REA’s 2026 mandatory verifiable CPD topic is not a soft skills elective. It is one hour on the use of digital technology in real estate transactions — covering risks and issues with digital tools, professional and legal obligations, AI, digital marketing, record-keeping, data privacy and confidentiality, and principles for using digital technologies.
That hour sits inside the usual annual structure: 10 hours of verifiable training (1 mandatory + 9 elective) plus 10 hours of non-verifiable sector-related training, all due by 31 December. Primary source: REA — Topics for verifiable CPD.
Approved providers, including REINZ’s online delivery of the mandatory topic, spell out learning outcomes that should feel familiar to anyone who has watched consumer AI tools walk into a listing lounge: Generative AI principles; privacy and confidentiality when information goes into open AI tools; digital record-keeping (including the risks of deleting or editing messages); agency policies; legal and professional obligations; and risks across AI, digital signatures, messaging apps, marketing, and online appraisal tools.
The CPD hour is the floor. It is not your agency policy. This piece is for agency leaders and licensees who need a Monday plan after the course — not another slide deck. It is not legal advice.
PropertyLM take
Completing the mandatory hour proves you sat the topic. Running governed workflows, evidenced outputs and human sign-off proves you understood it.
The problem: tools everywhere, controls uneven
Licensees already use digital technology every day. REA said as much when it set the 2026 mandatory topic. The practical problem inside agencies is unevenness: one salesperson drafts captions in a consumer chatbot; another pastes a LIM “just to summarise”; a third uses an online appraisal widget as if it were Rule 10.2; a branch has no written Gen AI policy; a supervisor cannot say which tools are approved.
That unevenness is exactly where complaints and privacy incidents grow. REA’s Generative AI guidance is blunt on accountability: you remain responsible for the real estate services you provide. An error made by a Gen AI tool is not a defence. Agencies and supervisors can also be pulled in when salespeople under their watch use tools badly. Careful human oversight is not optional colour — it is how you keep meeting the Act and the Code while the tooling changes.
The guidance maps existing duties onto new habits. Rule 5.1 still requires skill, care, competence and diligence. Disclosure and accuracy duties still apply to marketing and property information. Confidentiality and personal information still engage rule 9.17 and the Privacy Act. Appraisals still have to be in writing, reflect current conditions, and be supported by comparable sales under rules 10.2 and 10.3. Digital speed does not rewrite those limbs.
REA in plain NZ English
Gen AI creates; it does not understand. Treat outputs as drafts until a licensed human has checked accuracy, relevance and completeness — especially before anything touches a vendor, buyer or listing file.
What the mandatory hour is trying to install
Read the REINZ / Open Polytechnic-style learning outcomes as an operating checklist, not a quiz bank:
Know principles for evolving tech, especially careful, informed use of Generative AI.
Understand digital marketing risks, including Gen AI in promotional materials and the ethics that travel with them.
Describe privacy and confidentiality issues — particularly putting confidential or personal information into open AI tools.
Understand digital record-keeping duties, including risks from deleting or editing messages or failing to keep proper documentation.
Describe your agency’s policies and procedures for digital technology.
Understand legal, ethical, regulatory and professional obligations when using digital tools.
Identify risks across AI, digital signatures, messaging apps, marketing, and online tools such as property appraisals.
Notice the recurring pair: individual competence and agency policy. The course expects you to know both. If your agency has no policy, the learning outcome still names the gap.
How PropertyLM shows up: governed workflows, evidenced outputs, human sign-off
PropertyLM is built for the gap between “we did the CPD hour” and “we can defend how tools are used on live files.”
Governed workflows mean research and drafting happen inside patterns that assume review, not anonymous paste-into-the-void habits. Newton is designed for live New Zealand property context in Claude and ChatGPT in plain English — the kind of assisted research that should still be checked, cited and owned by a licensee.
Evidenced outputs mean Atlas-style CMA and appraisal support that keeps comparable sales and condition narrative visible. Online appraisal tools are named in the CPD learning outcomes for a reason: automation can gather, but Rule 10.2 still wants a written licensee opinion that realistically reflects current conditions and is supported by comps. REA’s appraisals guidance remains the reference for that standard.
Human sign-off is the non-negotiable. REA’s Gen AI guidance says improper, negligent or incompetent use can breach skill and care duties; Gen AI can sound right while being wrong; fact-check accuracy, relevance and completeness before client-facing use. Name the reviewer. Keep a trail for material outputs.
Privacy-aware research is the Monday habit that prevents the worst prompt mistakes: do not paste LIM PDFs, vendor personal information, negotiation notes or trust-account detail into open consumer chat tools. REA flags that providers may see inputs and outputs; government guidance recommends against putting personal and client information into external AI tools; rule 9.17 and the Privacy Act still apply. If you would not email the file to a stranger, do not paste it into a prompt.
Technologist framing
Problem → digital tools and Gen AI are already in the office, controls are uneven. How we solve by doing → governed workflows, evidenced outputs, human sign-off, privacy-aware research. For who → agency leaders writing policy and licensees doing the work. Proof → a one-page policy, induction that names banned inputs, and a verification habit before anything client-facing leaves the building.
For who: leaders first, then every licensee
Agency leaders and supervisors should treat the mandatory topic as a systems prompt. REA encourages agencies to have Gen AI policies. The CPD outcomes explicitly ask licensees to describe agency policies — which only works if those policies exist, are short enough to follow, and are taught in induction. Supervisors who cannot inventory tools in use are supervising theatre, not practice.
Individual licensees still carry personal accountability. Follow agency policy. Do not outsource judgement to a model. Fact-check. Keep records. Know that digital signatures, messaging apps and marketing copy sit inside the same professional frame as a face-to-face appraisal conversation.
Electives still matter. Social media and digital communications, communication channels and documentation, appraisals-related agency agreement content, and complaint handling all sit next to the mandatory hour. The mandatory topic is the spine for 2026 digital risk — not the whole skeleton.
What to do Monday: policy, induction, verification
Inventory tools in use. Include “personal” ChatGPT / Claude / Gemini accounts touching work files. Name what is approved, what is banned, and who decides.
Write or refresh a one-page Gen AI / digital tech policy. Allowed tools. Banned inputs (LIM/vendor/personal/confidential). Review rules for client-facing outputs. Record-keeping expectations for messages and material drafts. Escalation path. Point people at REA’s Gen AI guidance PDF.
Put it in induction and in the next branch meeting. CPD completion is not induction. Make new starters recite the banned-input list before they get production logins.
Separate research aids from appraisal opinions. Online estimates and AI drafts can help gather. They are not the Rule 10.2 appraisal. Keep written appraisals current, comparable-supported, and licensee-owned. See REA — Appraisals.
Install a verification habit. Before send: accuracy, relevance, completeness, privacy check, named human reviewer. Especially for marketing claims, disclosure-adjacent summaries, and anything that looks like valuation language.
Protect the record. Messaging apps and editable drafts are where files go missing. The CPD outcomes call out deletion and editing risk for a reason. Keep recoverable trails for material client communications.
Book the hour, then build the system. Complete the mandatory verifiable topic with an approved provider. Then schedule the policy session the same week — otherwise the certificate becomes the whole intervention.
Proof: the course is completed; the practice is evidenced
Proof for regulators and clients is not a Thrive completion tick alone. Proof looks like: a written agency policy people can find; induction records; appraisal files with comps and dates; marketing copy that was reviewed; prompts that never contained the LIM; supervisors who know which tools are live; and licensees who can explain — in plain NZ English — why they did not paste vendor data into an open model.
REA’s Gen AI guidance and the 2026 mandatory CPD topic are aligned on the same spine: opportunity exists, accountability stays with the licensee, privacy and accuracy are first-order risks, and agencies need policies. PropertyLM’s product stance matches that spine — governed workflows, evidenced outputs, human sign-off, privacy-aware research — so digital speed serves diligence instead of replacing it.
Do the hour. Then do the Monday work. The certificate is necessary. The policy and the verification habit are what keep consumers — and your licence — safer when the next tool arrives.
Primary sources: REA CPD topics; REINZ mandatory digital technology topic; REA Generative AI guidance.
— PropertyLM.
