AI in NZ Utilities: A Snapshot of Field Scheduling and Operational Transformation

Manual field scheduling is still a significant challenge for many New Zealand utilities. Across the sector, 30–50% of dispatched jobs require correction before work can begin, and another 10–15% stall mid-job due to planning oversights.

The result: crews spend hours troubleshooting administrative errors instead of delivering operational value, overtime costs increase, and service reliability suffers. Manual schedulers typically spend 4–7 hours weekly building and revising schedules that AI can generate in minutes.

AI is changing that but only when implemented with structured oversight and secure processes.

What Is Happening with NZ Utilities

New Zealand utilities face unique operational pressures that make transformation both urgent and complex:

• Reliability expectations: Outage response and restoration times are closely monitored, with customer expectations rising. Regulatory scrutiny has intensified following high-profile infrastructure failures, and utilities are under pressure to demonstrate continuous improvement in service levels.

• Aging infrastructure: Many assets are at or near end-of-life, increasing maintenance complexity. New Zealand's electricity demand is projected to grow 35–82% by 2050 (MBIE), while 30.7% of drinking water pipes are already in poor or very poor condition. This dual challenge - expanding capacity while maintaining deteriorating assets - strains operational resources.

• Workforce constraints: Limited crews and high field workloads make efficiency critical. With skilled trades shortages across the utilities industry and an aging workforce approaching retirement, organisations cannot simply hire their way out of inefficiency. Every hour lost to rework or poor scheduling compounds the problem.

• Manual processes: Scheduling errors, rework, and ad hoc adjustments reduce productivity. When schedulers lack real-time visibility into crew locations, skillsets, certifications, equipment availability, and job requirements, even well-intentioned decisions lead to costly mismatches. The downstream effect is failed first-time fix rates, extended outage durations, and frustrated field teams.

These challenges mean inefficiencies are not just inconvenient - they have measurable financial and operational consequences. Every failed dispatch costs twice: once in the field, once in overtime.

How AI Is Being Applied

Leading NZ utilities are adopting AI to tackle these challenges in practical ways:

• Smart Crew Allocation: AI matches crews with the right skills, certifications, and availability to each job. Rather than relying on schedulers' institutional knowledge or manual spreadsheets, AI evaluates hundreds of variables simultaneously - from required competencies and geographic proximity to crew fatigue levels and equipment compatibility. This ensures the right person arrives with the right tools, first time.

• Predictive Job Duration: Historical data helps estimate task times accurately, reducing rescheduling and overtime. By analysing thousands of completed jobs - factoring in job type, asset age, weather conditions, and crew experience - AI provides realistic time estimates. This prevents the common problem of schedulers underestimating complex jobs or overallocating resources to routine tasks.

• Route Optimisation: Dynamic routing reduces travel time and fuel costs, improving first-time job completion. AI continuously recalculates optimal routes based on real-time traffic, weather, and emerging priority jobs. What once required manual intervention now happens automatically, keeping crews productive and reducing vehicle operating costs.

• Proactive Resource Planning: AI anticipates part or equipment needs, ensuring teams are fully equipped before heading into the field. By cross-referencing job requirements with inventory data and historical parts usage, AI flags potential shortages before they cause delays. This seemingly simple capability eliminates one of the most frustrating causes of failed field visits.

Even early implementations are helping utilities recover hundreds of hours monthly - freeing crews to focus on critical operational tasks. Utilities using automated scheduling report crews spending 65% of their time on jobs, compared to just 44% with manual scheduling. That productivity gain translates directly to improved service levels and reduced operational costs.

Governance and Security Considerations

AI in critical infrastructure isn't just about efficiency - it introduces new responsibilities that utilities must address from day one:

• Data integrity: Scheduling relies on accurate crew, asset, and job data. If underlying systems contain outdated certifications, incorrect asset locations, or incomplete job histories, AI will amplify these errors at scale. Data quality must be addressed before - or during - AI implementation, not after problems emerge.

• Human oversight: Algorithms need validation, and operators must be able to override AI decisions when required. AI excels at pattern recognition and optimisation, but it cannot account for every contextual nuance - a crew member dealing with a family emergency, a road closure not yet in the system, or a customer relationship that requires special handling. Maintaining human judgment in the loop prevents automation from becoming inflexible.

• Secure integration: AI tools must connect safely with operational systems without expanding the attack surface. Cyberattacks on electrical substations jumped 70% in one year - a reminder that every new system connection is a potential vulnerability. AI platforms require secure authentication, encrypted data transmission, and network segmentation to protect critical infrastructure.

• Auditability: Regulatory and internal reporting require transparent, traceable decision-making. When an incident occurs, utilities must be able to explain why certain crews were dispatched, what information the AI used, and whether human oversight was applied. Black-box AI systems that cannot explain their reasoning are incompatible with regulated environments.

By integrating governance and security from the start, utilities ensure AI adoption strengthens operations without introducing new risks. This approach also builds stakeholder confidence - both internally with field teams and externally with regulators.

Real-World Proof: What's Working in New Zealand

NZ utilities industry operators are already demonstrating what's possible when automation is implemented thoughtfully.

When Spark New Zealand transformed its IoT capabilities - critical for utilities industry smart meters, sensors, and monitoring - they managed to onboard and maintain 2.3 million IoT connections with minimal increase in support staff. "The analytics and automation capabilities have been a game-changer," noted Matt McLay, IoT Sales Lead at Spark.

This proves that New Zealand organisations can scale operations dramatically through automation without proportionally scaling headcount - a critical advantage given workforce constraints across the utilities industry.

Lessons for Your Organisation

For utilities considering AI, early adoption can be practical and low-risk when approached systematically:

1. Map your problem areas: Identify scheduling bottlenecks and inefficiencies. Where do jobs fail most frequently? Which crew types are consistently overallocated or underutilised? What percentage of your scheduling time is spent on rework versus strategic planning? Quantifying these problems establishes your baseline and clarifies ROI expectations.

2. Assess data and system readiness: Ensure crew, asset, and operational data are accurate and accessible. Conduct a data quality audit: Are crew certifications current? Do asset records reflect actual conditions? Can job history data be extracted and analysed? Address gaps now to avoid undermining AI performance later.

3. Start with a controlled pilot: Test AI in a limited environment, monitor results, and refine processes. Choose a single crew type, geographic area, or service category where success can be clearly measured. Run AI-generated schedules in parallel with manual schedules initially, comparing outcomes before full deployment.

4. Embed governance and oversight: Define who owns AI decisions, maintain transparency, and secure integrations. Establish clear escalation paths for when AI recommendations seem incorrect. Document decision-making logic. Implement role-based access controls and audit logging from day one.

5. Track outcomes: Measure efficiency gains, overtime reduction, and reliability improvements to justify scaling. Track first-time fix rates, average job completion times, crew utilisation percentages, and customer satisfaction scores. Quantifiable improvements build the business case for broader adoption and secure executive support.

This structured approach minimises risk, builds organisational confidence, and creates a foundation for scaling AI across the enterprise.

Conclusion

AI is delivering measurable improvements in NZ utilities - eliminating scheduling bottlenecks, recovering field hours, and increasing operational efficiency. But the real opportunity is broader: when implemented with proper governance and secure processes, AI becomes a strategic lever that strengthens reliability, reduces risk, and sets up utilities for sustained transformation.

The gap between operators embracing AI and those relying on manual processes will only widen. Those who act now - with structure, oversight, and security embedded from the start - will build competitive advantages that compound over time.

See how AI could work for your utility with structured, secure implementation.

Start a conversation about AI readiness and governance today →

Key Takeaways

• 30–50% of manually scheduled jobs require correction before work begins

• Crews spend 65% of time on jobs with AI scheduling vs 44% with manual scheduling

• NZ electricity demand growing 35–82% by 2050 while workforce remains constrained

• Data quality, human oversight, and secure integration are non-negotiable for AI success

• Spark NZ scaled to 2.3M IoT connections with minimal staff increase through automation

• Start with a controlled pilot to prove ROI before scaling