Companies Plan AI Agent Deployment, Focus Shifts to Scaling

Businesses across German-speaking regions are increasingly planning to implement AI agents, but few projects successfully transition from pilot phases to production. According to a Lünendonk & Hossenfelder study commissioned by Reply, 43 percent of companies intend to deploy AI agents. However, only about one in four AI projects reach the production stage.

This situation raises questions about how AI applications can be reliably integrated into daily business operations and scaled across processes. Agentic AI enables the automation of complex, multi-step tasks. These agents can independently break down objectives, aggregate information from diverse sources, and interact with enterprise applications. Significant investments in technological and organizational foundations are deemed essential: 98 percent of companies plan to invest in data quality and integration, while 94 percent will focus on AI development and orchestration platforms.

The primary business objective is to enhance efficiency and reduce costs, with 96 percent of respondents expecting these benefits from AI agents. Furthermore, 89 percent see potential in modernizing legacy systems. AI agents can analyze and document existing applications, facilitating maintenance and future development. Governance is also becoming more critical, as 62 percent are planning a central catalog of approved AI agents.

Reply's customer projects, including those with Audi and Siemens Healthineers, illustrate practical applications and scaling of agent-based AI. Audi utilizes a multi-agent system for its software development, and Siemens Healthineers has developed an agent factory to streamline the creation of specialized AI agents. Filippo Rizzante, CTO at Reply, emphasizes that successful scaling requires AI agents to understand a company's specific business context and be securely connected to data and systems.