Research

Find the counterargument that should worry you

Builds the single best argument against your position at full strength, names the authority and the killer fact the other side will lead with, and commits to one of three ways to answer it.

About 14 minadvancedLitigation

Your prompt5,403 characters

Still to fill in: Your argument, Operative facts, Controlling court and law

RoleYou are the lawyer on the other side of this case, working without the loyalty. You build one counterargument at full strength rather than a list of weak ones, because a list of weak objections is what lawyers write to feel prepared. You refuse to strawman: if the best version of the argument against this client wins, you say so plainly and then tell them what to do about it. You also know which facts hurt most: usually the ones the client produced themselves.What I needFind the strongest counterargument to my position and build it at full strength. My argument is Your argument, the controlling court is Controlling court and law, and this gets tested at Where this gets tested.InputsMy argument: Your argument Operative facts: Operative facts Controlling court and law: Controlling court and law Authority in play: Authority in play Where this gets tested: Where this gets testedHow to work this1. Restate my argument in one sentence the way the other side would put it: accurate but uncharitable. If I would not recognize my own position in it, you have strawmanned me before you started. Fix it, then attack. 2. Build one counterargument in three to five sentences at maximum strength, written in their voice as an assertion. If a second theory is genuinely worth having, give it one line at the end labeled "runner-up" and move on. 3. Name the authority they lead with and quote the words they will pull from it. Mark anything not drawn from Authority in play as [UNVERIFIED - pull before relying]. 4. Identify the killer fact: the single item in Operative facts that does the most damage, and say how it reaches the record: a document, a deposition answer, an admission, or a gap they will exploit. 5. Give exactly three responses: distinguish the authority, reframe the rule so their case lands outside it, concede and pivot, and state the cost of each: what it gives away, and what it exposes at the next stage or on cross. 6. Commit to one response. Say why the other two lose here, and name the fact or ruling that would change your recommendation. 7. Close by assuming the counterargument wins. Name the fallback position, what it is worth, and whether it survives Where this gets tested.Ask me firstYou have my argument, the facts, and the authority in play above. Ask me these four questions (what the record alone will not tell you), then stop and wait: 1. What does the other side actually have? I want the counterargument they can make on the documents in their possession, not the one that would work if discovery had gone differently. 2. Which of my facts is worst, and where did it come from: my client's own document, a deposition answer, or an inference someone drew? The killer fact is usually one we produced. 3. Has opposing counsel already signaled a theory in a demand letter, an answer, a meet-and-confer, or argument at a prior hearing? Their real theory beats my guess at it. 4. If I lose this issue outright, what happens: the case ends, one claim drops, or a settlement number moves? That decides whether concede-and-pivot is even on the table. 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 formatThe counterargument in three to five sentences, written in the other side's voice. Then: the authority they lead with, with quoted language; the killer fact and its source in the record; three responses, each with its cost; a single recommendation with the reason the other two lose; and a short "if this wins anyway" fallback. Close with assumptions. End with one line naming the two of my answers that changed this counterargument the most, and which response you would have recommended without them. If an answer changed nothing, say so. That question should not have been asked.Never do this- If this counterargument would work against any party pressing any version of this claim, it is too generic. You have written a doctrine summary. Rebuild it from my facts and my record. - No hedging filler. Cut "arguably," "it should be noted," "the other side might contend," and "it depends." Write it as an assertion in their voice. Do not tell me to consult an attorney. I am the one this lands on. - Every case, statute, quotation, and record reference must come from my inputs or carry [UNVERIFIED - pull before relying]. Never invent a citation, and never invent a document that conveniently supplies the killer fact. - Where you do not know whether the record contains the fact their argument needs, say you do not know and mark it as something to run down in discovery. Do not smooth over the gap with fluent prose. - Do not pad. One counterargument fully built beats five sketched, and do not spread the recommendation across all three responses to avoid choosing. Length is not value.Before you answer- Did I steelman, or did I quietly write a version I could beat? - Is the killer fact a real item from the inputs, or something I assumed into existence? - Did I commit to one response, or hedge across all three? - Would this counterargument work against a different case on the same doctrine? It should not. - Is any authority unmarked and unverified?

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.

Assume the bench has read the moving papers first, wants this off the docket, and is looking for a clean reason to grant. The counterargument is halfway home before my response is opened. Read it from that chair, as the judge already leaning their way. Say where the response goes soft, what question from the bench would break it, and the answer that has to be ready. Then tighten the response once more so it does not fold under that questioning.

Go deeper

Pushes the work further once the basics are right.

None of this helps until it is a paragraph in the argument section. Draft the rebuttal paragraph using the recommended response: four to six sentences, leading with the reframed rule rather than with their case, applying it to my facts, and closing on the record cite that does the work. Ready to drop into the argument section, with any unverified authority still marked.

Before you run it

What to gather first

Your argument in two to four sentences, as you would put it to the court
The operative facts, including the ones that hurt
The controlling court and governing law
Any authority you or the other side are already relying on
Where this argument gets tested and who decides it

Watch for

The model invents authority for both sides with equal fluency. Pull and Shepardize every citation before you rely on it or plan around it.
A counterargument you can knock down easily is worthless here. If the output reads comfortable, it was strawmanned. Run it again with the worst facts foregrounded.
Concede-and-pivot is sometimes the right call. Do not reject it reflexively because it feels like losing ground.
This analysis is candid about your weaknesses. Keep it in work product, mark it accordingly, and do not paste it into anything discoverable or into a client email chain that may be produced.
The killer fact often sits in a document nobody has read closely. Confirm what the record actually says before you build a response around the model's version of it.

What comes back

A three-to-five-sentence counterargument written in the other side's voice, then the authority they lead with and the language they will quote, the killer fact and how it reaches the record, three responses each with its cost, one committed recommendation with the reason the alternatives lose, and an "if this wins anyway" fallback. Closes with assumptions.

See an example of what you’ll get

*(After you answer the four questions: on their side of the production they hold the calendar entry and the Q4 sales numbers and nothing else predating March 3; the February 14 entry is the worst fact and it came out of your own Bates range; their answer and their meet-and-confer letter both lead with the pre-existing-intent theory; and losing causation drops the Title VII count but leaves FEHA standing and moves the mediation number.)* The counterargument (their voice). Plaintiff's causation theory rests entirely on a four-day interval, and that interval proves nothing here because the decision predated her complaint. CEO Hayes's February 14 calendar entry ("VP Sales transition planning") is seventeen days before the March 3 HR complaint, and Hayes has testified without contradiction that he had already resolved to replace Plaintiff after the Q4 numbers came in. Under *Clark County School District v. Breeden*, an employer need not suspend a previously contemplated decision because an employee has since engaged in protected activity. Where the record shows the adverse action was in motion before the protected activity, temporal proximity carries no probative weight at all, and Plaintiff has nothing else. Authority they lead with. *Clark Cnty. Sch. Dist. v. Breeden*, 532 U.S. 268, 272 (2001) (per curiam) [UNVERIFIED - pull before relying]: "Employers need not suspend previously planned transactions upon discovering that a Title VII suit has been filed, and their proceeding along lines previously contemplated, though not yet definitively determined, is no evidence whatever of causality." The killer fact. The February 14 calendar entry. It is your client's own document, produced in your own Bates range, and it is the only item in the record that gives the defense a pre-complaint anchor. Expect it as Exhibit 1 to the Hayes declaration and as the first line of their statement of undisputed facts. Three responses. 1. *Distinguish.* *Breeden* involved a transfer already set in motion through documented steps; here there is one ambiguous eleven-word entry, no memo, no HR consultation, no successor identified. Cost: concedes that a documented pre-existing plan would defeat causation, which invites the defense to build one out of Hayes's testimony. 2. *Reframe.* The rule is not that any pre-existing intent neutralizes timing: it is that a decision "previously contemplated" must be identifiable from something more than a bare calendar entry, and *Breeden*'s own language turns on steps "previously contemplated" and communicated. Cost: requires the judge to read *Breeden* narrowly, and hands the defense a clean issue if you lose the motion. 3. *Concede and pivot.* Accept that Hayes was considering a change in February, then argue the complaint accelerated an indefinite plan into a four-day termination, and that acceleration is itself the adverse action. Cost: gives up the cleanest version of the prima facie case and shifts the fight to pretext, where you need the missing-PIP evidence to carry more weight than it comfortably will. Recommendation: reframe. Distinguishing concedes the governing principle and lets them rebuild it with declaration testimony. Concede-and-pivot signals weakness before the judge has read anything else. Reframing keeps *Breeden* confined to decisions with documented steps and lets the four-day gap keep its force. This changes if Hayes's deposition produced testimony about a specific successor candidate before March 3: in that case, pivot. If this wins anyway. The FEHA claim survives on a lower causation standard, and the failure to document any performance concern over six years still reaches a jury on pretext. That is worth substantially less at mediation, but it is not a dismissal. Assumptions. I assumed the February 14 entry is not accompanied by any email, memo, or HR record [verify - this is the whole motion]. I assumed Hayes has been deposed and gave the pre-existing-intent testimony [verify]. What your answers changed. Their signaled theory did the most. Because the pre-existing-intent argument is already on paper in their answer and their meet-and-confer letter, this is the counterargument you will actually face rather than my guess at one. Without it I would have gone at the protected-activity element instead, asking whether the March 3 complaint stated a reasonable belief of unlawful conduct, and the killer fact would have been the content of that complaint rather than the calendar entry. Your answer about what losing costs you is what keeps concede-and-pivot on the list as response 3: the FEHA claim survives and the number moves rather than the case ending, so a pivot has somewhere to land. If losing causation ended the case, response 3 comes off entirely, the menu is two options, and the “if this wins anyway” paragraph becomes a dismissal note rather than a valuation one. Which of your facts is worst changed nothing I did with it. The February 14 entry is the only pre-complaint item in the record, so it was carrying the killer-fact slot whichever of us named it, and knowing it came out of your own production bought one sentence rather than a different recommendation.

Why this prompt is built the way it is

## Framework 1. **One counterargument, fully built.** Three to five sentences at maximum strength. A list of weak objections is not the deliverable. 2. **Steel-manned in their voice.** Written as an assertion, the way the other side would write it, not as "they might argue." 3. **Authority-backed.** The case or statute they lead with, and the language they will quote from it. 4. **The killer fact.** The one item in the record that does the most damage, and where it comes from: a document, a deposition answer, an admission, a gap in the timeline. 5. **Three responses with costs.** Distinguish the authority. Reframe the rule so their case lands outside it. Concede and pivot. Each one states what it gives away. 6. **Commit.** One recommendation, why the other two lose, and the fact or ruling that would change the answer. 7. **Plan for losing.** If the counterargument simply wins, name the fallback position and what it is worth.