New Zealand lawyers are increasingly using AI tools — but most are still in the “cautious experimentation” phase rather than systematic adoption. The firms moving fastest aren’t using better tools; they’re using the same tools more deliberately.
Here’s where AI is delivering the most value in NZ legal practice in 2026, with specific tool recommendations for each use case.
The Legal AI Landscape in 2026
The tools available to NZ lawyers fall into two categories: general AI assistants (ChatGPT, Claude, Copilot) and legal-specific AI platforms (Harvey, Lexis AI, Thomson Reuters CoCounsel). The general tools are cheaper and surprisingly capable for most legal writing tasks. The legal-specific platforms have deeper integration with databases but higher cost and slower NZ adoption.
For most NZ law firms, the practical answer is: start with Claude or ChatGPT for writing and research assistance, and evaluate legal-specific tools once your team has basic AI fluency.
Use Case 1: Legal Research — What Works and What Doesn’t
What AI does well: Synthesising case summaries you provide, explaining legal concepts, identifying issues in a fact pattern, researching secondary sources and academic commentary, drafting research memos from cases you’ve identified.
What AI does poorly: Reliable case law retrieval. AI models have training cutoffs and do not reliably know recent NZ judgments. Never rely on AI to identify the leading case — it will hallucinate plausible-sounding citations.
The right workflow: Use Westlaw NZ, LexisNexis Pacific, or NZLII to find your cases. Then use AI to help synthesise and write. AI accelerates the analysis and drafting phase; traditional databases own the retrieval phase.
Use Case 2: Drafting — The Highest ROI Application
Document drafting is where AI delivers the clearest time savings for NZ lawyers. The model: AI produces a structured first draft from your instructions and notes; you apply legal judgment, review for accuracy, and refine.
Works well for:
- Client correspondence — advice letters, update emails, file closure letters
- Commercial contracts — first drafts of standard agreements from your brief
- Litigation documents — pleadings outlines, affidavit drafts from instructions, correspondence
- Property — standard form variations, correspondence to vendors/purchasers
- Employment — settlement agreements, employment agreements, correspondence
Recommended tool: Claude (Anthropic) for most drafting — consistently produces more careful, nuanced legal prose than ChatGPT. The 200K token context window also handles longer document review tasks better.
Essential caveat: AI drafts require careful review. AI does not know your specific client’s circumstances, the specific agreed commercial terms, or the jurisdiction-specific nuances that matter. It produces a starting point, not a finished product.
Use Case 3: Document Review and Analysis
AI can read a long document and answer questions about it — flagging unusual clauses, summarising key terms, identifying issues for consideration.
Effective applications:
- Due diligence — AI reviews disclosure documents and flags items requiring attention
- Contract review — summarising key obligations, unusual clauses, and missing standard provisions
- Lease review — identifying critical dates, rent review mechanisms, and unusual tenant obligations
- Long documents — extracting specific provisions from lengthy agreements quickly
Important: AI document review is a first-pass tool, not a substitute for thorough legal review. It catches what it catches — a trained lawyer still needs to review the full document for the things AI misses or mischaracterises.
Use Case 4: Client-Facing Communications
Legal writing is notoriously inaccessible to clients. AI can help translate — drafting client-facing explanations of complex legal concepts in plain English, with the lawyer reviewing for accuracy and appropriate qualification.
- Advice summaries — plain English summaries of complex advice for client comprehension
- FAQ documents — common client questions answered accessibly
- Status updates — regular client update emails drafted efficiently
- Website content — legal expertise explained accessibly for prospective clients
Professional Responsibility Considerations
The NZLS has not yet issued formal AI guidance, but the existing Rules of Conduct apply:
- Competence (Rule 3): Lawyers using AI must understand its limitations. Supervising AI output requires the same competence as supervising a junior lawyer’s work.
- Confidentiality (Rule 8): Client information must not be entered into public AI tools without client consent. This means no real client names, matter details, or identifying information in ChatGPT or Claude.ai.
- Supervision: Lawyers remain responsible for all work produced in their name, regardless of how it was drafted. AI-assisted work requires the same review as any other output.
The practical approach: use AI for drafting with anonymised facts and template language. Apply client-specific details during your review phase, not during the AI interaction.
Getting Started: A 4-Week Plan for NZ Lawyers
- Week 1: Sign up for Claude Pro (~$30 NZD/month). Use it only for client correspondence drafting — emails and update letters. Review every output carefully.
- Week 2: Add document review — upload a contract or disclosure document and ask Claude to summarise key terms. Compare with your own review.
- Week 3: Try a simple drafting task — a standard agreement or covering letter for a transaction you know well. Measure the time saving.
- Week 4: Build your context document — a prompt template that includes your practice area, your drafting preferences, and standard qualifications you always include. Reuse it.
GenAI Training NZ delivers AI training for NZ law firms. Half-day workshops covering practical AI use, professional responsibility, and workflow integration. Get in touch to discuss a session for your firm.
Also see: AI Training for Lawyers NZ | AI for Lawyers Christchurch | Context Engineering
