LB #2 - How to build an AI Agent with KNIME скачать в хорошем качестве

LB #2 - How to build an AI Agent with KNIME

This episode hosts Martin Dieste, a data professional with expertise in finance and accounting, to explore the creation of AI agents using KNIME, an open-source, low-code platform for data analytics. Martin introduces the concept of AI agents, explaining that they require a "brain" (typically a large language model), custom instructions, and a toolkit to perform tasks. He demonstrates building a stock information agent that retrieves stock values from an external API and manages a hypothetical stock portfolio. The session begins with setting up the agent infrastructure in KNIME, including creating folders and workflows, and connecting to a large language model like GPT 5.1. Martin walks through the process of converting workflows into tools that the agent can use, such as fetching stock ticker information and accessing a stock portfolio. He emphasizes the importance of detailed custom instructions and the ability of agents to interact with external systems through tools. Throughout the video, Martin shares insights into the practical applications of AI agents in business, highlighting their potential to automate tasks and increase productivity. He discusses the challenges and opportunities of implementing AI agents in real-world scenarios, drawing from his experience in developing an "offer to invoice" agent named Alan, which manages CRM processes and project setups. The video concludes with a Q&A session, where Martin discusses the future of AI agents and their role in various industries. The episode is part of the "My Data Guest" series, aimed at providing practical lessons on data analytics and AI. #ai #agent#agentic AI #knime