In this chapter
- getting started with the KNIME platform for running data transformation and advanced analytics
- exploring the integration of GenAI flows with external data sources and systems
- building hybrid KNIME and Langflow applications to pull large amounts of data, run LLM calls, and store the outcome for future use
So far, we’ve built GenAI applications using large language models (LLMs) to build conversational experiences. In all these cases, a human user interacted through chat with an LLM-driven entity. Conversational AI is definitely one of the most popular applications of AI these days, but it’s clearly just one of many.
Quite often, the input data doesn’t come from a human user at all. It may arrive from structured sources such as spreadsheets, enterprise applications such as SAP or Salesforce, relational databases, or other digital systems. This means that your AI applications must be integrated with the larger data ecosystem you use in your activities. Whether it’s an Excel file or a cloud-based customer relationship management (CRM) system, you can use a data source of any kind to fuel inputs for the LLM to do its magic. As illustrated in figure 7.1, this means going beyond the conversational paradigm we’ve seen so far, paving the way for a multitude of business applications.
Figure 7.1 The LLM’s input text might come from various sources—files, systems, or users.
