8. September 2025 By Stephen Lorenzen
Utilities meets AI – specific areas of application for AI in network operation
Network operators are under pressure: increasing complexity, regulatory requirements and a shortage of skilled workers are making operational management more difficult. At the same time, the amount of data in operations and maintenance is growing exponentially. Artificial intelligence (AI) opens up new opportunities here. Generative AI (GenAI) in particular can help make knowledge more accessible, automate processes and support data-based decisions. This article uses use cases from the areas of documentation, maintenance planning and planning processes to show how AI specifically strengthens network operations.
AI in network operations: more than automation
AI is not being introduced as an end in itself, but to provide targeted relief in network operations. Machine learning delivers forecasts and recognises patterns in large amounts of data. GenAI complements these capabilities with language comprehension, semantic analysis and interaction. The interaction of these technologies results in practical solutions – from automated fault analysis to virtual assistants that provide context-specific expertise.
Use case 1: Making documentation and knowledge available
Technical documentation is valuable, but often difficult to use: protocols, whiteboard photos, test reports and emails are available in different formats and in different locations. AI-supported pipelines transform this unstructured information into searchable knowledge.
- OCR and classification capture text even from scans or images.
- LLMs generate structured results such as task lists, tables or summaries.
- Semantic search enables queries such as ‘Which stations had repeated faults last year?’
An internal Q&A system based on GenAI can thus provide answers directly from maintenance logs and fault reports. This saves time, reduces duplication of work and creates transparency about the state of the network.
Use case 2: Predictive maintenance
Disruptions are expensive, both in terms of operational effort and supply interruptions. AI helps network operators identify risks early on and plan maintenance more effectively. Machine learning models analyse sensor values, status data and histories. GenAI explains the results in natural language and links them to recommendations for action.
A typical scenario:
A transformer shows increasing temperature deviations. The system reports:
‘Station T4 has a 42 per cent probability of failure in the next three months. Recommendation: prioritised maintenance in Q3.’
Technically, this process is based on a combination of health scores, failure probability and impact forecasts. The result is a prioritised maintenance programme that uses resources efficiently and minimises downtime.
GenAI in the energy industry
GenAI is driving the energy transition. It helps energy suppliers become more efficient, transform their processes and drive innovation.
adesso supports you with targeted offerings – from a quick start via Quick Check to tailor-made workshops, use case frameworks and the GenAI Factory. This ensures a safe, efficient and practical introduction to generative AI.
Use case 3: Securing approvals and planning
AI can also provide support in formal processes. For example, GenAI systems check construction documents, grid connection applications or planning documents for completeness and consistency. An AI system can, for example, automatically detect when performance specifications in text and circuit diagrams do not match and provide corresponding notes for processing.
Integration into existing processes means that checks are carried out more quickly, documentation is more complete and the quality of planning documents is improved.
Practical example: Value-based maintenance – transferable to grid operation
A current reference project shows how AI-based maintenance solutions can work: Together with RWE Generation, a value-based maintenance (VBM) system was developed that calculates the optimal maintenance time.
At its heart is a central dashboard (‘single source of truth’) that bundles three key services:
- Health Score Service – describes the current status of individual components.
- Failure Probability Service – predicts the probability of faults occurring.
- Failure Impact Service – assesses the impact of a possible failure, e.g. in terms of costs or security of supply.
The system is supplemented by mobile apps for documenting findings and a virtual assistant that answers questions about maintenance status, causes or calculation logic in real time.
While this project was initially implemented in plant operations, the principle can be transferred one-to-one to grid operations:
- Local grid stations can be prioritised based on health scores.
- Critical equipment is monitored in a targeted manner.
- Virtual assistants support technicians and dispatchers in their day-to-day business.
The result: greater transparency, better resource allocation and a measurable contribution to security of supply.
Conclusion: Start gradually, make the benefits visible
AI in network operations is no longer a dream of the future. From documentation and maintenance to planning, there are already concrete areas of application that offer measurable added value. The key is to take a pragmatic approach with small pilot applications that quickly deliver benefits and can then be scaled up.
Value-based maintenance is a proven model that demonstrates how AI can successfully support data-based maintenance. Applied to grid operation, this opens up enormous potential for greater efficiency, lower costs and higher security of supply.
We support you!
Together, we implement AI in network operations. From analysing your data landscape and developing suitable AI use cases to integrating them into existing processes, adesso accompanies you every step of the way.