The initial wave of euphoria surrounding artificial intelligence has given way to a much more sober assessment in many industrial companies. Whilst numerous AI projects have been launched in boardrooms and innovation departments in recent years, the operational benefits of many initiatives have so far fallen short of expectations. Pilot projects fizzle out, proof-of-concepts never reach production, and questions about the actual return on investment are becoming increasingly vocal.

This is precisely where Munich-based elunic AG comes in with its new platform, eluna Deep Agent. The company deliberately positions its solution not as yet another generic AI chatbot, but as an industrial automation platform for knowledge-intensive business processes – combined with a clear focus on digital sovereignty and European data sovereignty.

The real efficiency gap often lies outside production

Industrial manufacturing has been one of the most highly automated sectors of the economy for years. Production facilities, robotics, industrial control technology and sensor systems operate in a highly digitalised manner in many companies. At the same time, however, a striking contradiction is evident: in the adjacent office and administrative processes, manual procedures continue to dominate in many places.

Quotes are calculated manually, specifications checked by hand, complaints distributed across various systems and technical service processes only partially automated. Yet it is precisely these knowledge-intensive processes that cause significant time wastage, media breaks and staff strain.

According to elunic, this is where the industry’s real automation gap lies. The eluna Deep Agent platform is designed to address and automate precisely these processes – including quotation preparation, supply chain monitoring, complaints management, compliance audits and technical support.

The approach differs significantly from traditional AI assistants. Whilst many systems are primarily designed for dialogue-based interaction, eluna Deep Agent is intended to be able to independently understand objectives, derive action steps and execute processes across systems.

From Industrial IoT to an industrial AI platform

This development did not come out of nowhere. elunic AG has been working on software solutions for the connected industry for more than fifteen years. The starting point was the Industrial IoT platform shopfloor.io, later followed by ShopfloorGPT, the company’s first AI solution specifically designed for industrial applications.

With eluna Deep Agent, the company is now consolidating this development into a comprehensive AI automation platform. According to the company, more than 80 industrial firms – including GROB, Flender, Murrelektronik and Schunk – are already working with solutions from elunic.

AI investments should become measurable

What is remarkable here is not so much the technology itself as the implementation approach. Instead of simply providing companies with an AI tool, elunic focuses on an upstream analysis of the process landscape.

First, it examines which workflows offer the greatest economic leverage for AI automation. Only then does technical integration take place. The aim is to tailor AI projects specifically to concrete business problems and avoid typical ‘innovation theatre’ effects, where pilot projects generate attention but deliver no sustainable benefits.

Claudio Gusmini, Executive Director of elunic AG, describes this as a direct response to the growing disillusionment with AI in industrial companies. Many firms have realised that productivity gains do not automatically result from the use of large language models, but only through their integration into real business processes.

Digital sovereignty is becoming a strategic factor

Alongside process automation, a second topic is increasingly coming to the fore: digital sovereignty. Particularly among industrial SMEs, concerns are growing about excessive dependence on American cloud and AI providers.

Design data, technical documentation, customer contracts and production know-how are among the most sensitive information held by many industrial companies. Consequently, the question of where this data is processed and stored becomes critical.

eluna Deep Agent addresses this issue with an architecture that operates entirely without a mandatory US cloud connection. The platform can be operated on-premises, in a private cloud with EU hosting, or as a managed service. Furthermore, eluna adopts an open-source-oriented approach designed to give companies long-term control over data, models and source code.

Particularly in light of regulatory developments such as the EU AI Act or growing compliance requirements, this architecture is increasingly becoming a strategic competitive advantage.

AI is becoming an infrastructure issue for industry

The discussion surrounding industrial AI is thus shifting increasingly away from spectacular individual applications towards fundamental infrastructure issues. Companies today are asking less whether AI should be used, and more under what conditions it can be operated productively, securely and in a way that remains controllable in the long term.

It is becoming clear that the next phase of industrial AI development is likely to be characterised less by short-term demonstrations and more by deeply integrated platforms that automate operational processes, meet regulatory requirements and simultaneously safeguard European data sovereignty.

The key figures presented by elunic illustrate the economic potential of such approaches. According to the company, response times in technical support are reduced by 47 per cent, whilst escalations fall by 32 per cent. In AI-supported specification reviews, testing speed and efficiency are also expected to increase significantly.

For industrial SMEs, this could confirm a key insight from the current phase of AI: it is not the most spectacular AI that wins in the long term, but rather that which can be integrated into real business processes in a stable, secure and measurable way.