Agentic AI in Information Systems
What are the fundamental impacts of generative AI on information systems? How can LLMs be guided responsibly within a human-in-the-loop Agentic AI workflow, leveraging knowledge graphs in RAG to extract data efficiently and transparently?
Reliable information is the essential foundation for decision-making in market research, financial investigations, and legal domains. Regardless of the subject area, effective information retrieval involves identifying, aggregating, visualizing, and interpreting relevant information from diverse data sources.
As part of a human-in-the-loop Agentic AI workflow, LLMs can be integrated into information systems. Responsible use and a thorough understanding of AI tools are essential to ensure the generation of reliable insights grounded in factual knowledge.
Similar to how an orchestra is conducted or marionettes are controlled, AI agents can also be directed in a targeted manner where humans deliver capability and AI provides scalability, for example using the following tools as part of an information system:
- LangGraph, Agent Development Kit, AutoGen: Open-source frameworks for Agentic AI
- DeepResearch, NotebookLM, Perplexity Labs: LLM-supported research with RAG integration
- Model Context Protocol: Interface for connecting databases and analytics environments
"To think that we know everything is a condition of the human mind. The animal within us can not tolerate the possibility that knowledge is a matter of selection, judgement a matter of focus, clarity a consequence of exclusion. There is not one truth. Reality does not break along clean lines."