Technology is no longer just an isolated IT issue. It has become a decisive factor in economic competitiveness. Artificial intelligence is transforming business models, cloud infrastructures are shifting the boundaries of competition, and the global race for data, talent, and capital is accelerating noticeably. The Tech Innovation issue of Finanz und Wirtschaft is dedicated precisely to this development. It shows how technology is becoming the strategic foundation of modern companies today – and why executives must increasingly grapple with issues that were once the exclusive domain of IT.
Against this backdrop, Sanche Baskaran, CEO of Epic Fusion, discusses in an interview why many AI initiatives fail to achieve real economic impact despite significant investments, and what structural changes companies must make to leverage AI strategically.
Many companies are experimenting with AI – but only a few are generating real value. Why?
The most common misjudgments are surprisingly consistent: First, AI is viewed as a tool rather than a business capability that transforms processes, roles, and governance. Second, many companies start with exciting use cases but lack a solid foundation: data access, protection, governance, and operations. Third, adoption is underestimated – without trust and leadership, AI remains a gimmick. AI must be embedded in day-to-day business operations; otherwise, it remains just a show. If you don’t define impact, you end up with experiments. AI does not fail because of technology. AI fails due to a lack of strategic leadership.
What sets frontier companies that use AI strategically apart from others?
Pilot projects test technology. Strategic companies build capabilities. Frontier companies integrate AI into their value chain – not into isolated use cases. They invest in data architecture, governance, and organizational capabilities. Reactive companies buy tools. Strategic companies build structures. That’s the difference.
What conditions must companies create so that AI can deliver measurable impact?
AI initiatives rarely fail because of the model itself. They fail due to unclear processes, a lack of accountability, and insufficient integration. AI is not an IT extension. AI belongs in the corporate strategy. Many underestimate the organizational dimension. People budget for technology, but not for transformation. Solutions implemented quickly without governance result in higher costs in the long term – both operationally and from a regulatory standpoint. Sustainable AI strategies are not a sprint, but a structured scaling model.
Where will the greatest economic leverage lie in 2026: automation, assistance, or agents?
The greatest leverage lies in intelligent decision support and in the areas where companies currently waste the most time. It i s not just about automating tasks, but optimizing them. Pricing, risk analysis, resource allocation, service management – anywhere decisions are made on a daily basis. Those who systematically improve decision-making increase both margins and speed at the same time. In the short term, assistance may be the turbocharger. In the medium term, agents are the scaling lever because they take on tasks as digital colleagues and relieve the burden on teams.
How does Epic Fusion integrate intelligent systems into established IT landscapes?
Integration is an architectural issue, not a modeling issue. Instead of parallel AI worlds, we integrate along existing processes with clear interfaces and security mechanisms. Stability comes from clean design. We start pragmatically, with low risk and scalability – first assistance, then agents, then autonomous workflows. This keeps operations stable and risk manageable.
What role do data architecture and cloud strategy play in this?
AI is only as good as its data foundation. Without structured data flows, clear ownership, and scalable cloud architecture, AI remains fragmented. Those who fail to invest here are building on an unstable foundation. We consistently view data as a corporate asset. This means that aclear data lifecycle, clean data classification, and structured cloud migration form the foundation for a company’s ability to act. Only when data is secure and available can AI be deployed responsibly and effectively.
How are AI and automation changing cyber resilience?
AI not only increases productivity; it also changes the attack surface. Speed, connectivity, and autonomy affect both companies and attackers. That is why companies must rethink their security architecture, access controls, and monitoring. Autonomous systems require clear boundaries and oversight mechanisms. Resilience is becoming an integral part of every AI strategy. This is precisely why security must be embedded «by design»: from cloud security to organized security operations (SOC) to regular risk reviews. Security must not be reactive – it is a prerequisite.
Where is AI headed in the coming years?
From assistance to autonomy. AI is becoming less of a conversation partner and more of a system that coordinates, decides, and executes tasks. This requires a new interplay of leadership, governance, and platform operations. People set the direction. Autonomous agents are becoming digital colleagues who take on structured decision-making and coordination tasks. They work across systems, respond in real time, and significantly reduce the workload on business units. But: They need clear governance.
You’re currently developing your own product with autonomous AI agents called Frontier Engine. Why?
Because we are convinced that companies need more than off-the-shelf solutions. We are not waiting for hyperscalers to deliver solutions. With Frontier Engine, we’re deliberately developing our own platform that makes autonomous AI agents controllable, scalable, and compliant. Our customers today need solutions that can be adapted to their organizations. They must be able to securely integrate agents into existing systems, enforce governance, and scale at the same time. We build the product and orchestration logic ourselves, including our own components such as the MCP Server. This ensures that companies can integrate autonomous agents into their IT landscape securely and transparently. We do not build visions. We build what our customers actually need.
What problem do autonomous agents solve that traditional automation cannot?
Traditional automation follows fixed rules. Autonomous agents can make context-based decisions, prioritize, and coordinate. They connect systems, data, and processes dynamically – not statically. This is a fundamental difference. After all, added value is created when agents do not work in isolation but are orchestrated as a team.
Why is multi-agent orchestration the key to scaling?
A single agent solves a task. Orchestration coordinates many specialized agents along a value chain. These agents work together, exchange context, and execute tasks across systems. Scalability comes from structure – not from individual models. In our approach, this multi-agent orchestration is embedded into an existing platform architecture with components such as AI Foundry, Microsoft Fabric, and AI Search.
Why is change management crucial for AI projects?
Trust is built through transparency. Employees must understand what an agent does, and what it does not do. Leadership must communicate clearly, because AI does not replace responsibility. That is why we state clearly: Build trust, embed new ways of working, and redefine roles. In doing so, leadership must shift from control to coaching. Failing to actively lead change creates resistance.
If you could give a CEO one piece of advice today, what would it be?
Do not delegate AI. View AI as a strategic management tool for your company. Treat AI not as a tool, but as a strategic corporate capability.