From Chatbots to Agents, Loops and Graphs Executive Brief: A Narrative Evidence Synthesis and Management Recommendations
The Executive Brief presents chatbots, AI agents, loops and agent graphs as overlapping modes of use and architectural forms. It connects current research findings with practical recommendations for governing bodies and shows how people, organisation and technical systems can work together within a responsibly designed hybrid management system. It draws on the St. Gallen Integrated Management Concept and Agentonomics. The publication is a narrative evidence synthesis, not a systematic review.
Table of contents – English 1. Management before technology The management question comes first: AI as a matter of corporate leadership 2. Selecting AI tools and models according to the task The task rather than the brand, and the latest developments in AI models 3. Four overlapping modes of use and architectural forms Chatbots, AI agents, AI loops and agent graphs, including their current levels of adoption 4. From the model to responsible application Models and harnesses within the management system, everyday use, prompting and the Model Context Protocol 5. Curated research radar on AI agents
- From loops to graphs: how agent systems are structured
- Selected agent research: from ReAct to graphs
- Self-improving loops require safeguards
- Whether multi-agent systems help depends on the task
- Longer task horizons and limited safety evidence
- Five testable hypotheses for the further development of AI agents
- Six implications for management
6. Five decisions for governing bodies 7. The future is a management decision An interpretation based on the St. Gallen Integrated Management Concept and Agentonomics 8. Authors and sources Authors, selected references and the complete bibliography