DealIQ — Intelligent Sales Workspace

"It remembers your deals so you can focus on what happens next."

DealIQ is an AI-powered sales workspace prototype that uses long-term memory to help sales reps retain critical deal context, prepare for upcoming calls, generate targeted follow-ups, and contrast memory-informed intelligence with generic LLM responses.

---

Architecture Overview

                      DealIQ (Frontend - React + Vite + TS)
                                      │
                                      ▼
                      FastAPI Backend (app/main.py)
                                      │
        ┌─────────────────────────────┼─────────────────────────────┐
        ▼                             ▼                             ▼
SQLite Database               Hindsight Cloud                Groq LLM
(SQLAlchemy Models)           (Vectorize Memory)      (openai/gpt-oss-120b)
- Deals                       - Banks per deal        - Fallback: qwen/qwen3-32b
- Interactions                - Retain / Recall       - Structured prompting

• Frontend: React 19, Vite, TypeScript, Lucide Icons, Vanilla CSS Design System with rich typography (Inter + JetBrains Mono) and responsive layout.
• Backend: Python FastAPI, Pydantic v2, SQLAlchemy ORM, SQLite for prototype development (swappable to PostgreSQL).
• AI Engine: Groq API (openai/gpt-oss-120b primary with qwen/qwen3-32b fallback).
• Long-Term Memory: Hindsight Cloud (https://api.hindsight.vectorize.io).
• Currency: Indian Rupees (₹) formatted with the Indian numbering system everywhere.

---

Directory Structure

DealIQ/
├── backend/
│   ├── app/
│   │   ├── api/
│   │   │   ├── deals.py
│   │   │   ├── interactions.py
│   │   │   ├── ai.py
│   │   │   └── health.py
│   │   ├── database/
│   │   │   ├── database.py
│   │   │   └── seed.py
│   │   ├── models/
│   │   │   ├── deal.py
│   │   │   └── interaction.py
│   │   ├── schemas/
│   │   │   ├── deal.py
│   │   │   ├── interaction.py
│   │   │   └── ai.py
│   │   ├── services/
│   │   │   ├── deal_service.py
│   │   │   ├── memory_service.py
│   │   │   ├── llm_service.py
│   │   │   └── ai_service.py
│   │   ├── config.py
│   │   └── main.py
│   ├── tests/
│   │   └── test_api.py
│   ├── requirements.txt
│   └── .env.example
├── frontend/
│   ├── src/
│   │   ├── api/
│   │   │   ├── client.ts
│   │   │   ├── deals.ts
│   │   │   ├── interactions.ts
│   │   │   └── ai.ts
│   │   ├── components/
│   │   │   ├── common/
│   │   │   ├── deals/
│   │   │   ├── layout/
│   │   │   └── ai/
│   │   ├── pages/
│   │   │   ├── Dashboard.tsx
│   │   │   ├── Deals.tsx
│   │   │   ├── DealDetails.tsx
│   │   │   ├── Copilot.tsx
│   │   │   └── MemoryCompare.tsx
│   │   ├── types/
│   │   ├── utils/
│   │   ├── App.tsx
│   │   └── main.tsx
│   ├── package.json
│   ├── vite.config.ts
│   └── .env.example
└── README.md

---

Quickstart

1. Backend Setup

cd backend
python -m venv venv
venv\Scripts\activate      # Windows (or source venv/bin/activate on Linux/Mac)
pip install -r requirements.txt
cp .env.example .env       # Configure GROQ_API_KEY & HINDSIGHT_API_KEY
python -m app.database.seed
python -m uvicorn app.main:app --host 0.0.0.0 --port 8000

2. Frontend Setup

cd frontend
npm install
npm run dev                # Running on http://localhost:3000

3. Run Backend Tests

cd backend
pytest tests/test_api.py -v

---

10-Step End-to-End Workflow Demonstration

1. Dashboard / Deals: View seeded deals in Indian Rupees (₹15,00,000, ₹18,50,000, etc.).
2. Create Deal: Click + New Deal, enter details (Company: TechNova Solutions, Deal: AI Automation Platform, Value: ₹15,00,000, Stage: Discovery, Owner: Rahul, Next Call: 30 September).
3. Open Deal: Access /deals/:id to inspect deal metrics and previous interaction timeline.
4. Add Interaction & Save to Memory: Enter call notes and click Save to Memory to retain customer objections and requirements in Hindsight Cloud.
5. AI Copilot: Navigate to /copilot, select the deal, and ask questions (e.g., "What are the customer's main concerns?").
6. Structured AI Response: Review structured sections: SUMMARY, KEY FINDINGS, CUSTOMER CONCERNS, RISKS, RECOMMENDED NEXT STEPS.
7. Prepare for Next Call: Click Prepare for Next Call to generate a comprehensive sales briefing with call objectives, questions to ask, talking points, and next steps.
8. Follow-up Email: Click Generate Follow-up Email to draft an editable email tailored to recent discussions with 1-click [Copy] and [Regenerate].
9. Accumulated Memory: Record additional interactions; observe how the AI synthesizes historical context with newly provided facts.
10. Memory Compare: Navigate to /memory-compare and run side-by-side comparisons of Memory ON (context-aware intelligence) vs. Memory OFF (generic responses without deal history).