Knowledge discovery and collaboration meet agentic AI with Fluid Topics and Guru
Agentic AI brings new opportunities for KM: agents, proactive and autonomous, can pinpoint where gaps in knowledge occur, span multiple knowledge hubs to connect disparate pieces of information into usable knowledge, and make non-scripted decisions.
Heading the priority list for knowledge managers is enabling enterprise knowledge foundation for AI. Although agents don't need to wait for a prompt to act, humans are still necessary to provide guidance and oversight.
New knowledge workflows redistribute responsibilities between humans and AI, not only profoundly changing how knowledge is discovered but also the role of human KM teams.
KMWorld recently held a webinar, Agentic AI Meets KM: Revolutionizing Knowledge Discovery and Collaboration, with Fabrice Lacroix, CEO and founder at Fluid Topics and Hillary Curran, head of customer innovation and AI enablement at Guru, who discussed why different knowledge domains require different approaches and solutions.
According to Curran, the problem isn’t the AI models, it’s the knowledge behind them. Agents hallucinate, silos multiply, and errors amplify.
Knowledge Management is the 2nd highest reported use of agentic AI across industries and business functions, Curran explained. Inaccuracy is the most cited risk in production above security and compliance.
This is where Guru comes in. Guru offers a governed knowledge layer that can transform scattered, unstructured content into organized, verified, usable knowledge. The platform can enforce policy, permissions, citations, and audit trails—and keep improving over time. Guru can deliver trusted knowledge wherever work happens and power AI tools via MCP.
“AI without a verified knowledge layer is a liability. With one, it's leverage,” Curran said.
Your biggest content consumers will soon be AI agents, said Lacroix. People will only be a fraction of your traffic.
It must be clear to the LLM where to go, who to talk to for each need and action. You have multiple types of consumers; you need multiple rendering formats. Context is essential for accuracy. The solution is delineation and preparation. Build a unified repository of content optimized for AI, Lacroix noted.
“Our mission is to help tech industries gain value from their product knowledge and achieve true innovation in customer support and field service operations,” Lacroix said.
For the full webinar, featuring a more in-depth discussion, Q&A, and more, you can view an archived version of the webinar here.