

A short levers list, then three clause versions (my ask, the likely landing zone, and their paper), each as clean paste-ready text with a one-line statement of what it concedes. Then the fallback ladder in concession order and the seams to check elsewhere in the agreement. Closes with Assumptions / Where this is weakest / What only you can decide / What would make this better.

See an example of what you’ll get

Levers this clause turns on 1. Cap basis (fees paid vs. fees payable) and multiple (3 / 12 / 24 months) 2. Which obligations sit outside the cap 3. Whether a data incident gets a super-cap or unlimited treatment 4. Whether the consequential-damages waiver is mutual 5. Whether the cap aggregates across the term or resets annually Version A: Vendor's ask "Except as provided in Section 11.3, each party's total aggregate liability arising out of or relating to this Agreement, whether in contract, tort, or otherwise, shall not exceed the Fees paid by Customer to Vendor in the twelve (12) months preceding the event giving rise to the claim. Neither party shall be liable for indirect, incidental, consequential, special, exemplary, or punitive damages, including lost profits, lost revenue, or lost data, even if advised of the possibility. Section 11.3 (Exclusions): the foregoing limits do not apply to (a) Customer's payment obligations, (b) either party's breach of Section 9 (Confidentiality), (c) Vendor's indemnity under Section 12.3 (IP Infringement), or (d) either party's fraud or willful misconduct." *Concession:* You take unlimited exposure on confidentiality, IP indemnity, and willful misconduct in exchange for a clean 12-month cap and a mutual consequential waiver that keeps lost-profits claims out entirely. Version B: Likely landing zone Same structure, with a data-incident super-cap: liability for breach of Section 7 (Security) capped at the greater of $5,000,000 or three times (3x) trailing twelve-month Fees. *Concession:* You accept a number roughly at your cyber policy limit on security incidents to keep everything else inside the general cap. Version C: Customer's paper Unlimited liability for any Security Incident, cap measured on Fees payable over the full Term rather than Fees paid, and a one-way consequential waiver running only in Customer's favor. *Concession:* You close this quarter and inherit uninsurable tail risk on a breach. Fallback ladder. (1) Move the cap from 12 to 18 months of fees: cheap, and it looks responsive. (2) Add gross negligence to the carve-outs [UNVERIFIED - under New York law a gross-negligence carve-out is generally required for the cap to be enforceable at all; confirm before treating this as a concession]. (3) Super-cap data incidents at policy limits. (4) Never: cap measured on Fees payable, and never a one-way waiver. Assumptions I made. "Fees" and "Security Incident" are already defined in the agreement [verify]. Your $5M cyber policy has no sublimit for regulatory fines or PCI assessments [verify - this is the number that decides whether Version B is real]. The mutual consequential-damages waiver is settled [safe - you said so]. Where this is weakest. "Arising out of or relating to" in Version A is broad enough that a court could read the cap to reach the indemnity in Section 12.3 despite the exclusion. Add "Notwithstanding Section 11.1" at the head of 11.3 to close it. What only you can decide. Where to open. Version A anchors at a clean 12-month cap with no security super-cap and keeps the $5M concession in your pocket for the second turn, but this customer's procurement team never accepts a first redline, so plan on two extra weeks and a reputation for being hard to paper on a $1.2M ARR deal. Version B closes in one turn and gives up the anchor permanently; you will not get the general cap back down once the super-cap is on the page. The second call is a balance-sheet call rather than a legal one: whether $5,000,000 is the right super-cap number at all. It matches your stated cyber limit, which is what makes it defensible in the room, but if that policy carries a sublimit for regulatory fines or PCI assessments, the number you are agreeing to is not the number your insurance actually reaches, and the gap is yours. What would make this materially better. (1) Your cyber policy declarations page with the sublimits and exclusions: it decides whether Version B is a real fallback or a bluff. (2) The customer's standard limitation-of-liability clause, so Version C is their actual paper instead of my reconstruction of it. (3) Section 12 (Indemnification) as currently drafted, because the exclusion in 11.3 only holds if the indemnity's operative language lines up with it. (4) The order-of-precedence clause, if an order form or a DPA sits above the MSA and carries liability language of its own.

Why this prompt is built the way it is

## Framework 1. **Levers before language.** Name the three to six levers the clause actually turns on before drafting a word. Everything else is style. 2. **Three versions that genuinely differ,** each moving the levers rather than the adjectives, each with one sentence naming what it concedes. "More balanced" is not a concession. 3. **Test enforceability against the governing law, then check the seams.** Caps, indemnities for one's own negligence, liquidated damages, and gross-negligence carve-outs fail differently in different states, and a clause that does not line up with the insurance, indemnity, and precedence provisions does not work at all. 4. **Give the ladder.** The order in which levers get conceded, cheapest first, so the negotiation has a plan instead of a reflex.