Predictive B2B CRM and AI funnel analytics for a distributor
A unified CRM with customer 360, ML scoring for leads and deals, cross/upsell recommendations and data marts for sales and revenue teams.
Delivered functionality
360° profile, pipeline, tasks; lead scoring and stage win rates.
SKU/promo hints; manager/region reports; AR and activity alerts.
RAG assistant: call summaries and Q&A on history (ACL).
Business outcomes
Qualified lead-to-deal +8–14%; key-account churn −6–10% (pilot).
Sales leadership reporting time −35–50%.
Highlights
- Customer 360: orders, payments, AR, support tickets and meetings in one relationship graph.
- ML: win probability, churn risk, SKU recommendations from co-purchase patterns and seasonality.
- ClickHouse marts + dbt; funnel dashboards and manager attribution; RFM segmentation.
- Integrations: ERP, telephony, email (ETL), vector search over conversation history (RAG assistant).
Stack
Python 3.12
FastAPI
PostgreSQL
ClickHouse
Apache Airflow
dbt
Redis
Qdrant (vector search)
LangChain / LLM API (RAG)
scikit-learn / XGBoost
React + TypeScript
Metabase / Grafana
Want a similar architecture for you ?