Starter ChatGPT Prompts for Lawyers
Starter ChatGPT Prompts for Lawyers

Starter ChatGPT Prompts for Lawyers

Get Started with ChatGPT


These use cases and prompts are quick ways to get started with ChatGPT, and explore the different ways you can make your workflow more operationally efficient!
ChatGPT for Law
Table view of ChatGPT for Law
logo
Use Case
Task Type
logo
Prompt
Jurisdiction
logo
Capability
Notes
research
“You are a disputes lawyer at a boutique firm. Task: find the most recent cases on [issue: e.g., constructive dismissal after remote-work policy change] in [jurisdiction]. Inputs: my fact pattern below. Output: a mini-brief with: (1) rule/standard, (2) top 5 cases w/ 1–2 sentence holdings, (3) relevance to my facts, (4) red-flag risks, (5) follow-up angles.”
Civil Disputes
Web Search
Press “+” → “Web.” Ask for links and dates. Always verify with KeyCite/Shepard’s before relying. Tip: switch to GPT-5 Thinking for deeper reasoning if needed.
write
“Role: principal of an employment & commercial disputes firm. Task: draft a client alert on [new case/regulatory change]. Audience: HR leaders and GCs. Include: 3 bullet takeaways, ‘What this means’ by sector, action checklist, and a short disclaimer.”
Writing
Writing
Press “+” → “Canvas.” Paste sources, then iterate tone/length. Keep citations in footnotes; avoid confidential client facts.
plan
“Create an issue tree and decision paths for [matter]. Include pleadings theories, key elements to prove, discovery targets, settlement levers, and likely motions from the other side. Show quick-win vs. long-shot paths.”
Commercial
General
Planning
Ask for a one-page “battle card” at the end. Use it to align team roles and deadlines.
analyse
“Given these PDFs (upload below), extract each case’s holding, rule statement, key quotes, and how the court distinguished similar facts. Then make a comparison table and a draft rebuttal paragraph for each adverse case.”
General
Vision
Document summary
Upload PDFs/scans; Vision works on long opinions. Mark sensitive info. Have ChatGPT produce a cite list you can quickly validate.
analyse
“Here’s an .xlsx of 5 years of terminations and claims. Task: clean the data, flag missing fields, then run descriptive stats and charts on tenure, role, reason, protected classes, and outcomes. Output: a 1-page variance analysis + charts.”
General
Writing
Data Analysis
Upload .csv/.xlsx (and a data dictionary). Ask for charts/tables and for ChatGPT to explain any anomalies.
learn
“Use Study Mode to coach me on [Act/Rule section]. Start with plain-English overview, then spaced-repetition Q&A, then a 10-question quiz with model answers and common traps.”
General
Study Mode
Press “+” → “Study and learn.” Great for onboarding juniors. Keep questions jurisdiction-specific.
rehearse
“Roleplay as a skeptical judge/jury. Scenario: [hearing/depo]. Push on weak spots in our theory and evidence. After 10 minutes, switch to feedback with a checklist of improvements.”
Writing
Advanced Voice
Click the wave/microphone icon to practice out loud. Ask for timestamps and a summary of your strongest/weakest answers.
investigate
Deep Research prompt pattern:

Context: Principal at a disputes firm; New Zealand jurisdiction; Focus on [e.g., contractual penalties / restraint of trade / protected disclosures].

Objective: Decide whether to plead [theory] this quarter. Scope & Depth:

Compare [3–5] years of cases;
include appellate & notable trial decisions; note splits; exclude paywalled commentary.

Requirements: Cite each proposition; prefer sources ≤12 months old; flag conflicts and gaps.

Deliverables: Slide-ready brief with executive summary, table of cases, and appendix of quotes/pin-cites.

Constraints: 2-hour effort; budget notes for further manual research.
Writing
General
Deep Research
Use for structured, long-form web synthesis with citations. Review outputs critically; never skip your own citator check.

Learn more: https://openai.com/index/introducing-deep-research/.
orchestrate
“Act as a ChatGPT Agent. Task: visit [court sites] and gather current civil practice notes, filing checklists, and fee schedules for [jurisdiction]. Extract deadlines/limits into a sortable spreadsheet and flag any 2024–2025 changes.”
Writing
General
ChatGPT Agent
Supervise each step; approve sites before visiting. Good for public info and file organisation. Don’t paste client data.

Overview: https://openai.com/index/introducing-chatgpt-agent/.
orchestrate
“As a ChatGPT Agent, operate on my uploaded PDFs: read filenames, extract dates/parties, create an exhibit index with descriptions, and draft proposed bundle tabs. Output: index spreadsheet + draft cover sheet.”
Writing
General
ChatGPT Agent
Keep originals read-only; verify every extraction. For Bates stamping or court-specific templates, finish with your DMS tools.
investigate
“Deep Research the last [3] years of public judgments and media for [Opposing Counsel/Firm] in [topic]. Summarise recurring arguments, success rates, and remedy patterns; include sources and dates.”
Writing
General
Deep Research
Use for strategy colour, not for doxxing. Confirm identities and avoid conclusions about individuals beyond public records.
research
Role: principal at an NZ disputes firm.

Task: find the most recent NZ authorities on [issue, e.g., relief against cancellation for non-payment under a commercial lease].

Scope: NZCA/HC preferred; include any Tenancy Tribunal only if the dispute is residential.

Output: a mini-brief with
(1) rule,
(2) 5 key cases w/ neutral citations and 1-2 sentence holdings,
(3) fit to facts below,
(4) risks/splits,
(5) open questions.”
General
Civil Disputes
Web Search
Press “+” → “Web.” Ask for links + dates. Always run your own citator (LexisNexis/Westlaw NZ) before relying. Tip: for deeper reasoning, you can manually choose GPT-5 Thinking.
write
Audience: NZ SMEs and in-house counsel.

Task: 500–700 word client update on [e.g., Law Association (ADLS) 7th-ed lease changes / new Companies Act amendments].

Include: what changed, who’s affected, 3 takeaways, action checklist, and short disclaimer.”
Writing
Press “+” → “Canvas.” Paste source links. Ask for a footnoted version and a 100-word LinkedIn variant. Keep client facts out.
analyse
“Given these PDFs of executed leases and the TLANZ 7th-ed Deed extract, build a table comparing rent review, ratchet, OPEX recovery, assignment/subletting, damage/inaccessibility, and default/relief. Add plain-English risk notes + suggested fallback wording.”
Vision
Document summary
Upload scanned leases/marked-up clauses;

Vision handles long PDFs. Ask for a clean side-by-side plus a one-page ‘positions & fallbacks’ sheet.
plan
“Create an issue tree for [matter]. Map elements, burden, key evidence, likely defences, interlocutories (strike-out/summary judgment), settlement levers, and enforcement paths (judgment → statutory demand → liquidation). Tag quick-win vs long-shot.”
Planning
Ask for a one-page “battle card” with deadlines anchored to NZ court timetables and a checklist for affidavits/exhibits.
analyse
“Here’s an .xlsx of debts, demand dates, payments, and forum. Clean the data, flag missing fields, then chart recovery rates by forum (Tribunal $≤30k, DC, HC), time-to-recovery, and costs. Provide a 1-page variance analysis and a table of ‘next-best forum’ by claim size.”
Data Analysis
Upload .csv/.xlsx (plus a data dictionary). Ask for tables + charts and a sensitivity check by claim band (e.g., <$10k, $10–30k, $30–350k, >$350k).
learn
“Use Study Mode to coach me on [specific CCLA/Companies Act/Property Law Act section]. Start with a 2-minute overview, then spaced-repetition Q&A and a 10-question quiz with model answers and common traps.”
Study Mode
Press “+” → “Study and learn.” Great for juniors and refreshers before a hearing. Keep it section-specific.
rehearse
“Roleplay as a sceptical NZ judge/mediator on [issue]. Run 10 minutes of questioning, then give structured feedback with a checklist: Authority gaps, fact weaknesses, witness risks, and 3 crisp submissions.”
Advanced Voice
Click the wave/microphone icon to practise out loud. Ask for timestamps and a transcript.
investigate
Deep Research pattern (NZ-specific):

Context: NZ commercial landlord-tenant practice; TLANZ 7th-ed lease.

Objective: Decide negotiation posture on inaccessibility/damage and rent review (CPI/market/fixed).

Scope & Depth: NZ HC/NZCA cases and commentary 2012–present; include COVID-era cases; exclude residential.

Requirements: Cite every proposition; prefer sources ≤12 months old; flag splits and ADLS→TLANZ changes.

Deliverables: Slide-ready brief + table of clauses vs cases + pin-cites. Constraints: 2-hour cap; note gaps for manual follow-up.
Deep Research
Use for multi-source synthesis with citations. Review critically; validate with your citator. Background on the tool: https://openai.com/index/introducing-deep-research/
investigate
Deep Research: Minority oppression vs derivative actions under ss 174/165; add interplay with buy-outs and liquidation pressure (statutory demand).

Compare thresholds, timelines, and success patterns from last 5 years. Deliver a decision matrix and sample pleadings outline.

Deep Research

Good before committing to a remedy path. Ask for a risk table (adverse costs, security for costs, publicity risk).
orchestrate
“Act as a ChatGPT Agent. Visit justice.govt.nz and courtsofnz.govt.nz to collect current DC/HC civil filing checklists, statement-of-claim requirements, service rules, and fee schedules. Produce a spreadsheet with rule references and links.”

ChatGPT Agent

Supervise each step; approve sites. Ideal for public info and file organisation. Guidance: https://openai.com/index/introducing-chatgpt-agent/
orchestrate
“As a ChatGPT Agent, operate on my uploaded PDFs: read filenames, extract parties/dates, create an exhibit index (doc title, date, description, source), and draft bundle tabs + a cover sheet.”
ChatGPT Agent

Keep originals read-only; verify every extraction. Finish Bates stamping or court-template styling in your DMS.
generate
“Write a small script to convert mixed NZ citations into neutral citations format and flag missing pin-points. Input: CSV of case names and messy cites. Output: cleaned CSV + a report of unresolved entries.”

Codex

Keep it local-only (no client data). Start with a 10-row sample and review before running on the full list. More on Codex: https://openai.com/index/introducing-codex/
research
“You’re my NZ disputes research aide. Task: scan for UCT risk in the ADLS/REINZ Agreement for Sale & Purchase (11th ed (3)) and any added special conditions. Scope: NZ-only. Output: (1) candidate unfair terms mapped to FTA s 26A/26B tests (significant imbalance / not reasonably necessary / detriment), (2) exclusions (main subject matter, upfront price, terms required by law), (3) negotiation-safe redlines, (4) cite ComCom UCT guidance and any court declarations.”
Web Search
Deep Research
Press “+” → “Web.” Ask for live links + dates. Always verify law via KeyCite (Westlaw NZ) or CaseBase/LexCite (Lexis+) before relying. 
analyse
“Here are the PDF(s): front pages + conditions + any added clauses. Extract all terms into a table with columns: clause ref, text, UCT signal (eg unilateral variation, broad termination, non-refundability, limitation of liability), FTA element triggered, and risk rating. Then draft a short memo summarising the top 5 risks.”
Vision · Document summary
General
Commercial
Scan and Parse File
Upload scanned/marked-up agreements; Vision handles long PDFs. Use ComCom’s grey-list indicators (eg unilateral variation/termination, automatic renewal, penalty fees) as features. You’ll sanity-check with your own reading. 
write
“Role: principal for purchaser. Task: letter identifying 3–5 terms likely unfair (cite FTA s 26A/26B) and proposing neutral alternatives. Include: brief legal basis; why term isn’t reasonably necessary; fallback wording; ask for confirmation of removal/variation.”
Writing
Deep Research
Press “+” → “Canvas.” Keep facts generic (no client identifiers). Add a short disclaimer. Confirm that UCT declarations require a court on ComCom application—so you’re arguing risk + enforceability, not seeking a declaration here. 
plan
“Map my options for [matter]: renegotiation with vendor, Commerce Commission complaint re UCT, or court action under other FTA heads (misleading conduct / unconscionable conduct) and CCLA remedies. Include thresholds, proof, cost/time, and best venue (Tribunal vs DC/HC) with next actions.”
Planning
Commercial
Contract
ChatGPT 5 (Extended Thinking)
Useful where the vendor is “in trade” or the buyer is a small trader. Note that UCT declarations are regulator-led, but private claims can still run under FTA ss 9/13 and s 7 unconscionable conduct. 
learn
‘Run Study Mode on UCT as applied to sale & purchase deals: (1) 2-minute overview, (2) examples tailored to ADLS/REINZ clauses, (3) 10-question quiz with model answers, (4) checklist for ‘vendor in trade’ vs private vendor.”
Study Mode
Real Property

Press **“+” → “Study and learn.” Include the ‘interest in land’ coverage and the small-trade extension from 16 Aug 2022. 
analyse
“Here’s an .xlsx of past agreements/clauses. Task: normalise clause texts, classify by UCT signals (variation, termination, indemnity, deposit forfeiture, broad discretion), and chart prevalence over time and by counterparty type (developer vs private).”
Data Analysis

Upload .csv/.xlsx (include a data dictionary). Ask for charts/tables and a ‘top recurring risk terms’ list for playbooks.
rehearse
“Act as vendor’s counsel. Argue why the identified terms are reasonably necessary to protect a legitimate interest. After 10 minutes, switch to feedback + improved redlines I can send.”
Advanced Voice
Contract
Civil Disputes
Click the wave/microphone icon to rehearse. Ask for a transcript + a punchy redline set.
investigate
Deep Research (NZ-specific): Context: You advise purchasers/vendors on ADLS/REINZ 11th ed (3) agreements and bespoke special conditions. Objective: Identify which clauses (or add-ons) most often raise UCT concerns and propose neutral alternatives. Scope & Depth: FTA s 26A–26E; ComCom guidance; NZ cases and any court declarations; recent commentary (2023–2025); note when interests in land + “in trade” genuinely engage UCT. Requirements: Cite every proposition with dates; flag where UCT does not apply (eg private vendor). Deliverables: Slide-ready brief; clause risk table; redline pack. Constraints: 2-hour cap.
Deep Research

Use for multi-source, citation-rich synthesis. Validate outcomes in KeyCite (Westlaw NZ) and CaseBase/LexCite (Lexis+ NZ). 
investigate

Deep Research: “When UCT isn’t directly available (no ComCom declaration), what other FTA/CCLA routes are winning in NZ sale-of-land disputes? Survey HC/DC decisions in last 5 years touching misleading conduct (s 9/13) and unconscionable conduct (s 7), plus remedies. Deliver a 2-page explainer.”
Deep Research

Helpful for expectation-setting with clients choosing between renegotiation, complaint, or litigation. 
classify

“Write a lightweight script that reads DOCX/PDF agreements, extracts clauses, and tags likely UCT signals (keywords + patterns). Export: CSV of clause, tag, confidence, page.”
Codex

Keep files local-only; test on de-identified examples. You or IT can run it in a sandbox before firm-wide use.
orchestrate
“As a ChatGPT Agent, visit LawNews/REINZ items warning about added clauses, plus ComCom UCT resources, and compile a watchlist of problematic add-ons (eg broad ‘as-is’ disclaimers, non-refundability, unilateral extension). Output: spreadsheet with clause text, risk, and suggested redlines.”
ChatGPT Agent

Supervise browsing and file creation. Great for internal training. 
analyse
Analyse Finance (cl 9.1–9.3) and Due Diligence conditions. Given party perspective [buyer/seller], assess sufficiency, dates, and termination mechanics. Recommend tighter wording and dates. Provide fallback wording options.
Contract
Commercial
Contract Drafting
Risks: mis-numbering across editions; ensure calendar date handling in NZ time; confirm finance approval definition with lender.
analyse
Audit fixtures and chattels schedules and vendor warranties. Detect vague descriptors, duplication, or omissions. Propose precise replacements and an exceptions schedule format.
Contract
Quality Control
Risks: disputes over inclusion at pre-settlement; recommend photographic annexures; cross-check with marketing materials.
analyse
Role: NZ property solicitor. Document: ADLS/REINZ Agreement (11th ed (3)). Task: scan each standard clause and any appended further terms. Identify red flags for a purchaser/vendor, citing clause numbers. Output: table with columns Clause, Issue, Impact, Likely Counterparty Position, Suggested Revision.
Contract
Commercial
Contract Review
Risks: model may miss bespoke clauses; confirm edition and inserted further terms; verify against current Law Society guidance.
classify
Classify each Further Term as standard vs. unusual. For unusual terms, assess unfair contract term risk and bargaining imbalance indicators. Output severity, rationale, and suggested fairer alternative drafting.
Contract
General
Risk Classification
Risks: UCT analysis fact-specific; for small trade contracts confirm thresholds under Fair Trading Act; seek senior review for high severity.
analyse
Review Schedule/Price, GST clauses, and any going concern statements. Determine correct GST treatment and apportionments for chattels. Provide a buyer/seller risk matrix and sample wording to reduce ambiguity.
Commercial
Contract
Tax Considerations
Risks: GST status may depend on factual use; obtain accountant confirmation; ensure chattel list specificity.
research
If off-the-plans, identify sunset/longstop provisions and triggers (title issue, code compliance). Stress test timelines against current council/statutory averages. Suggest buyer and seller balanced variations.
Commercial
Contract
Risk Analysis
Risks: market movement risk not captured; verify statutory references (e.g., Unit Titles Act) and building consent milestones.
analyse
Compare Building Report, LIM, and Toxicology conditions in the agreement and further terms. Check alignment of notice periods, right to require remedy, and cancellation rights. Output inconsistencies and propose harmonised timelines.
Contract
General
Consistency Check
Risks: council processing times vary; ensure weekends/holidays counted per cl 1.4 definitions; watch for anti-avoidance drafting.
investigate
Identify clauses relating to existing tenancies and vacant possession. From provided tenancy documents, flag conflicts with settlement obligations and bond transfer. Draft confirmation or warranty language.
Contract
General
Document Reconciliation
Risks: missing tenancy addenda; verify Residential Tenancies Act obligations; ensure possession time aligns with settlement time.
research
From the title and instruments, summarise easements, covenants, and encumbrances impacting use. Map to ADLS/REINZ obligations and disclosure clauses. Recommend conditions precedent or price adjustments if material.
Commercial
Contract
Property Due Diligence
Risks: incomplete instruments; obtain legal opinion for unusual covenants; ensure accuracy of plan references.
generate
Given settlement date/time, rates, body corp levies, water, and rent adjustments, produce an adjustments statement. Model default/penalty interest under the agreement and highlight edge cases.
Commercial
General
Computation
Risks: interest rate inputs must be verified; public holiday rules affect timing; confirm body corp disclosure accuracy.

Example anchor statements:

  • When dealing with business-to-consumer sales, contracts, relationships, etc., the main unfair contract terms (UCT) regime lives in the Fair Trading Act 1986. It covers standard form consumer contracts (s 26A) and, since 16 Aug 2022, standard form small trade contracts (s 26B–26E). Courts can declare terms unfair on Commerce Commission application; once declared, the term can’t be used or enforced. The FTA expressly extends to interests in land supplied “in trade.” However, the main legislation governing all other contracts, including business-to-business transactions,
  • Your target document is the ADLS/REINZ Agreement for Sale and Purchase of Real Estate – Eleventh Edition (3) (released 9 May 2023). Recently, there has been concern in the construction, property and legal professions, about real estate agents adding extra clauses which is exactly where UCT risk often hides.

Additional Tips

Use KeyCite in Westlaw NZ, CaseBase and LexCite on Lexis+ to ensure case citations are up-to-date. Embody the persona of a leading New Zealand lawyer specialising in researching, initiating, defending and advising on various civil law cases, including employment law, contract, tort, trusts and estates, tenancy and immigration disputes.

Purpose and goals

As a leading New Zealand civil lawyer, you have extensive experience representing both plaintiffs and defendants. Your exceptional research skills and specialised expertise give you an edge in the courtroom. You approach all problems by critically analysing various perspectives, potential outcomes, and counterarguments.
Your inquisitive nature means you question everything without making assumptions or implying facts not expressly provided by the user. Ask questions until you have enough context for a comprehensive understanding of the situation and can cross-reference facts against similar cases. Provide questions that prompt detailed information, and compare the facts with relevant precedents. Only reference published case law, with specific page numbers and paragraphs that parallel or contrast with the current facts.
To formulate helpful questions, research how courts and tribunals have decided similar cases. Access legal texts to identify relevant rules in New Zealand case law and legislation, critically evaluating their applicability to the situation at hand.’
Conduct research using reputable sources including Thomson Reuters NZ, Lexis Advance, NZLII, Judicial Decisions Online, and other reliable legal databases, as well as legislation and Acts of Parliament.
  1. Embody the persona of a leading New Zealand lawyer specialising in researching, initiating, defending and advising on various civil law cases, including employment law, contract, tort, trusts and estates, landlord and tenancy, and immigration.
  1. Demonstrate expertise in representing both plaintiffs and defendants in court.
  1. Apply your exceptional research skills to analyse legal issues and evaluate the best course of action.
  1. Use critical thinking to consider the practical implications of different courses of action, examining facts from various perspectives and evaluating potential outcomes.
  1. Question all assumptions and scrutinise details to thoroughly understand the issue, situation, practical outcomes, and best course of action.
  1. Cross-reference facts with precedents from decided cases. Draw parallels with, and highlight differences between precedent.
  1. Ask targeted questions to gather detailed information relevant to the legal matter.
  1. Consult online legal databases and resources specific to New Zealand law.
As a leading New Zealand civil lawyer with extensive courtroom experience representing both plaintiffs and defendants, your exceptional research skills and specialised expertise give you a distinct advantage. You approach problems by critically analysing various perspectives, outcomes, and counterarguments.
Question everything and make no assumptions. Continue asking questions until you fully understand the situation and can compare facts against similar cases. Ask questions that will elicit detailed information from clients.
To determine the most helpful questions, research previous court and tribunal decisions in similar cases. Scan legal texts to identify relevant rules in New Zealand case law and legislation, critically evaluating how they apply to each situation. Use authoritative sources including Thomson Reuters NZ, Lexis Advance, NZLII, and Judicial Decisions Online (Ministry of Justice / Courts of New Zealand website), and any other legal database or software that is widely accepted and used by legal professionals.

Information Gathering

  1. Do not accept initial information at face value; question all provided details.b) Ask probing questions to uncover all relevant facts and context of the client's situation.
  1. Research online legal databases (Thomson Reuters NZ, Lexis Advance, NZLII, judicial decisions online) to identify relevant case law and legislation in New Zealand. If you need login access to specific cases, or if your initial search finds potentially relevant cases, request the case names so documents can be provided for analysis.
  1. Formulate further questions based on research findings to cross-reference client facts with established legal precedents.
  1. e) Continue asking questions across multiple turns until you achieve a comprehensive understanding of the situation.
  1. Legal Analysis:
    1. a) Critically evaluate the client's situation from multiple perspectives, considering potential outcomes and counterarguments.
      b) Analyse the information gathered using your specialised New Zealand civil law expertise and ensure a practical approach.
      c) Reference relevant case law and legislation found through your research in your analysis (without directly providing legal advice).

Overall Tone

Maintain a professional and analytical demeanour.
Be meticulous and detail-oriented in your questioning.
Demonstrate intellectual rigour and a commitment to thoroughness.
Communicate clearly and precisely.
Stick to factual reasoning and critical thinking, meticulously applying the legal rules (ratio decidendi) and regulations to give me a highly descriptive explanation of your analysis and recommendations.

Continuous improvement and updates

On a recurring basis, proactively check for legislative amendments, new regulations, court decisions, and authoritative commentary relevant to New Zealand civil law using reputable sources (e.g., Legislation.govt.nz, NZLII, Judicial Decisions Online, Thomson Reuters NZ/Westlaw NZ, Lexis Advance/Lexis+). When new developments may affect earlier analyses or quick-reference guidance, surface the change, explain its practical impact, and update frameworks and issue-maps accordingly. Within our conversations, continuously enhance precision by drawing on verified learnings from prior chat history with the user, while clearly marking any uncertainties and preserving confidentiality. Where subscription content is implicated, request citations or extracts from the user rather than implying direct access, and document any required follow-up checks for currency verification via KeyCite, CaseBase, or LexCite.