Slovenské Elektrárne, Slovakia’s largest electricity producer, was dealing with increasingly complex operations across **nuclear, hydro, and renewable plants**. As the business grew, several issues became clear:

• Fragmented information: Critical knowledge was scattered across multiple systems and documents. Employees spent significant time just trying to find information, which slowed decision making and increased operational risk.
• Slow, manual processes: Teams often had to manually reconcile project changes and component histories. As Peter Kertys, Head of Data Science, noted, it took so long to find items that there was little time left for actual analysis.
• Rising technical debt: Aging on-premises infrastructure made it difficult, slow, and expensive to scale or introduce new capabilities, especially in fast-moving areas like AI.
• Need for better decisions at scale: With operations spread across many plants and departments, the company needed a way to connect data, scale expertise, and support faster, more confident decisions.

To address this, Slovenské Elektrárne chose to build a **cloud-native data and analytics platform** using **Microsoft Foundry (formerly Azure AI Foundry)**. This allowed them to:

• Consolidate data and reduce fragmentation.
• Introduce AI agents and a chatbot for natural language access to information.
• Scale infrastructure on demand without being constrained by legacy systems.
• Modernize while maintaining strict control over security, compliance, and operational risk.

As their IT leader Gabriel Farkaš put it, AI and cloud transformation became a practical way to rethink how the organization works, not just a technology upgrade.

Security and compliance were treated as non‑negotiable from the start. Working with Microsoft, Slovenské Elektrárne designed a **secure, hybrid architecture** that balances regulatory requirements with the flexibility of the cloud. Key elements of their approach include:

• Hybrid by design: The solution connects on-premises systems with Microsoft Cloud services, but in a controlled way. This lets them modernize and use AI while keeping sensitive operational systems tightly governed.
• Separation from nuclear plant operations: Cloud and AI services run outside core plant systems and are not connected to nuclear operations. This separation helps manage risk in a safety-critical environment.
• Strict governance and access control: The environment uses strong governance practices and role-based access controls so that only the right people can access specific data and tools.
• Human validation in the loop: AI-driven extraction and automation are combined with human review. This ensures that information remains both accessible and trustworthy, which is essential for technical and regulatory work.

As Solution Architect Maroš Bubán explains, every solution is designed with security first. This approach allowed Slovenské Elektrárne to reimagine how they use data and AI without compromising on safety or compliance.

The impact of Azure AI and Microsoft Foundry at Slovenské Elektrárne has been both practical and measurable across productivity, knowledge access, and decision making. Time savings and efficiency

• Overall, the company has saved **more than 2,100 workdays** through AI-driven automation and assistants.
• In 2025 alone, AI automation in technical and regulatory documentation processes saved around **1,800 workdays**.
• Quality-plan processing improvements added another **340 workdays** saved.
• AI-powered summarization now saves **more than 90 workdays annually**.

Improved knowledge access and learning

• More than **1,200 employees each month** use the internal AI chatbot, supported by **10 specialized agents** that connect ERP systems, IT tools, and company documentation.
• Over **4,000 domain-specific test questions** were generated from operational manuals across **12 hydro power plants**, helping standardize knowledge and speed up training.
• Employees can now ask questions in natural language and get contextual answers from documentation, enterprise systems, and historical data, reducing time spent searching and enabling more proactive analysis.

Better use of existing data

• Using Azure Vision, Slovenské Elektrárne converted **29 terabytes of scan data** into a structured digital map of **327 assets**, improving operational decision making.
• Previously unusable or hard-to-access data is now integrated into everyday workflows.

Beyond the numbers, work has shifted from reactive data searching to proactive analysis. Teams spend less time on low-value tasks and more time on higher-impact activities, while having faster, more reliable access to the information they need.