Gains:
- Ability to produce a regular draft skeleton for the case/petition in the fact-law-demand structure
- Ability to strengthen the petition by questioning counter-arguments and possible objections in advance
- Ability to mark the boundaries of procedural rules and evidence base and prepare them for human inspection
Writing a petition is turning a messy file into a compelling narrative: events line up, each fact is anchored in law, a clear demand is finally formulated. This is mentally intense and time-consuming work. Artificial intelligence (AI) is a powerful aid in scaffolding this process, pre-empting counter arguments and editing the text. However, a petition is a formal document filed with the court; Every fact in it must be based on evidence, every reference must be based on a primary source and must comply with procedural rules. In this unit, we will learn how to use AI end-to-end in petition drafting, but maintain human control at every step.
Let's clarify the terms. A petition is an official written document submitted to a court or an authority, containing the request and justification. Procedure is the rules that govern how a case will be conducted; such as deadlines, formal conditions, competent court. Facts are the concrete facts on which the dispute is based. Evidence is a document, witness or record that proves a fact. Petitum is the concrete decision requested from the court at the end of the petition. There is also a framework that we will use frequently: the fact-law-demand structure; In other words, first what happened, then what legal basis, and finally what is required.
Establishing the Petition Skeleton
The safest and most efficient use of AI is not to make up content, but to put the facts and bases you provide into an orderly framework. Step by step:
- Give the facts chronologically and neutrally. What happened, when, on what document is it based?
- You provide the legal basis. Because of the AI risk of attribution fabrication (see Unit 5), you enter items and decisions as verified.
- Have the skeleton produced. An orderly outline in the fact-law-demand structure.
- Question counter arguments. Consider possible objections and the response to each in advance.
- Highlight procedural and evidence gaps. Which claim depends on which evidence, which period is critical.
- Prepare for human inspection. The draft is a beginning that the lawyer will review and bear responsibility for.
Your role: a legal assistant who drafts the petition. Below I provide the facts and VERIFIED legal basis. Turn these into an organized petition skeleton with the structure of "fact - legal basis - result of request". Rules: - Use only the facts and grounds I have given; new fact or reference FABRICATION.- Mark [EVIDENCE: ...] on which evidence each claim is based; If the evidence is unclear, write [EVIDENCE MISSING]. - Be formal, clear and assertive in language; avoid exaggeration and emotional expression.<facts>[chronological facts and the documents on which they are based]</facts><legal_basis>[verified article/decision references]</legal_basis><request>[requested from the court]</request>
Tip: You give the legal basis to the AI, don't ask it. This single rule prevents the most serious error in petitions (fabricated precedent) from the very beginning. AI is a good writer and editor; It is not a reliable legal database.
Considering Counter-Arguments in Advance
A strong petition anticipates and answers not only its own request but also the other party's possible objections. AI can work like a “devil's advocate” and show you your weaknesses.
Look at my draft petition below from the other party's perspective. 1) Write the 5 strongest objections that can be raised against this request and the facts. 2) For each objection: what its basis may be, what weakness it targets in our draft. 3) Suggest the answer we can give to each objection and what additional evidence we will need. Invent a new legal reference; focus only on the logic of the argument.
It is also useful to examine the shortcomings of the petition in terms of procedure and form. AI asks “do these exist?” with a checklist logic. he asks; But the accuracy of the deadlines and authority rules must be confirmed by the legislation.
Scan this petition with a checklist for form and completeness: party information, subject, open request result, connection of facts to evidence, legal basis, date and signature area. List those that are missing or unclear. Mark the duration and authority rules as "must be verified by legislation"; Don't make up your own time/authority rule.
Another safe use of AI in petition writing is in editing language and tone. A raw draft is often either too emotional or messy; AI can simplify text and tone it down to a formal tone without changing the legal content. Yet control of meaning remains human.
Rewrite the following petition paragraph in more formal, clear and consistent language WITHOUT CHANGING the legal content and request. Translate emotional, exaggerated or accusatory statements into objective statements. Mark any change that might affect the meaning as "needs semantic check". Adding new facts or claims; just improve the language of existing content.
Tip: Aggressive or emotional language in a petition often reduces its persuasive power. An objective, fact-based tone is stronger; AI is a good help in capturing this tone, but the lawyer controls the legal accuracy of the text.
Weak Prompt / Strong Prompt
Weak prompt: Write a petition on this issue and include precedent decisions.
The result: a narrative made up without knowing your facts, unverified (possibly fake) precedents, and a procedurally unreliable text. If used directly, it is risky in terms of both litigation and professional liability.
Powerful prompt: [assistant role + your given facts and verified bases + fact-law-claim skeleton + [EVIDENCE] signs + "fabricated fact/citation" + separate counterargument analysis + figure checklist]
The result: a draft based on your actual file, with evidence links marked, prepared for counter-objections, and suitable for human review.
Layers of the Petition Draft
layer
The role of AI
man's role
facts
Chronological arrangement
Truth and evidence bond
Legal basis
Embed in text
Providing and verifying attributions
argument
Outline and counterargument
Strategy and choice
As a result of the request
sharpening
legal compliance
Procedure/form
checklist
Confirm duration and authority
final text
—
Approval and responsibility
Three Mini Cases
Case 1 — Closing the weak point. An attorney tested a debt lawsuit petition with a counterargument prompt. AI showed that the "statute of limitations" objection may be the strongest counter move, and that this issue is not addressed at all in the draft. The lawyer added the document proving that the receivable is not statute-barred to the petition from the beginning. At the hearing, the other party actually claimed statute of limitations; The prepared party responded to this with documentation.
Case 2 — Evidence gap. A legal team noticed three allegations in an employment lawsuit that the AI had flagged as [EVIDENCE MISSING]; These allegations were strong but not backed by any documentation. The team collected witnesses and records for two allegations, and removed the third allegation, for which no evidence was found, from the petition. Instead of a petition full of evidence-free claims, a more solid text was presented, every claim of which can be proven.
Case 3 — Preventing fabricated precedent. A new intern had asked the AI for “supporting rulings” for a petition and was about to add the incoming citations. The bureau's "we give the AI the resources, we don't ask it" rule came into play; The references were checked from the official source and two were found to not exist. Thanks to the rule, a petition containing a false precedent never reached the court. Measured result: no corrections/retractions due to unverified attribution that year.
Common mistakes
- Asking for a precedent decision from AI and adding it without verifying it. The most dangerous mistake in petitions; You provide and verify the bases.
- Making AI adapt the facts. AI does not know your file; The facts must come from you, neutral and based on evidence.
- Not thinking of counter arguments. A petition that only writes its own request is weak in the first objection.
- Bypassing the evidence link. Every claim must be supported by evidence; [EVIDENCE MISSING] signs should be taken seriously.
- Relying on AI for time and authority rules. Procedural rules must be confirmed by legislation.
- Considering the draft is final. The responsibility for the petition lies with the lawyer; The outline is just the beginning.
In summary
AI is a powerful assistant that organizes facts in the petition draft, suggests counter-arguments, and scans the text with a checklist. But the petition is an official document: the facts must come from you and based on evidence, you must provide and verify the legal basis, and the procedural rules must be confirmed by the legislation. Use AI to build the fact-law-claim skeleton and point out weak points; The responsibility for the accuracy of the content and the text always remains with the lawyer. AI generates draft; The competent professional decides what to present to the court.
Application task
Choose a conflict scenario. (1) Give the facts chronologically and in evidence, with verified bases, and have a draft petition produced with the skeleton prompt; Examine all [EVIDENCE] signs. (2) Pull out the 5 strongest objections and their answers with the counter-argument prompt; strengthen the draft accordingly. (3) Run the shape checklist prompt and fill in the missing items. (4) Include all signals regarding duration and authority in the "to be verified from legislation" list and make a plan for verification from the real source.
checklist
- [ ] Were the facts given from you chronologically and based on evidence?
- [ ] Are the legal bases provided by you and verified references?
- [ ] Was the draft organized in the fact-law-demand structure?
- [ ] The evidentiary link of each claim has been marked, have those with [EVIDENCE MISSING] been addressed?
- [ ] Have the counterarguments and their responses been studied in advance?
- [ ] Have the duration and authority rules been confirmed by the legislation?
- [ ] Did the final text result in attorney approval and liability?