CRM System

Operational customer relationship management for DaisyAI outreach.

Lifecycle Stages

Prospects

People we've identified but never contacted. Structured data only.

Leads

People we've contacted but haven't had a two-way conversation with. Tracks outreach history.

Relationships

People we've had actual conversations with. Rich context in markdown files.

Workflow

Apollo Search → Prospect → (First Email) → Lead → (Response) → Relationship

Prospecting Workflow (Natural Language)

How It Works

You say: "Find 200 CEOs at workers' compensation organizations with 50-500 employees"

Claude does:

1. Translates your request to Apollo API parameters
2. Generates and runs a script
3. Shows you the results
4. You review and approve
5. Converts approved results into CRM prospects

Why Natural Language

The Apollo People Search API has 30+ parameters. Natural language requests get translated dynamically:

• "CEOs" → person_titles[] + person_seniorities[]
• "workers' comp orgs" → industry/organization filters
• "50-500 employees" → organization_num_employees_ranges[]

No rigid abstraction - Scripts generated on the fly, iterate quickly, formalize when ready.

API Reference: See crm/docs/apollo-people-search.md for translation reference.

Core Operations

1. Prospecting - Find contacts via Apollo, deduplicate, add to CRM
2. Campaign Setup - Pull prospects/leads, generate Apollo sequences
3. Execution - Send emails, track engagement, update status
4. Promotion - Graduate leads to relationships when they respond

Canonical Identity

Email address is the unique identifier across all systems.

Safety Rails

• Relationships are NEVER included in automated campaigns
• All campaigns deduplicate against existing contacts before sending
• State transitions are tracked (prospect → lead → relationship)

---

System under construction - structure will evolve as we build