An agentic AI system that evaluates B2B sales leads and generates personalized outreach messages using a locally hosted LLM.

• Scoring Agent: evaluates lead quality
• Messaging Agent: generates outreach content
• Orchestrator: controls decision flow and thresholds

• Agent separation (decision vs execution)
• Configurable score thresholds
• Explainable decision traces
• Local LLM inference (Ollama + Mistral)
• CSV-driven, testable workflow

• Python
• Ollama (Mistral-Instruct)
• Pandas

1. Load leads from CSV
2. Score and prioritize leads
3. Generate outreach for qualified leads
4. Save results with audit trail

python outreach_generator.py