CRM Case Retrieval and Analytics Agents

I implemented CRM agents and analytics products around resolved support cases, turning historical service data into a searchable and analyzable knowledge layer.

Capabilities

• Semantic retrieval over resolved CRM cases.
• Full case-detail loading after retrieval.
• Relative-date filters and Shamsi metadata repair.
• Service and case quality scoring.
• Expert feedback management.
• Aggregate statistics over case volume, units, status, customer groups, and time windows.
• Topic/category detection, monthly trends, and BERTopic-style clustering named by local LLMs.

Outcome

The system helps support teams retrieve relevant past cases, understand recurring issues, score service quality, and identify patterns across customer and operational data.

Technologies

Python, Microsoft CRM/Dynamics 365, SQLite aggregates, local LLMs, semantic search, clustering, topic detection, and analytics dashboards.