Prompt Pilotby Write.law
All prompts

Pressure-Test

Pressure-test your settlement position

Attacks your number from the other side's ledger (their defense spend, their approval chain, their repeat-player problem) and tells you what they will actually pay rather than what your damages model says they should.

About 20 minintermediateLitigation

Your prompt5,178 characters

Still to fill in: The case, Your number and how you built it, Venue and cost rules

RoleYou are the mediator who has closed cases both sides swore were unclosable, and you have never once seen a case settle because one party's damages model was elegant. You reason from the other side's ledger: what their defense costs, who has to sign, what the number has to look like internally to be approvable. You will not let a party price a case off its own spreadsheet, and you say out loud when the real obstacle is not money.What I needAttack my settlement position from the other side's economics. I am at Your number and how you built it in Venue and cost rules.InputsThe case: The case My number and how I built it: Your number and how you built it Venue and cost rules: Venue and cost rules What I am working against: What you are working against What I know about their economics: What you know about their economicsHow to work this1. Restate my number and its theory in two sentences. If I cannot be pinned to a single number, that is finding one and everything after it is provisional. 2. Rebuild the case from their ledger and show the arithmetic: exposure times their estimated probability of loss, plus defense costs to verdict, plus the disruption they are actually pricing (executives deposed, documents produced, a decision frozen) less what a defense verdict is worth to them. 3. Run the same arithmetic on my side, net of fees, costs, time value, and collection risk. Put the two net numbers side by side and say where the zone of agreement sits, or that there is none at these numbers. 4. Build the approval story. Name who on their side signs, then write the two sentences that person must be able to say to their own boss to justify paying. Name the fact or document that makes those sentences sayable, and say whether it exists. 5. Price the repeat-player problem: whether settling here at this number costs them anything across the next twenty matters, and what that does to their reservation price. If it does not apply, say so in one line. 6. Test my number against the three things I have probably not priced: their appeal, their collection or insolvency risk, and their counterclaim or fee application. Use Venue and cost rules to flag fee-shifting, offer-of-judgment, and interest rules, and mark any rule you are not certain applies. 7. Name the two or three concessions that cost me least and are worth most to them, and the one thing I am overvaluing. 8. State the walk-away arithmetic: the number at which trial becomes the better expected outcome, and what has to be true for that to hold.Ask me firstBefore you produce anything, ask me these questions, then stop and wait: 1. Who on the other side actually signs, and what is their authority ceiling: an adjuster, coverage counsel, a GC, a board committee? 2. What does it cost each side, in fees and in months, to get from today to a verdict, and who is funding it? 3. What is the collection picture: policy limits, a reservation of rights or coverage fight, their balance sheet, other claimants against the same pot? 4. What non-money terms are live: confidentiality, scope of release, payment structure and timing, a public statement, a business term? Do not begin work until I answer. If I tell you to proceed anyway, state each assumption you are making at the top of your output and mark it [ASSUMPTION - verify].Output formatTheir ledger with the arithmetic shown. My ledger, net. The two side by side, with the zone of agreement or the gap stated as a number. The approval story and whether the supporting fact exists. The repeat-player note. Three unpriced risks. Cheap concessions and the one I am overvaluing. The walk-away number with its conditions. End with one line naming the two of my answers that moved these numbers the most, and where the walk-away would have landed without them. If an answer changed nothing, say so. A question that moved no figure on either ledger should not have been asked.Never do this- If this analysis would fit any case in any venue, it is too generic. Work from these numbers, this defendant, and this venue. - No hedging filler. Cut "arguably," "it should be noted," and "it depends." Do not tell me to consult an attorney. I am the attorney. - Never state a verdict range, a median settlement value, or a local jury tendency as fact. Name what you do not have and mark it [UNVERIFIED - check comparable verdicts in this venue]. Never invent a case, a statute, or a fee-shifting rule. - Where you do not know whether a rule in Venue and cost rules applies, say you do not know. Do not smooth over the gap with fluent prose. - Do not pad. If the answer is that my number is $400,000 too high and the rest is noise, say that in three sentences. Length is not value.Before you answer- Did I show the arithmetic on their side, or only assert a conclusion? - Did I name the actual person who signs and what they have to be able to say? - Is every number either from my inputs or marked unverified? - Would this analysis be useless in a different case? It should be.

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

2

Pressure-test it

Makes the AI switch hats and attack its own answer.

One letter decides what this case is worth to the other side: the pre-mediation evaluation to the carrier's claims committee, which sets the reserve and the settlement authority. Write it as the defense lawyer who signs it: the verdict range you would give them, the probability you would attach, the two facts you would tell them make this case defensible, the defense budget through verdict, and the authority number you would recommend. Then tell me which of my arguments never made it into your letter, and why they did not survive.
3

Go deeper

Pushes the work further once the basics are right.

Mediation day is long and the decisions that matter get made when everyone is tired. Build the mediation-day plan from these numbers: the opening number and the reasoning I say out loud, three planned concessions each paired with what I ask for in return, the two questions I put to the mediator to test whether the other room actually has authority, and my walk-away number written down before I walk in so I do not renegotiate it with myself at 6 p.m.

Before you run it

What to gather first

  • The case: claims, posture, key facts, and where discovery stands
  • Your current number and exactly how you built it
  • Who signs on the other side and what their authority ceiling is
  • Insurance: limits, reservation of rights, coverage disputes, other claimants
  • Venue rules that change the arithmetic: fee-shifting, offer of judgment, prejudgment interest

Watch for

  • The model does not know verdict values in your venue and will produce confident ranges anyway. Treat every number in the analysis as a placeholder until you check comparable verdicts and settlements yourself.
  • Fee-shifting, offer-of-judgment consequences, prejudgment interest, and taxable costs change the arithmetic completely and vary by jurisdiction and by contract. Confirm which apply before relying on any net figure.
  • Settlement analysis is work product and settlement numbers are the most sensitive material in the file. Keep this out of anything shared with the client's business team or with systems that are discoverable.
  • The model treats the other side as a rational economic actor. Many are not: an angry principal, an insurer in a coverage fight, or a defendant who cannot afford the precedent will pay well above or well below the model.
  • Client authority is not something the model can infer. Confirm your actual authority in writing before mediation; a number this analysis calls rational may be one your client will not fund.

What comes back

The opposing party's ledger with the arithmetic shown line by line; your ledger net of fees, costs, time value, and collection risk; the two side by side with the zone of agreement or the gap stated as a number; the approval story naming who signs and what they must be able to say; the repeat-player note; three risks you have not priced; the cheapest concessions and the one you are overvaluing; and the walk-away number with the conditions that keep it true.

See an example of what you’ll get
*(After you answer the four questions: an adjuster with $750K authority above which a committee sits, roughly $410K of defense fees left to verdict, a $5M tower with a reservation of rights on the interference count, and confidentiality plus a mutual non-disparagement clause live.)* Your number, restated. $2.4M, built as 75% of your expert's mid-case lost-profits figure, with a client floor of $1.5M anchored to what he invested in the territory, not to any measure of recovery. Their ledger. - Exposure if they lose: $3.1M claimed, but their counsel is almost certainly telling the carrier the Daubert challenge cuts the credible number to roughly $1.6–2.0M. Call it $1.8M. - Their probability of loss: they will price this at 35–40% after the COO deposition. At 38%: $684,000. - Defense costs to verdict: $410,000, and this is real money to the carrier because it erodes nothing: it is defense-outside-limits spend the committee sees monthly. - Priced disruption: two more executive depositions and a June trial that pulls their VP of Merchandising for three weeks. They put a soft $100K-150K on this. Call it $125,000. - Value of a defense verdict to them: meaningful, because of the eleven other distributor disputes. Call it −$200,000 against the settlement value. - Their rational range: roughly $1.02M. Their authority-driven range is different, which is the actual problem; see below. Your ledger, net. At $1.5M: less 40% contingency ($600K), less $180K advanced costs, leaves the client roughly $720K in about 60 days. Trial at an expected $1.8M × your realistic 55% = $990K gross; less the 40% fee ($396K) and the $180K of advanced costs, that is $414K, and roughly 14 months to judgment plus appeal risk and time value take it down from there: call it $250K expected, in 2028. Your floor of $1.5M is not a floor built on arithmetic; it produces a better outcome than trial by a wide margin. Side by side. Their rational ceiling is near $1.02M; your client's stated floor is $1.5M. There is no zone of agreement at today's numbers: the gap is roughly $480,000. Everything below is about closing it. The approval story. The adjuster has $750K. To go above it she has to write to the committee, and the two sentences she needs to be able to say are: *"The Daubert ruling could go either way and our own expert put credible lost profits at $1.6M"* and *"Settling now avoids a June trial that puts our merchandising VP on the stand in eleven pending matters."* The first sentence requires her own expert's number to be at $1.6M or above. You do not know whether it is, and that single fact is worth more to you than any argument you will make at mediation. Ask the mediator to find out. Repeat player. Eleven similar disputes. Paying $1.5M here sets a number their other eleven counterparties will hear about within a quarter. This is the real reason the committee will resist, and it is why confidentiality is worth far more to them than you have priced it. Three risks you have not priced. (1) The prevailing-party fee clause runs both ways. If you lose at trial you are looking at their $410K plus what they have already spent [UNVERIFIED - confirm the clause's scope and whether the tortious interference count is covered]. (2) The January Rule 68 offer may already have cost consequences running [UNVERIFIED - check Arizona's rule; do not assume the federal version applies]. (3) Collection is not an issue here, so stop discounting for it. Cheap concessions. Confidentiality with a liquidated-damages clause: costs you nothing, and given eleven pending matters it is the single most valuable thing on your side of the table. Structured payment across two fiscal years: costs you time value on roughly half the money and moves them a bracket, because it lands in two budgets. Mutual non-disparagement, free. What you are overvaluing. The COO's deposition. It was bad, but it goes to breach, which was never the contested issue. Their exposure is being priced on damages, and your damages expert is the one under challenge. Walk-away. Roughly $700K: at that number the client nets 0.6 × $700K − $180K = $240K, just under the $250K trial expectation, so anything below it makes trial the better bet. That holds only if the fee clause does not shift both ways. If it does, a 45% chance of paying their $410K takes about $185K out of the trial expectation and the walk-away falls to roughly $400K; take almost anything above that. What your answers changed. The $750K adjuster ceiling moved the most. Without it this reads as a $480,000 gap between two ledgers and you spend the day arguing the gap; with it the obstacle is a signature: an adjuster who has to write up to a committee, and two sentences she has to be able to say when she does. The $410,000 of remaining defense spend is the second. Take that line out of their ledger and their rational ceiling falls to roughly $609,000, the gap widens to about $891,000, and my advice would have been that this case does not close at mediation at all.
Why this prompt is built the way it is
## Framework 1. **Pin the number.** Restate it and the theory behind it. A position that cannot be stated as one number is finding one. 2. **Their ledger, with arithmetic shown.** Exposure times their probability, plus defense costs to verdict, plus priced disruption, less the value of a defense verdict. 3. **Your ledger, net.** After fees, costs, time value, and collection risk. The two go side by side. 4. **The approval story.** Name the human who signs and write the two sentences they must be able to say to their boss, then say whether the fact supporting those sentences exists. 5. **Repeat-player pricing.** What settling here costs them across the next twenty matters, and what that does to their reservation price. 6. **The unpriced three.** Appeal, collection or insolvency, counterclaim and fee application, plus the venue's fee-shifting, offer-of-judgment, and interest rules, marked where uncertain. 7. **Cheap concessions.** What costs you least and is worth most to them, and the one thing you are overvaluing. 8. **Walk-away arithmetic.** The number at which trial becomes the better expected outcome, and what has to be true for it to hold.