Edison on Microsoft Azure gives BKW a platform for hands-on AI

BKW is systematically advancing its digital transformation, seeking to get technological innovation broadly embedded throughout the company. Artificial intelligence (AI) is seen not only as a driver of efficiency, but also as a tool to prepare for growth and new business potential. In a heavily regulated environment where data protection, security and compliance are at a premium, however, the company had no illusions: Responsible and scalable AI use had to be built on a very firm foundation. 

That is where Edison came in. Edison, an AI-supported platform developed internally on the basis of Microsoft Azure, combines Generative AI functions with semantic search capabilities. The platform enables staff to analyze internal information in context, to generate content, and to handle recurrent tasks more efficiently. “We see AI not just as a driver of efficiency, but as a strategic tool to help us grow where we want to grow,” explains Yamshid Farhat, Lead AI & Technology at BKW. “For that, we need a technical environment that is secure and flexible and can be used across the enterprise – by any team and in any place.” Edision provides precisely this framework. Today, it has come to symbolize the company’s new and strategically embedded approach to AI: rooted in technology, fully compliant and open to practical use in the daily routine. 

Internal skills + Microsoft expertise = a scalable IT platform 

A conscious decision was made to build Edison with internal resources. Supported by Microsoft, BKW trusts its own core team to develop technological sovereignty. “We don’t just want the technology to work: We want to understand what we are doing,” explains Raffael Hilber, head of the Group-wide AI initiative. “Only then will we in future be able to decide autonomously what partnerships make sense for us and where we need to call in specific areas of external expertise.” One central success factor – and a key reason why Edison could be developed in only eight weeks – was BKW’s wealth of in-house technological knowledge. 

The technical architecture is based on a modular Python backend that orchestrates the various Azure components – including Azure AI Search, Azure AI Foundry and Azure AI Document Intelligence. The platform uses Azure AI Search to access structured and unstructured content in BKW’s SharePoint system. This is then merged with Generative AI on the basis of Azure OpenAI Services and Azure AI Foundry to facilitate context-relevant responses. This interplay delivers fast, robust results and meets the strictest security, scalability and regulatory conformity requirements. The architecture can be expanded flexibly – up to and including advanced scenarios such as the use of agentic AI. 

From the outset, the project team worked closely with Microsoft on both the technical and the strategic levels. The leaders benefited from regular discussions with Azure experts, the experience of other industries and selective deep dives into the topics of architecture, governance and security. This close cooperation meant that key questions could quickly be cleared up, proven practices adopted and the platform successfully developed. 

Hands-on launch thanks to active staff participation 

The project team planned a phased roll-out from the word go. The first set of specialized applications was tested in small groups with a clear thematic focus. This added tangible value and allowed specific feedback to loop back into the process. Digital training courses, peer sessions and hands-on workshops then opened up the outcomes to the entire workforce. “We didn’t want a solution that was just for a handful of people. We wanted a platform that would genuinely add value for all our staff – and that could be integrated organically in our routine work,” Yamshid Farhat explains. 

The team crafted the first concrete use case in corporate communication: The aim was to accelerate the editorial process and respond to inquiries more efficiently – based on internal content such as media inquiries, blog articles and publications. The results were impressive: Media inquiries were processed 50% faster, and the quality of communication improved tangibly. 

Building on this platform, Edison was launched as a fully compliant chatbot for all employees. It is complemented by unit-specific solutions that are also based on Edison. 

Only two months after enterprise-wide roll-out, 8% of staff were already using Edison actively – a clear sign of how well it had been received. At the same time, a pipeline of more than 40 documented use cases was drawn up: Potential new applications – from customer service to operations and technology – are being derived from it all the time. 

Edison – The starting point for a long-term AI strategy 

Edison has given BKW a robust platform that brings technological development, culture change and strategic orientation together in the context of AI usage. The solution supports a wide variety of applications, can be expanded continually and receives active support from a growing number of departments. “For us, Edison was the starting point for an enterprise-wide journey,” Hilber explains. “We don’t want to use AI just here and there: We want to systematically embed it wherever it genuinely adds value.” New areas of application are coming into view one by one, especially in core business processes such as energy trading. Edison is thus creating conditions that, in future, will let AI be integrated even more deeply in operating business. 

BKW is also preparing to deploy additional AI-based solutions that build on the company’s heightened understanding of the technology. Examples include plans to roll out Microsoft Copilot, the development of specialized models and scenarios in the energy business, and the provision of AI-based wizards to help on-site technical staff maintain infrastructure. 

In the meantime, the team is now testing new approaches such as agentic AI systems that comprise multiple models and work together to solve complex tasks. The medium-term objective is to involve AI in dealing with demanding use cases such as dynamic resource planning and multi-tiered customer inquiries. In the shape of Edison, BKW has now given itself a future-proof foundation on which it can continue to develop AI responsibly and selectively: practical, scalable solutions that are deeply embedded in the organization. 

BKW Energie AG

Headquartered in Berne, BKW AG is an energy and infrastructure service company with an international reach. Its roughly 12,000 employees develop, build and operate energy, telecommunication, infrastructure and environmental systems for companies, government agencies and private individuals. In all these fields, BKW pursues a clear strategy: to grow and develop sustainably and profitably, to tap new business models, and to actively shape the digital transformation through 2030.