Das Wichtigste in Kürze
AI Agents: From Pilot Project to Productive Use
Why Data Quality and Governance Are Crucial
Microsoft Foundry, Fabric, and MCP Connectors Explained Simply
Four Steps for a Successful Entry into the Agent Economy
Measure Business Value Instead of Technical Metrics
How do agentic systems impact business? The Microsoft Cloud & AI Frontier Week provided answers.
Gradually, companies are moving beyond the AI phases of curiosity and experimentation and are now implementing it. While still cautious, initial examples that Microsoft showcased in its presentations hint at where the journey is headed.
For example, with a guest bot that answers questions, or in the energy sector, provides better weather and production forecasts. Similarly, AI agents are already reducing the increasing administrative burden in a heavily regulated market. Employees gain more time for genuine customer inquiries. At a bank in Germany and a tire manufacturer in France, every employee has a digital assistant that leverages the company's entire knowledge base.
Why many AI initiatives don't get past the testing phase
So-called Frontier Firms are already a step ahead: They are developing business processes deeply integrated with AI and built on a trusted data foundation. The focus is no longer on the technology itself, but on measurable business value.
Especially in Europe, however, companies face significant challenges. Regulatory requirements, the evaluation of AI models, and their economic scalability complicate the path to productive deployment. Added to this is a widespread phenomenon: the pilot project paralysis. Many AI initiatives get stuck in the testing phase because the concrete added value cannot be demonstrated.
Microsoft therefore recommends a holistic approach where data, applications, AI, and security work closely together. With solutions like Microsoft Fabric und OneLake for the data platform, Azure for applications, and Microsoft Foundry for developing AI solutions, a foundation is created to sustainably and securely embed AI within the company.
Microsoft Foundry is particularly exciting: Companies can access thousands of AI models and deploy them flexibly via standardized interfaces. Pre-built connectors also simplify the integration of existing systems. Thanks to open-source components, technological sovereignty is maintained, allowing companies to operate their own AI agents even in an on-premises environment.
The crucial success factor is not just technology, but a clear strategy, a solid data foundation, and a focus on measurable business impact. Only then can an AI pilot project become a sustainable innovation.
Tips for starting your agentic future
Four steps to get started:
- Unify your data: Consolidate all data into a single analytics platform (OneLake, Fabric, Foundry). Utilize WorkIQ (operational data) and FabricIQ (semantic business context).
- Build for the intent/agent economy: Make data, services, and applications machine-readable.
- Invest in employees' AI competencies: Every person needs an AI assistant, just as every process needs an AI agent.
- Prioritize trust, governance, and security: Involve legal, compliance, and risk early on. Principles for responsible AI enable adoption.
For Renato Petrillo from Baggenstos, the first step is primarily about a clear strategic alignment with a focus on people and business. "What do SMEs expect from AI and its agents, where are customer processes affected, and how will collaboration specifically take shape – without losing sight of costs?" AI agents are not about cheaply replacing people, but about creating better customer-centric processes: "Business is done by people for people."
Henrique Moniz de Aragão, Head of Cloud & AI, Enterprise Commercial, UK, Microsoft, concluded with an unmistakable message to the audience and to the European (and thus Swiss) economy: "Don't let caution turn into stagnation."
More information: Microsoft Cloud & AI Frontier Week
