Every mid-market enterprise in New Zealand wants AI capability. Almost none of them can hire dedicated AI engineers at current market rates. The good news is that hiring a full AI team is not the only path, and for most organisations at this stage, it is not the right one. Here is how to build AI talent in a market where there is not enough talent to go around.
The NZ Talent Reality
New Zealand produces roughly 2,000 computer science graduates a year. Of those, a fraction specialise in AI and machine learning. Of that fraction, a further fraction want to work in enterprise (versus academia or the startup ecosystem). Of that further fraction, a portion leave for higher salaries in Australia, the UK, or remote roles with US companies.
The result is a genuine scarcity. Senior AI engineers with enterprise experience in New Zealand can be counted in the low hundreds. The demand from government, corporates, and growing AI companies far exceeds supply.
340%
increase in NZ job postings mentioning AI skills between 2023 and 2025
Source: Seek NZ, AI Skills Report, Q2 2025
This is not a problem you can solve by offering more money. Even if you win the bidding war, you are hiring one or two people into an organisation that has no AI infrastructure, no AI culture, and no support structure for AI work. Isolated AI hires in non-AI organisations have a high attrition rate because the work is not sustainable.
The Four Paths
Path 1: Partner
Best for: Organisations starting their AI journey, or building their first two to three AI capabilities.
Work with an external AI partner for your initial capabilities. They bring the expertise, the tooling, and the lessons from previous deployments. Your internal team focuses on domain knowledge, change management, and learning.
Advantages: Immediate capability. No hiring risk. Knowledge transfer builds internal skill over time.
Risks: Dependency if knowledge transfer is not managed. Cost per capability is higher than internal delivery at scale.
When to use: Your first twelve to eighteen months of AI work. This is the fastest path to production AI, and the partner's experience prevents the costly mistakes that first-time AI teams inevitably make.
Path 2: Train
Best for: Organisations with strong existing technical teams who want to add AI capability.
Upskill your existing engineers, data analysts, and domain experts. AI skills are more accessible than they were two years ago. A competent software engineer can become a competent AI engineer in six to twelve months with the right training and project experience.
Advantages: Leverages existing domain knowledge. Builds capability that compounds. Lower cost than external hiring.
Risks: Takes time. Requires investment in training and mentoring. Not all technical professionals want to (or can) transition to AI work.
What to train: Prompt engineering and evaluation (weeks). RAG architecture and implementation (months). Model integration patterns (months). AI operations and monitoring (months). The progression is incremental. You do not need to master everything before delivering value.
Path 3: Hire
Best for: Organisations with established AI programmes that need to scale.
Hire dedicated AI professionals once you have enough AI work to sustain them. One AI engineer in an organisation with no AI programme is a recipe for frustration. Two AI engineers in an organisation with three production AI capabilities and a clear roadmap is a recipe for growth.
Advantages: Dedicated capability. Deep expertise. Long-term cost efficiency at scale.
Risks: Expensive in the current market. High attrition risk if the role is isolated. Slow to hire (three to six months for a senior AI engineer in NZ).
What to look for: Breadth over depth. Enterprise AI engineers need to work across the full stack: prompts, retrieval, integration, evaluation, operations. A narrow specialist in model fine-tuning is less useful than a broad practitioner who can build, deploy, and maintain end-to-end capabilities.
Path 4: Fractional
Best for: Organisations that need strategic AI guidance without full-time commitment.
Engage a fractional AI lead or advisor. Someone who works with your team one to two days a week, providing direction, reviewing architecture decisions, and mentoring your team. They bring experience from multiple organisations without the cost of a full-time senior hire.
Advantages: Senior expertise at a fraction of the cost. Breadth of experience from working across organisations. Flexible commitment.
Risks: Limited availability. Not embedded enough for deep technical work. Works best alongside internal capability, not as a replacement.
The Recommended Sequence
For most NZ mid-market enterprises:
Phase 1 (months 1-12): Partner + Train. Work with an AI partner for your first capabilities. Simultaneously train your existing team on AI fundamentals. Insist on knowledge transfer as part of every partner engagement.
Phase 2 (months 12-24): Train + Fractional. Your internal team takes on more AI work, supported by a fractional AI advisor. The partner engagement reduces to specific, complex capabilities or architectural guidance.
Phase 3 (months 24+): Hire + Train. With established AI work and a capable internal team, hire dedicated AI professionals. They join a functioning programme, not an empty one. Attrition risk is low because the work is real and supported.
What Most Organisations Get Wrong
Hiring too early. The impulse is to hire an AI engineer as the first step. But an AI engineer without AI infrastructure, AI culture, and AI work to do is an expensive, unhappy employee.
Underinvesting in training. Training existing staff is slower than hiring, but it produces more durable capability. Your domain experts who add AI skills are more valuable than AI experts who lack domain knowledge.
Not insisting on knowledge transfer. Every external engagement should leave your team more capable than when it started. If your partner delivers a black box that only they can maintain, you have bought a product, not built a capability.
Treating AI talent as purely technical. The best AI professionals we have seen combine technical skills with communication, business understanding, and user empathy. Hiring purely for technical AI credentials misses the people who actually deliver enterprise value.
The NZ Advantage
New Zealand's constraints are also advantages. Our small market means closer relationships between organisations and talent. Our strong community means knowledge sharing happens naturally. Our pragmatism means less hype and more focus on what works.
The AI talent challenge is real, but it is navigable. The organisations that build AI capability systematically, through a combination of partnering, training, and eventually hiring, will have a durable advantage over those that try to buy their way in.