An HR-Specialist AI Handling Labor & Personnel Inquiries

Within a railway group company, we supported the consolidation and replacement of HR and labor consultation work — previously handled in fragmented fashion by multiple staff and locked in individuals — with a "labor-response AI (clone)" embodying the expertise of a senior HR specialist. We trained the AI on the extensive history of past interactions and manuals, and built a system that automatically responded to employee inquiries. We designed an operating model in which the AI handled all baseline responses, with staff supporting only the complex cases the AI could not address. This eliminated variability in responses across staff, substantially reduced workload, and established a cycle in which the AI's response quality improved over time as it continued to learn from new interactions.

[Challenges]

• The internal labor and HR consultation desk was staffed by multiple people (creating variability in workload allocation and responses, and fragmented knowledge).
• Coordination and email handling created heavy workload (multiple staff felt the operational burden and could not focus on their core work).
• Know-how was locked in individuals (with multiple staff handling different cases, response know-how was also fragmented).

[Results]

• Deployed a labor-response AI that replicated the HR specialist (absorbing interactions and retaining the full response record).
• Workload was reduced by making AI the primary respondent with humans in a supporting role (the clone handled baseline responses; staff stepped in only when needed. Workload was reduced by 25%).
• Building on the response history, the AI absorbed new learnings and grew continuously (response records were preserved for review, and the clone evolved based on ground-truth data).