Sales Intelligence Research Automation

A multi-agent AI crew built with CrewAI that researches a prospect before a sales call. Give it an email address and it pulls together company research, CRM history, past email threads, and delivers a ready-to-use sales intelligence report.

Want to try it right now? Open demo.ipynb and run it cell-by-cell -- no setup beyond API keys required.

What it does

Given a prospect's email, four agents work sequentially:

Agent	What it does	Tools
Company Research Specialist	Researches the prospect and their company online	Serper (web search)
HubSpot Data Analyst	Pulls CRM records, deal history, and past interactions	HubSpot API
Email History Analyst	Reviews previous email threads for context and tone	Gmail API
Sales Intelligence Report Writer	Synthesizes everything into an actionable briefing	Context from all above

The final output is a markdown report with an executive summary, prospect profile, recommended approach, conversation starters, and objection handling.

Quick start

Option A: Interactive notebook

The fastest way to run the crew end-to-end. Everything is self-contained -- agent definitions, task configs, and execution are all inline so you can read, tweak, and run without touching the project structure.

jupyter notebook demo.ipynb

API keys are prompted via getpass so nothing gets saved in the notebook.

Option B: CLI

pip install uv
crewai install
crewai run

This uses the YAML configs under src/sales_intelligence_research_automation/config/ and reads API keys from a .env file. Copy .env.example or create .env with:

OPENAI_API_KEY=sk-...
SERPER_API_KEY=...
NVIDIA_NIM_API_KEY=...

Project structure

├── demo.ipynb                          # Self-contained notebook demo
├── src/sales_intelligence_research_automation/
│   ├── config/
│   │   ├── agents.yaml                 # Agent roles, goals, backstories
│   │   └── tasks.yaml                  # Task descriptions and expected outputs
│   ├── crew.py                         # Crew orchestration class
│   └── main.py                         # CLI entry point
├── knowledge/                          # Knowledge base resources
└── .env                                # API keys (not committed)

Customizing

• Agents: Edit config/agents.yaml to change roles, goals, or backstories
• Tasks: Edit config/tasks.yaml to change task descriptions or expected outputs
• Tools & logic: Edit crew.py to swap LLMs, add tools, or change the process type
• Inputs: Edit main.py to change the default prospect email

Requirements

• Python >= 3.10, < 3.14
• API keys for OpenAI, Serper, and NVIDIA NIM
• HubSpot and Gmail integrations require CrewAI Apps platform authentication

Support

• CrewAI Documentation
• GitHub
• Discord