Mortgage broking in New Zealand is a relationship business built on trust — and it comes with significant compliance obligations under the Financial Advisers Act 2008 (now the Financial Markets Conduct Act) and the Credit Contracts and Consumer Finance Act. Between client fact-finds, lender research, loan structuring, disclosure documents, and client reports, there’s a substantial amount of writing involved for every deal.
AI doesn’t change the advice — it changes how quickly and professionally you can document and communicate it.
Where AI Adds Real Value for Mortgage Brokers
1. Client Advice Letters and SOA Drafting
Statements of Advice (SOA) and loan recommendation letters follow predictable structures — client situation, objectives, recommendation, rationale, disclosure. AI can help you draft these documents significantly faster once you’ve gathered the client’s information.
The approach: describe the client’s situation in general terms (income, debt, property goal, risk profile), describe the recommended structure, and ask AI to draft the recommendation section. You verify compliance, check numbers, add lender-specific details, and finalise. The substance is yours; AI handles the writing.
2. Lender Research and Comparison
With dozens of lenders, changing rate cards, varying serviceability calculators, and different policy nuances (LVR limits, self-employed treatment, income types), keeping up with the market is a job in itself. AI can help you research and compare options faster — summarising information, identifying edge cases in lender policy, and helping you explain the differences to clients in plain language.
Important: Always verify rate and policy information directly with lenders or your aggregator. AI knowledge can be outdated, and rate card information changes frequently.
3. Client Communications
Pre-approval updates, conditional approval explanations, settlement countdown emails, requests for outstanding documents, rate review reminders — most mortgage brokers send dozens of similar emails every week. AI can help you build a polished template library that goes out faster and reads more professionally.
AI also helps with difficult conversations — explaining a declined application, discussing a credit issue, or outlining why the client’s borrowing capacity is less than they expected. These messages require care; AI can help you strike the right tone.
4. Compliance Documentation
Financial advisers under the FMCA regime need to maintain records of advice, client consent, disclosure, and conflicts of interest. AI can help draft and maintain internal compliance documents, update your engagement letters when regulations change, and build checklists for your compliance review processes.
If you’re creating or updating an AI use policy for your practice, AI can help draft that too — specifying which tools your team can use, what client data must never be shared with AI systems, and how AI-assisted advice must be reviewed before it reaches clients.
5. Marketing and Referral Content
Educational content — “what does a mortgage broker actually do?”, “how much deposit do you really need in NZ?”, “fixed vs. floating: what’s right for you in 2026?” — builds trust with prospective clients and drives organic search traffic. AI can help you produce this content consistently without it taking hours every week.
Social media posts, email newsletters, and referral partner updates can all be drafted with AI, edited to your voice, and sent. The goal is to stay visible to your network without the content creation overhead stopping you from doing it.
6. Property Market Research
Clients frequently ask for your view on the market — interest rate direction, property value trends, what’s happening in specific suburbs. AI can help you quickly synthesise publicly available information into a coherent briefing that positions you as informed and professional. Never present AI-generated market commentary as guaranteed fact — present it as your informed perspective based on current data.
Privacy and Regulatory Considerations
Mortgage brokers handle some of the most sensitive personal financial data that exists: income, assets, debts, credit history, employment, relationship status. Under the NZ Privacy Act 2020 and FMCA obligations, this information must be protected and used only for the purpose the client consented to.
- Never paste real client financial data into consumer AI tools (ChatGPT, Claude.ai, Gemini). Income figures, credit scores, liability schedules, and property addresses are sensitive personal information under the Privacy Act.
- Use anonymised scenarios when drafting — “a couple earning $180K combined, with a $50K deposit, purchasing their first home in Hamilton” — then add the client’s name to the finalised document yourself.
- Check your CRM and advice software for any built-in AI features and their data handling policies. Where is the data processed? Who has access?
- Your FAP (Financial Advice Provider) obligations don’t change based on how you produced a document — you remain responsible for the accuracy and suitability of every piece of advice you give.
For brokers handling high volumes of client data, a local AI setup keeps all information on your own hardware, eliminating the risk of client financial data leaving your control.
The Compliance Line: What AI Must Not Do
AI cannot provide financial advice in the regulated sense — it cannot assess a client’s actual situation, recommend a specific product, or take responsibility for an advice outcome. Every piece of client-facing advice content must be reviewed and signed off by a licensed adviser before it goes to the client.
AI is your drafting assistant and research accelerator — not your co-adviser.
Getting Started
The fastest win: pick your three most common client email types (conditional approval explanation, document request, rate review) and use AI to build professional templates this week. Send the next five through those templates. See how much time you save and whether the quality is better.
For a broader programme — training your team on AI workflows, building a compliant AI use policy for your FAP, and identifying where AI can improve your deal conversion rate — an AI Assessment provides a clear, actionable roadmap. We work with financial services businesses across New Zealand on AI adoption that meets your compliance obligations.
Frequently Asked Questions
Can I use AI to write my Statements of Advice?
AI can draft the structure and narrative sections of an SOA significantly faster than writing from scratch. However, the financial figures, product recommendations, and compliance disclosures must be completed and verified by the licensed adviser. AI-assisted SOA drafting is fine; AI-only SOA generation is not.
What does the FMA say about AI in financial advice?
As of 2026, the FMA hasn’t issued specific guidance on AI in financial advice workflows, but the existing conduct obligations — act in the client’s interest, ensure advice is suitable, maintain accurate records — apply regardless of what tools you use. The FMA has signalled it will issue guidance on AI in financial services; watch their website for updates.
Is it safe to use ChatGPT for mortgage research?
For generic research (explaining how LVR works, comparing fixed and floating concepts, drafting educational content) — yes, with care. For client-specific advice or anything involving real client data — no. And always verify any specific rate or policy information against current lender sources; ChatGPT’s training data has a cutoff and doesn’t know today’s rate cards.
My aggregator has AI tools built in — should I use them?
Check what data they access and where it’s processed. Your aggregator’s AI tools likely have access to your client data through the CRM — understand their privacy policy, data retention, and whether the tool is being used to train AI models. Ask directly; your aggregator relationship manager should be able to answer.
