Large organizations often struggle with knowledge silos, where critical expertise is trapped with individuals or isolated teams. Employees waste time searching for internal experts or unknowingly duplicating work that already exists elsewhere. While document-based knowledge systems like wikis help, they fail to capture the "who knows what" problem—leaving valuable tacit knowledge untapped.

How It Could Work

One approach could be a dynamic, searchable directory of employee expertise, similar to Bridgewater's "baseball cards" concept. Instead of relying on static HR profiles, this system could:

• Auto-populate expertise data from existing tools like project management software, email/chat keywords, or code repositories.
• Infer tacit knowledge using machine learning, such as analyzing frequent collaborators or niche terms in documents.
• Visualize expertise networks to show how skills and knowledge cluster across teams.

Integration with workplace tools like Slack, Microsoft 365, or HRIS could minimize manual upkeep. A simpler version might start with employee self-tagging and peer endorsements.

Potential Benefits and Challenges

For large enterprises—especially those with dispersed teams like tech firms or consultancies—this could reduce redundant work and speed up decision-making. Employees might benefit from greater visibility for their skills, while managers could optimize team utilization. However, adoption could face hurdles:

• Privacy concerns could be addressed by anonymizing search results by default and allowing opt-in visibility.
• Data overload might be mitigated by AI highlighting high-signal expertise (e.g., "Reviewed 5+ ML patents last quarter").
• Tool fatigue could be minimized by integrating with existing systems rather than requiring a standalone platform.

Execution Strategies

A phased rollout might help validate the concept:

1. MVP: A basic web app with self-tagged skills and project history, tested with a pilot department.
2. Phase 2: Integration with one core system (e.g., Microsoft Graph API) to auto-generate expertise hints.
3. Phase 3: Expanded integrations (Slack, Jira) and ML-driven recommendations for finding experts.

Unlike existing tools that focus on static documents or Q&A, this approach would prioritize connecting people—making it easier to find not just answers, but the right person to ask.