The most common question we get from NZ accounting firms considering AI training: “What are other firms doing?” It’s the right question. Here’s an honest answer based on what’s actually happening in the NZ accounting market in 2026.
The Current State: A Split Market
NZ accounting firms in 2026 fall into three groups:
- Early movers (~15%) — firms that have run structured AI training, built templates and workflows, and have AI as a normal part of practice. These firms are already operating at a structural productivity advantage.
- Active experimenters (~40%) — firms where one or two individuals use AI enthusiastically but there’s no systematic adoption. Uneven results, no shared learning, and the productivity gains are individual rather than firm-wide.
- Wait-and-see (~45%) — firms that are aware of AI but haven’t moved substantively. Often waiting for clearer guidance from CAANZ, for “the right tool,” or for the market to settle. The risk is that by the time they move, the gap with early movers will be significant.
What the Early Movers Are Doing Differently
The accounting firms capturing genuine value from AI share three practices that the wait-and-see firms don’t have:
1. They defined specific use cases before training
Early movers didn’t tell their team to “use AI more.” They identified the three to five highest-volume, most time-consuming writing tasks in the firm — management accounts commentary, client letters, research memos, meeting follow-ups — and built training around those specific tasks. Staff knew exactly what they were supposed to do differently after day one.
2. They built a shared prompt library
Individual AI users build their own prompts through trial and error. Firms build a shared library. The difference is compounding: every partner’s discovery benefits every staff member. Early mover firms typically have 20–50 reusable prompts covering their most common tasks, developed over 3–6 months of systematic use.
3. Leadership used it visibly
In firms where AI adoption has succeeded, partners and senior managers used AI visibly and talked about it. In firms where it stalled, it was positioned as a tool for junior staff. AI adoption follows status — if senior people aren’t using it, junior people don’t prioritise it.
The Use Cases Delivering the Most Value
Based on what’s working in NZ accounting firms:
Highest ROI (use these first)
- Management accounts commentary — drafting narrative commentary for monthly management packs. Average time saving: 45–90 minutes per reporting period per entity.
- Client letters and correspondence — advice letters, engagement letters, and ongoing client communications. Average time saving: 15–30 minutes per letter.
- Annual report narrative — director’s report sections, financial highlights, and performance commentary.
Good ROI (add in month 2–3)
- Technical research summaries — IRD rulings, binding rulings, and accounting standards interpreted for client application (always with partner review).
- Meeting preparation and summaries — pre-meeting research on clients and post-meeting action summaries.
- Grant and R&D tax incentive applications — first draft narrative for Callaghan Innovation R&D applications.
Proceed carefully (get protocols right first)
- AML/CFT documentation — AI can help structure procedures, but compliance requirements mean careful human review.
- Tax advice letters — AI can draft structure, but technical accuracy must be fully verified.
Professional Obligations: The Practical Picture
CAANZ has not issued formal AI guidance as of early 2026, but the existing Code of Ethics applies:
- Competence: Members must only perform services they are competent to perform. Using AI doesn’t lower this standard — it means you must be competent to supervise AI output, which is different from writing everything yourself.
- Confidentiality: Client financial information cannot be entered into public AI tools without appropriate data agreements. The practical rule: no client names, specific revenue/profit figures, or identifying transaction details in ChatGPT or Claude.ai.
- Due care: AI drafts require the same review as any other work product. “The AI wrote it” is not a defence.
Enterprise AI options (ChatGPT Enterprise, Claude for Business, Microsoft Copilot within your Microsoft 365 tenancy) offer stronger data handling for practices working with sensitive client information.
A Practical 60-Day Plan for Accounting Firms
- Days 1–7: Identify your top three use cases. Where does your team spend the most time writing? That’s where you start.
- Days 8–14: Half-day training session — all client-facing staff. Hands-on, use-case-specific, with the privacy protocols your firm needs.
- Days 15–30: Use AI for the defined use cases. Document what’s working. Start building the shared prompt library.
- Days 31–60: Expand to two more use cases. Share the prompt library firm-wide. Identify a champion in each team to drive ongoing adoption.
GenAI Training NZ delivers AI training tailored to NZ accounting firms — covering CAANZ professional obligations, data privacy for client information, and the specific use cases that deliver fastest in accounting practice. Get in touch to discuss a session for your firm.
Also see: AI for Accountants Christchurch | AI for Accountants Auckland | AI for Accountants Wellington | Context Engineering
