KNIME, the open-source data analytics and AI company, has unveiled its AI companion, K-AI, designed to empower users to co-create advanced data workflows with AI.

Figure 1. KNIME AI Companion: Empowering Data Workers to Unlock Insights.

K-AI offers features like answering questions, providing recommendations, and building or extending entire data workflows based on user prompts. By accelerating the time to insights, K-AI ensures complete transparency and control for users over its AI-driven processes. Figure 1 shows KNIME AI Companion: Empowering Data Workers to Unlock Insights.

This innovation represents a significant leap in automating data tasks, enabling users of all skill levels to derive insights more efficiently. KNIME’s visual workflows provide a clear, step-by-step representation of how AI contributes to each process, acting as comprehensive documentation and including narrative text to simplify review.

The introduction of K-AI reflects a paradigm shift for data workers, moving from complex, code-heavy environments to transparent, well-documented workflows accessible to all. It also supports skill development across organizations by reducing reliance on internal experts.

K-AI operates in two primary modes:

1.Q&A Mode:

Users can ask technical or theoretical questions related to KNIME, receiving detailed answers to understand node functionality and improve workflows.

2.Build Mode:

K-AI assists users directly within the workflow canvas by adding, connecting, and configuring nodes based on user input. It can construct workflows with a single prompt, suggest next steps, style visualizations, and generate Python or R scripts. Throughout, K-AI explains its actions, ensuring users understand the additions and their purpose.

In Q&A Mode, K-AI also suggests extensions, workflows, and nodes from the KNIME Community Hub, streamlining the discovery of existing solutions for reuse and adaptation.

According to KNIME CEO Michael Berthold, “KNIME’s mission has always been to make data accessible to everyone, regardless of technical background. K-AI represents a significant advancement in enabling individuals to build workflows faster while learning throughout the process. By integrating AI assistance into our open-source platform, we are accelerating data work and empowering users to grow their expertise. We’re eager to see the new possibilities K-AI will unlock for users.”

First introduced as an experimental extension in early 2024, K-AI has now been fully integrated into KNIME’s Analytics Platform 5.4. This latest release includes enhanced LLM connectivity, LLM evaluation tools, and a more streamlined user experience:

• Expanded AI extension support for Databricks LLMs, GPT4All models on IBM Granite, Microsoft Phi 3.5, and continued compatibility with OpenAI, Azure, Hugging Face, and Ollama models.
• Tools for detecting vulnerabilities such as bias, hallucination, and data leakage in AI-augmented workflows, leveraging Giskard LLM and Giskard RAG nodes.
• Connectivity improvements for Tableau, Box, and Databricks, alongside KNIME’s extensive integration library of over 300 data sources.

Source: Database

Cite this article:

Priyadharshini S (2024),"KNIME Introduces AI Companion To Help Data Workers Uncover Insights from Their Data",AnaTechmaz, pp. 61