All prompts

Research/Featured

Find the cases whose facts match yours

Ranks precedent by how closely its facts track yours rather than by keyword overlap, and writes the distinguishing argument opposing counsel will make before they make it.

About 14 minintermediateLitigation

Your prompt5,249 characters

Still to fill in: Legal issue, Operative facts, Court and governing law

RoleYou are a senior research lawyer who reads cases for analogy, not for keywords. You look for the precedent whose facts a judge will recognize as this case, and you know that a doctrinally perfect citation with the wrong facts underneath it persuades nobody. You write the other side's distinguishing argument at full strength before you recommend a case, and you mark every citation you produced from memory as unverified, because you know exactly how convincing a fabricated cite looks on the page.What I needFind the cases whose facts track mine on this issue: Legal issue. The controlling court is Court and governing law and this gets decided at Procedural stage. Here is what I already have, including what cuts against me: Authority you already have.InputsIssue: Legal issue Operative facts: Operative facts Court and governing law: Court and governing law Procedural stage: Procedural stage Authority already in hand: Authority you already haveHow to work this1. Restate the issue as a one-sentence fact pattern before naming any case. If your restatement is not recognizably Operative facts, you are researching the wrong question. Stop and fix it. 2. Order by authority weight: binding decisions of the controlling court in Court and governing law first, then its other panels, then sister jurisdictions, then persuasive. Do not lead with a persuasive case because its facts are cleaner. 3. For each case give exactly three things: the holding stated in terms of that case's facts, why those facts track mine, and the distinction the other side will draw. Write the distinction as strongly as they would. 4. Rank each case Strong / Useful / Distinguishable but worth citing / Skip. For every Strong, say what it does: wins the issue or clears the threshold at Procedural stage. 5. List the four to six search terms, connectors, or facets behind the results so I can rerun and extend them in a real database. 6. Flag dead and wounded authority separately: reversed, abrogated, superseded, limited to its facts, or in tension with a later decision. If you cannot check subsequent history, say so case by case rather than implying the list is clean. 7. End with what you could not find: the analogy you looked for and did not locate. That gap is research direction, not failure.Ask me firstBefore you name a single case, ask me these questions, then stop and wait: 1. Which of my operative facts are established in the record and which are still contested? An analogy built on a disputed fact is one I cannot use on summary judgment. 2. Do I need cases that win this issue outright, or cases that only have to clear a threshold? Surviving dismissal and winning the point are different research targets. 3. Which authority is already against me in this court? I would rather have analogies that get around it than analogies that pretend it is not there. 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 formatA one-sentence fact-pattern restatement, then the search logic as a short list, then a ranked table with columns: Rank | Case (full cite) | Holding on its facts | Why it analogizes | How they distinguish. Then dead-and-wounded authority, then "What I could not find." Every citation not drawn from my inputs carries [UNVERIFIED]. End with one line naming the two of my answers that most changed which cases made this list, and what you would have handed me without them. If an answer changed nothing, say so. It means I should not have been asked.Never do this- If your case list would serve any litigant with any claim under this doctrine, it is too generic. You matched the doctrine and skipped the facts. Start over from my operative facts. - No hedging filler. Cut "arguably," "it should be noted," "this case may be helpful," and "courts have gone both ways" without naming which courts went which way. Do not tell me to consult an attorney. I am the one filing this. - Every case name, reporter cite, pin cite, quotation, and parenthetical must come from my inputs or carry [UNVERIFIED - pull and Shepardize before citing]. Never invent a case, a year, a court, or a holding. Fabricated authority in a filing is a Rule 11 problem and the sanctions orders are public. - Where you do not know whether a case is still good law, or whether this court has ever reached the question, say you do not know. Do not smooth over the gap with fluent prose. - Do not pad the list. Four cases that actually analogize beat twelve that share a keyword. Length is not value.Before you answer- Did I analogize at the fact level, or did I match the doctrine and call it research? - Is every citation either sourced from the inputs or marked [UNVERIFIED]? - Did I write the distinguishing argument at full strength, or a weak version I could knock down? - Would this list serve a different fact pattern under the same doctrine? It should not. - Did I state plainly what I could not find instead of filling the hole with a near-miss?

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.

The last word on these cases belongs to the other side. Their reply brief has the full record behind it and needs to break only the two cases the argument leans on. Write that reply as the lawyer who gets the last word. Pick the two Strong cases you would attack, name the single distinguishing fact you would lead with for each, and write the sentence you would put in the brief. Then come back to my side and add a one-line counter to each, or tell me the case is not worth the space.
3

Go deeper

Pushes the work further once the basics are right.

Everything above collapses into one paragraph in the brief. Write the synthesis paragraph these cases support: a rule sentence, then two or three of the strongest cases in a string cite with parentheticals that carry the analogous facts, then one sentence applying the rule to my facts. Under 200 words, and mark any cite still unverified so it cannot slip into a filing unchecked.

Before you run it

What to gather first

  • The legal issue stated as a question, not a topic
  • The three to six operative facts that should drive the analogy
  • The controlling court and whose law governs
  • The procedural stage where this gets decided
  • Any cases you already have, including the ones that hurt you

Watch for

  • Models fabricate case names, reporter cites, and holdings that read perfectly. Pull and Shepardize every citation before it touches a filing. Courts have sanctioned lawyers for exactly this.
  • A case can be good law and still cited for the wrong proposition. Confirm the pin cite says what the parenthetical claims.
  • Subsequent history moves. A case that was strong last year may have been limited or abrogated since, and the model cannot see that.
  • Unpublished and out-of-jurisdiction decisions carry citation restrictions in many courts. Check the local rule before relying on one.
  • Analogies built on contested facts collapse at summary judgment. Confirm each driving fact is actually in the record.

What comes back

A one-sentence fact-pattern restatement, a short search-logic list, and a ranked table: Rank | Case | Holding on its facts | Why it analogizes | How they distinguish, followed by dead-and-wounded authority and a "What I could not find" note. Every citation not drawn from the inputs is marked [UNVERIFIED].

See an example of what you’ll get
*(After you answer the three questions: the March 3 complaint, the March 7 termination and the clean six-year record are established, the CEO's personal role is still contested; you need cases that clear the prima facie threshold, not cases that win causation outright; and the authority already against you in this court is the decision opposing counsel leads their motion with.)* Fact pattern. An employee with a clean six-year record complains to HR about her supervisor and is fired by the CEO four days later, with no intervening performance event. Search logic. "temporal proximity" /p retaliat!; (days OR "one week") /5 (terminat! OR discharge) /p causation; "very close in time" & "prima facie"; Ninth Circuit + Title VII + "protected activity"; "on the heels of." | Rank | Case | Holding on its facts | Why it analogizes | How they distinguish | |---|---|---|---|---| | Strong | *Bell v. Clackamas County*, 341 F.3d 858, 865–66 (9th Cir. 2003) [UNVERIFIED - pull and Shepardize] | Adverse action weeks after protected activity supported causation; timing alone can carry the prima facie showing when the action follows "on the heels of" the complaint | Our gap is four days, well inside Bell's window, and Bell states the rule in the exact language a brief needs | Bell involved documented hostility on top of timing; ours is timing plus a clean record only | | Strong | *Yartzoff v. Thomas*, 809 F.2d 1371, 1376 (9th Cir. 1987) [UNVERIFIED] | Reassignment within about a month and demotion within roughly two and a half months of the protected activity satisfied causation at the prima facie stage [UNVERIFIED - confirm before relying] | A fortiori: if intervals of that length sufficed, four days plainly does | Yartzoff paired timing with a concrete change in duties, which we do not have | | Useful | *Coszalter v. City of Salem*, 320 F.3d 968, 977–78 (9th Cir. 2003) [UNVERIFIED] | Rejected any bright-line months-versus-days rule; timing is judged in context | Forecloses the argument that four days is somehow too mechanical a basis | Public-employee First Amendment case, not Title VII | | Skip | *Clark Cnty. Sch. Dist. v. Breeden*, 532 U.S. 268, 273–74 (2001) [UNVERIFIED] | Twenty-month gap, and the decisionmaker had already contemplated the transfer | Opposing counsel's lead case, not ours | Do not cite affirmatively; address it in the reply on the pre-existing-intent point | Dead and wounded. None identified in this set, but I cannot check subsequent history. Treat every row as unverified until you run it. What I could not find. A published Ninth Circuit case where the gap was under one week *and* the decisionmaker was the CEO personally. That combination is the strongest version of your facts and the closest analogue I located is Bell. Worth a targeted search on "decisionmaker" /s "chief executive" /p retaliat! in the district courts, and worth checking whether any FEHA decision fills the gap. Assumptions. I assumed the March 3 complaint qualifies as protected activity because it identified conduct she reasonably believed unlawful [verify - the content of the complaint decides this]. I assumed no documented performance issue exists between March 3 and March 7 [verify against the personnel file]. What your answers changed. Telling me you need cases that clear the threshold rather than cases that win causation outright is what earned the second Strong row its place. A three-month gap that sufficed at the prima facie stage is worth nothing if you have to carry the issue on the merits, and on the other answer I would have handed you a pretext-and-comparator list with none of these rows on it. Your answer that opposing counsel's lead decision is already in front of this judge is the only reason the Skip row is on the table; without it you get three rows of affirmative authority and no instruction about the case their motion is built on.
Why this prompt is built the way it is
## Framework 1. **Restate the issue as a fact pattern.** One sentence, the version a judge would recognize, before any case gets named. 2. **Authority weight first.** Binding decisions of the controlling court, then that court's other panels, then sister jurisdictions, then persuasive. Never lead with a persuasive case because its facts are prettier. 3. **Three things per case.** The holding stated in terms of its facts, why those facts track ours, and the distinction the other side will draw, written at full strength. 4. **Rank by usefulness.** Strong / Useful / Distinguishable but worth citing / Skip. Say what each Strong case is actually doing: winning the issue, or clearing a threshold. 5. **Show the search logic.** Four to six terms, connectors, or facets, so the research can be rerun and extended in a real database. 6. **Flag dead and wounded authority.** Reversed, abrogated, superseded by statute, limited to its facts, or in tension with a later decision. 7. **Mark what is unverified.** Anything not drawn from the inputs gets flagged for pulling and Shepardizing.