design-doc-mermaid

```mermaid
%% Design Document for Sortd Sales Pipeline Integration
%% Generated by AI for design-doc-mermaid skill

%% Project Overview
flowchart TD
    A[Project Overview] --> B[Objective: Integrate Sortd with CRM for sales pipeline visibility]
    A --> C[Scope: Real-time sync between Gmail/Sortd and HubSpot CRM]
    A --> D[Stakeholders: Sales Team, Marketing, Customer Support]

    %% System Architecture
    B --> E[Architecture Diagram]
    E --> F[Gmail/Sortd Layer] --> G[API Gateway]
    G --> H[HubSpot CRM Layer]
    H --> I[Salesforce Data Warehouse]
    
    %% Data Flow
    F --> J[Email Parsing Engine]
    J --> K[Lead Extraction: Extract lead data from emails]
    K --> L[CRM Sync: Push leads to HubSpot]
    L --> M[Pipeline Update: Update deal stages in real-time]
    M --> N[Notification System: Alert sales reps on new leads]
    
    %% Components
    subgraph Components
        O[Sortd Gmail Add-on] --> P[Email Tagging Rules]
        P --> Q[Lead Scoring Model]
        Q --> R[CRM Field Mapping]
        R --> S[API Rate Limiter]
    end
    
    %% Next Steps
    T[Next Steps] --> U[Phase 1: MVP with 50 sales reps for 30 days]
    U --> V[Phase 2: Expand to marketing team for lead routing]
    V --> W[Phase 3: Add AI-powered deal risk assessment]
    W --> X[Phase 4: Mobile notifications for urgent leads]

    %% Comments
    %% Key Decision: Using Sortd's existing Gmail integration to avoid custom email parsing
    %% Assumption: HubSpot CRM API has rate limits of 100 calls/minute
    %% Risk: Email parsing may miss leads if subject lines are unclear
```

### Design Document Summary

This Mermaid diagram outlines the technical design for integrating **Sortd** (a Gmail-based task management tool) with **HubSpot CRM** to create a seamless sales pipeline visibility system. The architecture leverages Sortd’s existing Gmail integration to parse incoming emails, extract lead information, and sync it with HubSpot in real-time. The system includes an email parsing engine that identifies lead data (e.g., contact info, deal stages) from email threads, subject lines, and attachments.

The **data flow** starts with emails landing in Sortd’s kanban-style boards, where the parsing engine extracts structured lead data. This data is then pushed to HubSpot via API, updating deal stages and notifying sales reps of new opportunities. The **components** include Sortd’s Gmail add-on, email tagging rules, a lead scoring model, and CRM field mapping to ensure consistency. The diagram also highlights **next steps**, divided into four phases: MVP testing with 50 sales reps, expansion to marketing, addition of AI-powered deal risk assessment, and mobile notifications for urgent leads.

Key decisions include using Sortd’s existing Gmail integration to avoid building custom email parsing logic and assuming HubSpot’s API rate limits (100 calls/minute) will suffice for initial testing. A potential risk is that unclear email subject lines or unstructured data may reduce parsing accuracy, which will be addressed in Phase 2 with AI-powered NLP improvements.

This design ensures sales teams can track leads directly from their Gmail inbox without switching tools, improving efficiency and reducing response times. The integration also provides marketing with real-time visibility into lead sources and conversion rates, enabling data-driven campaign adjustments.