Law prompt of the day

Fact-pattern issue list: claims, defenses, and research questions from an anonymized intake

A new matter arrived with a messy fact summary. Turn it into an organized issue list so research starts from the right questions.

You are a research attorney building an issue list from an anonymized fact pattern for a licensed lawyer's review — an organizational aid, not legal advice. I will paste the facts as we know them.

Produce:

A) FACT ORGANIZATION — the facts in chronological order, each tagged as documented, client-reported, or unknown.

B) POTENTIAL CLAIMS AND DEFENSES — the legal issues the facts may raise, each with the elements stated in general terms and which facts appear to bear on each element. Mark elements with no supporting fact yet.

C) RESEARCH QUESTIONS — for each issue, the specific question to research, written as "RESEARCH NEEDED: [jurisdiction + question]".

D) FACT GAPS — the facts that would most change the analysis, and the document or witness likely to supply each.

E) TIME-SENSITIVE FLAGS — any issue where a deadline might matter, flagged for the attorney to verify; never compute the date.

Inputs: [ANONYMIZED FACT SUMMARY] · [JURISDICTION] · [CLIENT'S GOAL] · [DOCUMENTS RECEIVED SO FAR]

Rules: Anonymize before you paste and keep privileged or confidential client information out of a consumer LLM (ABA Formal Op. 512). Do not invent facts, elements, or authority — state elements generally and route every jurisdiction-specific point to a research question. Verify all elements and any limitation period against current law in the jurisdiction before relying on the list. This is an organizational aid; the attorney decides what issues exist.

Why it works — Early research goes wrong when it starts from the first theory that comes to mind rather than from the facts. Tagging each fact by its source, mapping facts to general elements, and turning every jurisdiction-specific point into a research question gives the attorney a structured starting point without the model asserting law. Flagging deadline risk without computing dates keeps the most dangerous error class with a human.