AI for Enterprise KM & Collaboration
Delia Lazarescu, Tech Unicorn & Technical Consultant, Google, believes that the future of KM will involve knowledge shifting from human readable to AI-digestible. Knowledge will be optimized for AI first, then delivered to humans. This is a shift we will have to adjust to.
In her keynote talk at the KM & AI Summit, she said that, for KM, AI matters due to the current situation of the cost of time wasted trying to find information and reinvent knowledge due to lack of access. Lost productivity leading to human frustration is another reason to look at AI adoption.
In 2023, a breakthrough year, models neared human performance, costs exponentially decreased, and investment in AI increased dramatically. Soon AI will have the computational power of all humanity, given the crazy improvement in LLM capabilities, she thinks. Suddenly we can do so much more with so much less.
She singled out several best practices regarding central/simplified documentation, including having a single source of truth, structured formatting, LLM-friendly language, and regular updates and version control. Standards are important because LLMs by themselves are useless—they can only predict. They become useful when combined with MCP (Model Context Protocol), an open protocol to attach external data sources and tools.
Semantic structuring for AI calls for clear metadata, consistent headings, semantic linking, natural language optimization, and regular semantic audits. AI's impact on KM and collaboration is evident when knowledge discovery is optimized, workflows streamlined, collaboration empowered, and knowledge sharing enhanced. Overall, these drive innovation.
A real world example of AI-powered knowledge search and discovery is Deep Research, which includes smart semantic search, context-aware recommendations, voice search, and chatbots. Another is automated knowledge curation and organization with AI-auto-tagging, AI-powered summarization, and duplicate and redundant content detection. She also discussed AI-powered knowledge extraction and synthesis, knowledge graph and AI-assisted decision-making (which is a bit like having a second brain), and AI for knowledge sharing and collaborating
Key challenges to unlocking the potential of AI include data silos and hallucinations, resistance to change, and AI governance and ethics. To meet these challenges, she suggested breaking down silos, patiently fostering AI adoption, and establishing AI governance. This may be easier said than done.
The KM & AI Summit, held March 17-19 in Scottsdale, Arizona, heralds a new era of KM.
The KMWorld conference returns to the J.W. Marriott in Washington D.C. on November 17-20, 2025 https://www.kmworld.com/Conference/2025
KMWorld 2025 is a part of a unique program of five co-located conferences, which also includes Enterprise Search & Discovery, Enterprise AI World, Taxonomy Boot Camp, and Text Analytics Forum.