Artificial intelligence (AI) is rapidly transforming the management of asset information across the energy sector. As organisations seek to leverage AI to drive operational efficiency, reliability, and innovation, the importance of high-quality, reliable asset information has never been greater. AI is only as effective as the data it consumes; without robust, accurate, and well-managed asset information, even the most advanced AI applications will fail to deliver value.

This article examines the implications of AI for asset information management, drawing on insights from a recent International Energy Agency (IEA) report on AI and energy. The focus adopted here is on how AI is being used to manage electricity networks, the resulting challenges and opportunities for asset information, and the evolving capabilities required within asset information teams. This discussion is from a New Zealand perspective, reflecting both global trends and local examples.

AI Adoption in the Energy Sector

The IEA report provides a comprehensive overview of AI's impact across energy sectors, including oil, gas, and electricity. Within the electricity sector, AI adoption is analysed across three main domains:

Generation
Networks
End users

This has direct implications for asset information professionals who are supporting the adoption of AI in managing electricity networks. Here, AI is being applied in diverse ways, each with distinct implications for asset information and management practices.

How Utilities Are Using AI: Global Survey Insights

The recent survey of transmission and distribution system operators from 13 countries by the IEA highlights how AI is being deployed across four key areas:

Real-Time Operations (e.g., control room activities)
Asset Operations (work delivery, capital works, operational work management)
Maintenance (renewal and maintenance planning)
Grid Development (network/system development and planning)

Source: https://www.iea.org/reports/energy-and-ai

1. Real-Time Operations

Real-time operations, particularly in control rooms, currently see the lowest adoption of AI, with only 23% of survey respondents using AI in this context. The limited uptake is attributed to the need for rapid, high-stakes decision-making and the lack of established benchmarks or industry standards for real-time AI validation. There is greater acceptance of AI in scenarios with extended decision timeframes, where human oversight remains critical.

2. Asset Operations

AI adoption is highest in asset operations, with around 70% of respondents utilising AI for work delivery and management. For example, the State Grid Corporation of China employs AI to optimise the scheduling of work and outages. In this context, AI often functions as an "auxiliary assistant," providing decision support by analysing large volumes of data and offering recommendations, while leaving final decisions to human experts.

3. Maintenance

Maintenance and renewal planning also see high AI adoption, with approximately 70% of respondents leveraging AI to optimise maintenance schedules and equipment replacement. AI-powered predictive maintenance, based on condition monitoring, is increasingly common. Around 30% of respondents use image recognition AI for asset monitoring and vegetation management. For instance, the grid operator in Ireland utilises machine learning for predictive maintenance, extending asset lifespans and reducing unplanned outages.

AI is also used for post-mortem analysis such as performing detailed investigations into asset failures. This activity benefits from AI's ability to generate insights without operational time pressure. Notably, 8% of respondents use generative AI for fault diagnosis, highlighting the growing sophistication of AI applications in this area.

4. Grid Development

AI is being used for long-term scenario planning and network development. By processing complex datasets, AI enables power system operators to simulate a wide range of scenarios, supporting informed decisions about grid investments and operations. AI's ability to balance multiple, competing objectives is particularly valuable in complex network planning and in integrating renewable energy sources.

However, a key limitation is the lack of transparency and auditability in AI-driven decision-making. Regulatory requirements often mandate detailed documentation of decision rationales, but current AI systems struggle to provide clear, written explanations for their actions.

Implications of AI for Asset Information Management

The integration of AI into electricity network management has profound implications for how asset information is managed. Five key areas are particularly impacted.

1. Data Quality and Integration

The adage "garbage in, garbage out" is especially true in the context of AI. While human experts may spot and correct obvious data errors, AI systems lack this intuitive quality assurance and will process whatever data they are given. This underscores the need for rigorous data governance, ensuring that critical data is accurate, consistent, and trustworthy.

Effective AI applications also require seamless integration of diverse datasets. This demands consistent use of unique identifiers across systems, robust data architecture, and strong data management practices to enable reliable, cross-system analytics.

2. Feature Extraction from Unstructured Data

AI excels at extracting structured information from unstructured data sources such as images, videos, LiDAR scans, and natural language text or audio. This capability may reduce the need for manual data structuring—such as as-built documentation or inspection records—by enabling the direct derivation of asset information from raw data. The shift towards AI-driven feature extraction will require new approaches to data capture, storage, and validation.

3. Predictive Maintenance

Predictive maintenance leverages real-time data analysis to anticipate equipment failures and trigger proactive interventions. This approach offers significant benefits, including extended asset lifespans, reduced downtime, and optimised maintenance schedules. The latest ISO 55000 asset management standards now include predictive maintenance as a requirement, reflecting its growing importance.

AI's ability to analyse large datasets and model diverse end-of-life scenarios is driving a shift from periodic inspections to continuous condition monitoring. This evolution will fundamentally change how asset information is collected, managed, and utilised, with a focus on continuous data streams and real-time analytics.

4. Data Volume and Complexity

AI applications thrive on large, complex datasets. As a result, asset information management is shifting from the oversight of individual records to the management of metadata, aggregated datasets, and exception reporting. This requires new tools and processes for data aggregation, quality assurance, and insight generation.

5. Regulatory and Security Needs

The increasing volume and diversity of data used in AI applications heightens regulatory and security challenges. Organisations must ensure compliance with data privacy, disclosure, and auditability requirements. For example, tracing the source of an error—such as an incorrectly recorded switch number—becomes more complex when data is unstructured and distributed across multiple systems.

Robust data lineage and audit trails are essential to confidently investigate and address data-related incidents. Regulatory bodies, such as the Commerce Commission, highlight the need for accurate asset data, such as the recent warning issued about the impact of data inaccuracies on asset valuations.

Capabilities for an AI-Enabled Asset Information Team

The emergence of AI-driven asset management is fundamentally reshaping the expectations and requirements of asset information teams. As AI becomes increasingly embedded in operational and strategic decision-making, new skills, roles, and capabilities are rising to the forefront. Five key areas stand out as particularly important for organisations looking to future-proof their asset information functions.

Data Governance and Strategy is the foundation of effective AI adoption. Within this area, roles such as the Asset Data Manager and Asset Information Data Architect are critical. The Asset Data Manager is responsible for overseeing data standards, quality rules, and governance frameworks to ensure the integrity of asset data. Meanwhile, the Asset Information Data Architect designs scalable, integrated data architectures that support both AI-readiness and alignment with broader enterprise strategies. This focus on governance and strategy is essential, as AI systems are highly sensitive to data quality and consistency; without strong governance, organisations risk errors that can undermine the value of AI initiatives.

Digital Engineering and Models is another area gaining prominence. Here, roles like the Digital As-Built Coordinator and Digital Asset Modelling Specialist are crucial. The Digital As-Built Coordinator manages the transition from construction documentation to operational digital records, ensuring that raw files are properly digitised and structured for AI use. The Digital Asset Modelling Specialist develops and maintains digital twins or 3D representations of assets, providing the structured visual context required for AI-driven insights. While AI can extract features from unstructured data, this information must be systematically organised and modelled to be truly effective in asset management.

Predictive and Condition-Based Maintenance is rapidly becoming a central capability for asset information teams. The Asset Condition Monitoring Specialist is tasked with deploying and managing sensors to collect real-time condition data from assets. Complementing this, the Asset Lifecycle Data Scientist analyses condition and performance data using AI and machine learning techniques to predict asset behaviour and optimise maintenance schedules. Predictive maintenance is key to achieving cost savings, improving reliability, and extending asset lifespans. Real-time data and advanced analytics are essential to unlock these benefits.

GIS and Spatial Intelligence are also increasingly vital. The GeoAI Data Scientist develops AI models that leverage geospatial data for predictive analytics, anomaly detection, and asset optimisation, while the Spatial Integration Specialist integrates real-time IoT sensor streams with geospatial platforms. This enables continuous monitoring and harmonisation of large volumes of spatial and operational data. Spatial data provides critical context for asset management, and integrating GIS, remote sensing, and IoT data significantly enhances operational efficiency and decision-making.

Integration and Cybersecurity round out the list of essential capabilities. The IoT Asset Integration Engineer connects sensors, edge devices, and smart assets into enterprise systems, enabling real-time insights that power AI applications. The Asset Cybersecurity and Compliance Manager is responsible for overseeing data security, privacy, and regulatory compliance, ensuring robust traceability and protection of sensitive information. As AI systems rely on interconnected, real-time data flows, ensuring secure integration and compliance is vital. The complexity and risk associated with data usage increase as AI becomes more deeply embedded in operational decision-making, making these roles more important than ever.

In summary, the rise of AI-driven asset management is driving a shift in the skills and capabilities required within asset information teams. Organisations that invest in these areas will be well-positioned to harness the full potential of AI, while maintaining the integrity, security, and value of their asset information.

Real-World Applications of AI in New Zealand Electricity Networks

AI is already being used in various ways by New Zealand electricity network operators:

Image Recognition: Computer vision is being used to identify asset condition, defects and safety risks.
LiDAR Analysis: Advanced analysis of LiDAR data supports asset inspection and condition assessment.
Natural Language Processing: AI tools are being used to interpret reports from inspectors and fault responders, extracting structured information from narrative text.
Document Refinement: AI-powered language models are assisting in refining Asset Management Plans (AMPs) and other regulatory documentation.

These examples illustrate the growing integration of AI into everyday asset management practices, highlighting both the opportunities and challenges ahead.

Conclusion

The adoption of AI in asset information management is reshaping the energy sector, offering significant benefits in efficiency, reliability, and decision-making. However, the value of AI is fundamentally dependent on the quality, integration, and governance of asset information. As AI applications become key consumers of asset data, organisations must invest in robust data management practices, new digital capabilities, and specialised skills.

A future-ready asset information team will combine expertise in data governance, digital engineering, predictive analytics, spatial intelligence, and cybersecurity. By embracing these changes, energy sector organisations in New Zealand and beyond can harness the full potential of AI while ensuring the integrity, reliability, and value of their asset information.

Contact Us

Thank you for reading. If you have any questions on this article or would like tailored advice on this topic, please complete the form below and we will be in touch as soon as possible.

AI Implications for Asset Information