Pressure-Test

Pressure-test a contract before you sign it

Reads the agreement as the counterparty's litigator two years into a dispute: finds the discretion, the silences, and the unenforceable standards they will exploit, walks three exits, and gives paste-ready fixes ranked by what they protect.

About 20 minadvancedTransactional, In-house

Your prompt5,319 characters

Still to fill in: Contract text, Who you are in this deal, Governing law and forum

RoleYou are the counterparty's litigator two years from now, not their deal lawyer today. You have the signed agreement, a client who wants out, and every incentive to find the reading that helps. You hunt for the three places a contract hands someone permission (discretion, silence, and undefined standards) and you never confuse a term being market with a term being safe.What I needRead the agreement below against Who you are in this deal and tell me what Counterparty's lawyers do with it in a dispute under Governing law and forum. Weigh everything against the real stakes: Deal context.InputsContract: Contract text My side: Who you are in this deal Counterparty: Counterparty Governing law and forum: Governing law and forum Deal context: Deal contextHow to work this1. Inventory every place Counterparty holds unilateral discretion: "sole discretion," "as it may determine," "reasonably acceptable to," approval rights, and any right to change terms, pricing, or policies after signature. Quote the words and the section. 2. Inventory the silences: assignment, change of control, data return and format, key personnel, insurance, survival, audit, prepaid amounts on termination. Treat each silence as a right they will claim. 3. Inventory the standards that could not be enforced ("commercially reasonable," "material," "promptly," "best efforts") and for each say what Who you are in this deal would have to prove to win on it. 4. Walk three exits as timelines: they terminate for convenience; they allege our material breach; a change of control puts them inside a competitor. Say what Who you are in this deal has on day 1, day 30, and day 90. 5. Trace one realistic claim through cap, exclusions, carve-outs, indemnity, insurance, and forum. Say what is actually collectible at the end, in dollars. 6. Read notice and cure as mechanics they will use: who must notify, in what form, to what address, within how many days, and what they may do with no notice at all. 7. Give every hole a pre-signature fix as language I could paste, marked must-fix (I do not sign without it), trade, or nice-to-have. 8. Where Governing law and forum changes the answer: enforceability of the cap, liquidated damages, a consequential-damages waiver, an indemnity for one's own negligence. Say so rather than assuming.Ask me firstYou have the contract, my side, and the governing law above. Ask me these three questions (the ones the paper cannot answer), then stop and wait: 1. What is the realistic worst case in dollars if this counterparty fails, underperforms, or ends up owned by someone you would never have signed with? Every clause gets measured against that number, not against market practice. 2. If they stopped performing on a Tuesday, what breaks and how many weeks until a replacement is running? Switching cost is what makes their discretion expensive. 3. What has already been conceded or declared non-negotiable, so I do not spend your leverage on ground that is gone? Do not begin until I answer. If I tell you to proceed anyway, state each assumption at the top of your output and mark it [ASSUMPTION - verify].Output formatWhere they have permission: discretion inventory, quoted, by section. What the contract does not say: each silence and what it lets them do. Standards you could not enforce. Three exits walked: day 1, day 30, day 90. The remedy stack, traced: one claim from breach to dollars collected. Pre-signature fixes: paste-ready language, each labeled must-fix / trade / nice-to-have. End with one line naming the two of my answers that decided which fixes are must-fix, and what you would have ranked first without them. If an answer changed nothing, say so. It means I should not have been asked for it.Never do this- If this would fit any vendor contract anywhere, it is too generic. Anchor every finding to this agreement's sections and this deal's numbers. - No hedging filler. Cut "arguably," "it should be noted," "this is fairly standard," and "it depends." Do not tell me to consult an attorney. I am counsel. - Every statute, case, or rule you invoke must come from my inputs or carry [UNVERIFIED - confirm under the governing law]. Never invent a citation or quotation, and never quote contract language that is not in what I pasted. - Where you do not know how Governing law and forum treats a cap, a waiver, or an indemnity, say you do not know and name the question. Do not smooth over the gap with fluent prose. - Do not confuse unusual with dangerous. An odd clause with no consequence is a footnote; an ordinary cap below real exposure is the headline. - Do not pad. If three holes matter, name three. Length is not value.Before you answer- Did I quote the operative words and name the section for every finding? - Did I test the cap against the actual exposure number rather than in the abstract? - Does every fix come as language that could be pasted into the document tonight? - Did I walk all three exits on a timeline, and would this analysis be useless applied to a different contract? It should 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.

Their paper, their deal desk, their quarter to close. Take the negotiation call as the counterparty's general counsel, who has a standard answer to every ask on this list. Go through the pre-signature fixes and sort them: refuse outright, accept without thought, or accept for a price. For each refusal, write the sentence you would actually say on the call, then give me the fallback language you would sign instead. Finish by naming the one ask on the list you would quietly have given up if pushed twice. That is where the leverage should go.

Go deeper

Pushes the work further once the basics are right.

You have the issue list; what you do not have is the order to spend leverage in. Write the negotiation plan: the two must-fix items with their exact language and the business reason you would give for each rather than the legal one, the order to raise them, what you are prepared to trade for each, the fallback position if the first ask is refused, and the line at which you recommend walking away.

Before you run it

What to gather first

The full agreement, including exhibits, order forms, and any incorporated policies
The realistic worst-case dollar exposure if this counterparty fails or exits
How long it would take to replace them and what breaks in the meantime
What has already been conceded or declared non-negotiable
Who on your side is willing to have a fight over a term

Watch for

The model reads only what you paste. Order forms, exhibits, and policies incorporated by URL usually carry the discretion and the real obligations. Omit them and the analysis is wrong.
Enforceability of caps, consequential-damages waivers, liquidated damages, and indemnities for one's own negligence varies sharply by governing law. Confirm the rule in the actual jurisdiction before relying on any conclusion here.
Cross-references defeat the model. It will misread a carve-out that points to a section that points to an exhibit. Verify every chain by hand before you cut or accept a term.
An adversarial read produces a long list. Signing is a business decision, and a review that flags everything protects nothing. Insist on the must-fix ranking.
Do not paste client-confidential agreements unless your firm's AI policy and the engagement terms permit it.

What comes back

A discretion inventory quoted by section; the contract's silences and what each one lets the counterparty do; the standards you could not enforce and what proving each would take; three exit scenarios walked at day 1, day 30, and day 90; one claim traced through cap, carve-outs, indemnity, insurance, and forum to dollars actually collectible; and pre-signature fixes with paste-ready language, each labeled must-fix, trade, or nice-to-have.

See an example of what you’ll get

*(After you answer the three questions: worst case is roughly $6M in claims-handling disruption and a regulatory finding, four to six months to migrate off, and pricing is already settled.)* Where they have permission. - § 3.4: "Vendor may modify the Service at any time in its sole discretion, provided such modification does not materially degrade core functionality." They define both "materially" and "core functionality." This is a license to change the product you built your intake workflow around. - § 5.2: Vendor may update the Acceptable Use Policy "from time to time" by posting. Posted terms bind you on notice you will not receive. - § 9.1: Suspension for suspected breach "immediately and without notice." With a four-to-six-month migration path, suspension is functionally termination. What the contract does not say. - Nothing about change of control. § 14 lets Vendor assign to a successor without consent; you may not assign at all. Given a PE-backed Series D vendor in a consolidating space, you may wake up hosting your claims data with a competitor and have no exit. - Nothing about the format or timing of data return. § 11 says data "will be made available" for 30 days. Available how: a paginated API at 500 records per call? That is not a migration path for four years of claims. - No insurance requirement anywhere, for a vendor holding regulated policyholder data. - No survival clause. As drafted, confidentiality and the indemnity end at termination, and their counsel will read it exactly that way. Standards you could not enforce. "Commercially reasonable efforts" on uptime (§ 6.1) with no metric: to win you would have to prove industry practice through an expert, in an arbitration, against a vendor who will testify their practice is the industry. "Promptly" for security-incident notice (§ 7.3): meaningless against NYDFS's 72-hour clock, which is yours to satisfy, not theirs. Three exits walked. - *They terminate for convenience (§ 14.3, 60 days).* Day 1: notice arrives. Day 30: you are still three to five months from a replacement being live and have no contractual right to transition assistance. Day 90: contract is 30 days dead, data available only through the § 11 window, adjusters working manually. - *They allege your material breach (§ 9.1).* Day 1: suspension without notice; intake stops. Day 30: you are in a cure dispute over an undefined term while your claims backlog is a regulatory problem. Day 90: you are arguing about the cap. - *Change of control.* Day 1: acquisition announced. Day 30: no consent right, no termination right, no data-segregation commitment. Day 90: your claims data sits with a competitor's affiliate and your only lever is not renewing in year three. The remedy stack, traced. A 10-day outage causing $2.1M in disruption: § 12.2 caps liability at three months of fees, or $350,000. § 12.3 excludes consequential and indirect damages, and business disruption is where they will put your $2.1M, leaving direct damages of about $180,000 in duplicated staffing. Carve-outs at § 12.4 cover confidentiality breach and indemnity, but not gross negligence or willful misconduct, so even a deliberate shutdown is capped. Indemnity at § 10 is third-party only. No insurance requirement to reach behind the cap. Realistic recovery: $180,000 against $2.1M, after an AAA arbitration that costs you six figures. [UNVERIFIED - whether New York enforces a cap this asymmetric against a regulated entity, and whether the gross-negligence omission survives, needs confirming under New York law.] Pre-signature fixes. 1. Must-fix: § 12.4 carve-outs. Add: "The limitations in Sections 12.2 and 12.3 shall not apply to Vendor's gross negligence or willful misconduct, Vendor's breach of Section 7 (Security), or Vendor's indemnification obligations under Section 10." This is the single change that moves recovery from $180,000 to something real. 2. Must-fix: § 11 data return. Add: "Upon termination for any reason, Vendor shall provide a complete export of Customer Data in a documented, machine-readable format within 15 days, and shall provide transition assistance at then-current rates for up to 120 days upon Customer's request." Your migration timeline requires it. 3. Trade: § 14 change of control. Add a customer termination right on assignment to a competitor, with 90 days' transition assistance. Worth giving up the audit right you asked for. 4. Trade: § 6.1 uptime. Replace "commercially reasonable efforts" with 99.9% monthly uptime and a service-credit table. 5. Nice-to-have: § 7.3 notice. Replace "promptly" with "within 48 hours of discovery", a full day inside the 72-hour NYDFS clock that is yours to satisfy, not theirs. What your answers changed. The $6M figure is what made § 12.4 a must-fix rather than a trade. Drop the worst case to $250,000 and a capped, consequential-excluded recovery is roughly proportionate to the harm. The carve-outs become something I would trade for the uptime metric, and the § 11 data-return language would have ranked first instead. The four-to-six-month migration is what made § 11 the second must-fix and what sets the 120-day transition-assistance ask; at a two-week switching cost, § 9.1 suspension without notice is an inconvenience rather than a functional termination, the three exits get walked in days instead of quarters, and data return drops to a nice-to-have. Your "pricing is settled" answer changed nothing here: not one of the five fixes touches price, so the ranking is what it would have been if I had never asked.

Why this prompt is built the way it is

## Framework 1. **Inventory the discretion.** Every "sole discretion," "as it may determine," "reasonably acceptable to," approval right, and post-signature change right, quoted with its section. 2. **Inventory the silences.** Assignment, change of control, data return, key-person or continuity, insurance, survival, audit, effect of termination on prepayments. A silence is a right the other side will claim. 3. **Inventory the unenforceable standards.** "Commercially reasonable," "material," "promptly," "best efforts", and what you would have to prove to enforce each. 4. **Walk three exits on a timeline.** Termination for convenience, alleged material breach, change of control. Day 1, day 30, day 90. 5. **Trace one claim through the remedy stack.** Cap, exclusions, carve-outs, indemnity, insurance, forum: from breach to what actually gets collected. 6. **Read notice and cure as a weapon.** Who notifies, in what form, to what address, in how many days, and what they can do with no notice at all. 7. **Every hole gets paste-ready language**, marked must-fix, trade, or nice-to-have. 8. **Flag where governing law changes the answer** rather than assuming it does not.