Advisory Standards
What to check before hiring an AI consultant in NZ
The number of people selling AI advice has grown fast. An experienced buyer needs a sharper test than a credentials page to protect their operations and budget. This guide covers the questions that separate useful advice from software reselling.
How are they paid?
Determine if their revenue comes from your consulting fee or a software vendor commission.
Who does the work?
Check whether the senior principal leads the engagement or delegates it to juniors.
Where is the data stored?
Verify data residency policies and local support availability.
Who owns the intellectual property?
Confirm that your business retains ownership of the configured models and workflows.
Proceed with advisers who charge for problem resolution rather than software subscriptions.
No hidden software commissions
Direct access to the senior practitioner
Handover and ownership terms documented
The Reseller Risk
Why independence matters more than technical knowledge
The most common mistake businesses make when hiring an AI consultant is paying for strategy and receiving a software pitch. Many operators in the advisory market take referral fees or recurring commissions from software vendors. This commercial reality changes the advice they give.
An adviser who earns a commission on a specific platform will naturally steer your business toward that platform. They are financially incentivised to find a software solution, even when a process fix, a checklist, or a feature within your existing accounting software would solve the problem faster. When regional businesses are looking for an AI consultant in Taranaki, local access often matters as much as technical capability. They need someone who understands the margin squeeze affecting local industries, not someone applying a global tech template to a regional reality.
The recommendation is always new software
A true consultant will frequently advise you to fix a process or use a tool you already pay for. A software reseller will always recommend buying a new licence.
The problem definition is rushed
A consultant wants to understand where your operations are failing. A reseller wants to match your symptom to their platform’s feature list as quickly as possible.
The ongoing costs are minimised during the pitch
Implementation is usually priced clearly, but the ongoing token costs, subscription tiers and maintenance requirements are often left vague until the business is locked in.
Questions to ask before you hire an AI consultant
A capable AI adviser will welcome these questions. A software reseller will usually provide vague answers or attempt to redirect the conversation.
Do you receive vendor commissions?
Ask directly if they receive any financial benefit, referral fee, or ongoing margin from the software they recommend. Complete independence is the only way to trust the advice.
Will you tell us if AI is the wrong answer?
A competent consultant will regularly advise against technology. If a problem stems from poor management or an undefined process, adding AI creates faster errors.
Who actually does the work?
Confirm whether the senior principal you speak with during the pitch will lead the implementation, or if the work gets delegated to junior staff immediately after signing.
Who owns the intellectual property?
Your business must own the specific prompts, workflows and data models built for you. Ensure the consultant cannot lock you out or resell your configurations to competitors.
Where is our data stored?
Data residency matters. Check whether information is processed locally or sent to offshore servers, and ensure the approach complies with the Privacy Act 2020.
What happens after the handover?
Determine who owns the system when a vendor updates their model and the outputs change. A clear maintenance boundary prevents unexpected breakages.
What a sensible first engagement looks like
A reliable AI consulting process starts small, proves value, and establishes boundaries before committing to a large build.
The problem definition
The engagement must begin by isolating the operational delay or cost. The consultant should map the real workflow, including the unlogged handovers and workarounds.
- Identify the affected process
- Map current user behaviour
- Document existing bottlenecks
- Calculate the cost of the problem
The existing tool audit
Before recommending new software, the consultant should check whether your current accounting, CRM, or Microsoft 365 platforms already contain unused features that solve the problem.
- Review current software licences
- Test native automation features
- Identify configuration gaps
- Maximise existing investments
The governance check
Implementing technology requires clear rules. The consultant must review data privacy, security settings and human oversight requirements before anything goes live.
- Review data classification
- Assess privacy implications
- Define human approval steps
- Establish clear accountability
The defined pilot
The first build should be a restricted pilot that tests a single use case. This proves the technology works in your specific operational environment before full rollout.
- Select a single measurable task
- Run a time-boxed test
- Measure the baseline performance
- Compare results against costs
Local Reality
Why New Zealand context matters
I have watched dozens of AI projects fail because the advice assumed a Silicon Valley operating model. New Zealand businesses operate under specific conditions. Our market is characterised by tight margins, small teams where staff wear multiple hats, and persistent labour shortages in skilled trades.
A consultant who understands this reality will prioritise administrative relief and margin protection over experimental technology. They will structure the AI strategy around what a twelve-person business can actually run and maintain. They will also follow established local frameworks, such as the MBIE Responsible AI Guidance for Businesses, to ensure compliance with New Zealand regulations.
The Accountability Gap
Who owns the problem when the system breaks?
Every technology deployment eventually encounters an issue. A vendor updates an API, a model changes its output format, or a staff member enters an unexpected data type. The true value of a consultant becomes obvious during these moments.
The handover document
An independent consultant provides comprehensive documentation detailing exactly how the system works, where the data flows, and who inside the business is responsible for maintaining it.
Avoiding vendor lock-in
Advice should maintain your operational flexibility. A good consultant designs systems so you can swap out underlying AI models if prices rise or better options emerge.
Training your internal team
The goal is to build capability inside your business. The consultant should train your staff to manage the new workflow, rather than forcing reliance on an ongoing maintenance retainer.
Direct access to the principal
When you work with a principal-led agency, the person who designed the strategy remains accountable for its operation. The advice is grounded in personal experience, not a generic playbook.
Questions
Hiring an AI Consultant
Common questions New Zealand business owners should ask to ensure they receive independent, valuable advice.
Why is it a problem if my consultant resells software?
A consultant who earns a commission from a software vendor is financially incentivised to recommend that specific product. This conflict of interest means they are less likely to suggest a process improvement or a free configuration of your existing tools.
What should a consultant do before recommending a tool?
A capable consultant maps your current process first. They must define the operational bottleneck, determine the root cause of the delay, and calculate the cost of the problem before suggesting any technology.
How can I tell if an adviser understands the New Zealand context?
They will frame their recommendations around local realities, such as small team sizes, margin pressures, and the Privacy Act 2020. They will focus on practical administrative relief rather than abstract global trends.
Should a consultant guarantee a specific financial return?
No honest consultant guarantees a return before completing a discovery phase. They should instead help you establish a baseline measurement, run a restricted pilot, and prove the return through actual performance data.
Who owns the intellectual property from an AI project?
Your business should retain full ownership of the data models, prompts, and specific workflows developed during the engagement. Ensure this is clearly stated in your contract.
What happens if AI is not the right solution for my business?
An independent consultant will tell you directly if AI is the wrong answer. They will advise you to fix the underlying process, improve team communication, or use simpler automation instead.
Need independent advice for your operations?
Start with a structured conversation about your specific business problem. We provide clear, objective advice without any vendor commissions or software reselling agendas.