WorkPeople

Problem

Employees often know what they need, but not who inside the company can help. The product needed to turn fuzzy requests into useful routing without requiring users to understand org charts or internal systems.

Approach

• Framed the core job around people discovery rather than document search.
• Mapped the AI flow from user intent to skills, teams, context, and recommended contacts.
• Designed the product as a lightweight agent layer that could eventually connect to internal directories, collaboration tools, and knowledge bases.

What shipped

• Conversational people-finder experience for routing employees to internal experts.
• Agent workflow model for translating vague employee needs into better people matches.
• Prototype direction for making internal knowledge and expertise easier to access.

Stack

Constraints handled

• Had to work from vague employee requests instead of perfectly tagged internal data.
• Needed a path for enterprise permissions, source visibility, and recommendation guardrails.
• Required a prototype shape that business stakeholders could understand before deep integrations existed.

Next proof to add

The next proof layer should include deployment context, connected data sources, guardrail design, and employee task-completion feedback.