Socially Interactive Agents for Preserving and Transferring Tacit Knowledge in Organizations

AI agents can help preserve knowledge that employees struggle to put into words: The paper demonstrates that Socially Interactive Agents (SIAs) can extract tacit knowledge—the deep, experience-based insights that experts often can't easily articulate—through empathic, natural-language conversations, addressing a significant challenge organizations face when experienced workers leave.

Building trust is essential for AI-facilitated knowledge sharing: The research emphasizes that employees are hesitant to share tacit knowledge unless they have assurance of confidentiality and appreciation, meaning organizations must design these AI systems with trust-building features rather than treating them as simple extraction tools.

AI can scale knowledge transfer beyond what human facilitators alone can achieve: While traditional human-led knowledge transfer is effective, it's resource-intensive and doesn't scale well; SIAs offer a way to provide personalized guidance and structured knowledge capture to many employees simultaneously, particularly valuable during onboarding and when retaining institutional knowledge.

Multiple AI technologies work together to make this approach work: The system relies on Large Language Models for natural dialogue, Retrieval-Augmented Generation to connect individual insights to broader organizational knowledge, and Chain-of-Thought prompting to guide employees through structured reflection—no single technology solves the problem alone.

Practical applications extend across critical organizational functions: The paper identifies concrete use cases including employee onboarding (personalized guidance) and knowledge retention (structured interviews with departing experts), suggesting immediate value for organizations concerned about knowledge loss.