KMWorld Conference Wrap-Up and Look Ahead
KM Practices and Strategies
Talks during the main conference also revolved around AI technologies, tempered with insights regarding human involvement. The practical side of integrating GenAI into KM projects was a constant throughout the conference. Seth Earley and Giovanni Piazza outlined a project Earley Information Science did for a global pharmaceutical company that used existing SharePoint infrastructure but brought in large language models (LLMs), taxonomies, ontologies, and retrieval-augmented generation (RAG) to evaluate its drug portfolio. Earley stressed the importance of information architecture in conjunction with AI.
Nonprofit organizations need KM just as much as those in the for-profit sector. Jennifer Anna, CHG Healthcare and formerly with the World Wildlife Federation (WWF), and Jessica DeMay, Enterprise Knowledge, ran through all the steps they took to build a KM program for the WWF from scratch. High-level solutions for search and findability, a knowledge portal, KM governance, and content strategy were required. Lessons learned included the importance of adaptability and leadership vision, the foundational nature of KM for broader organizational goals, the need for systems and processes to be designed with scalability in mind, and how well-organized content is crucial.
Based on his long career spanning multiple companies, nonprofits, and universities and living in four countries, Scott Leeb, now with Fragomen, has identified five challenges for KM: converting tacit to explicit knowledge, demonstrating ROI and articulating the value of KM, the visibility and find- ability of knowledge, collaboration, and viewing technology as an enabler not a solution. How these are handled depends upon the individual organization, its culture, and its structure.
Along with the ubiquitous (obligatory?) mentioning of AI and GenAI throughout the week was a track all about the value of storytelling. Knowledge shared by humans through the stories they tell has a unique impact on other humans in a fashion that GenAI, at least so far, cannot equal.
Communities of Practice (CoP) are also human-centered. Setting up CoPs within organizations contributes to knowledge transfer and breaks down knowledge silos. Organizations such as the World Bank have multiple CoPs. At USAID, knowledge managers use the Milton-Lambe knowledge manage- ment framework—Discuss, Synthesize, Document, and Find/ Review—from the Knowledge Manager’s Handbook, by Nick Milton and Patrick Lambe—to define and embed effective knowledge management in events and learning mechanisms. The Discuss framework works particularly well within CoPs.
Taxonomies and Enterprise AI
Taxonomists, claimed Nike’s Aaron Lehnert, are “ghosts in the machine.” They act as mediums, psychics, and interpreters of language used by subject matter experts, stakeholders, and consumers, which they skillfully transform into taxonomies. He also warned about potential bias in taxono- mies and called for them to be “models for truth.”
Metadata was of interest to Richard Huffine. At the FDIC, Huffine relates metadata to governance, data privacy, deduplication, search and findability, and data protection. A semantic layer is central to data catalogs and data platforms.
Companies and Suppliers Mentioned