EliseAI GTM Lead Enrichment Platform

Live demo: https://elise-ai-jade.vercel.app/ — use test@test.com / password123 to log in.

A full-stack tool that automates the inbound lead process for sales teams. Paste in a lead's basic contact info, and the platform automatically enriches it with real market data, scores it, and generates personalized outreach — replacing 20–30 minutes of manual SDR research with a single click. In this demo, it is a shared dashboard between all SDRs. Each SDR can see who is working on which lead.

Note on API limits: All data sources used in this project are free-tier APIs (NewsAPI, Tavily, Gemini, Census, FRED, HUD, etc.). If the enrichment pipeline returns incomplete data or AI generation fails during testing, it is likely due to rate limits or daily quotas being exhausted — not a code failure. In that case, please refer to the prerecorded demo video below.

Presentation video: https://www.youtube.com/watch?v=rPMZCpOIqjo

Demo video: https://www.youtube.com/watch?v=PVk8olu4dtQ

Project Plan: https://docs.google.com/document/d/11IDMz2ijoi90tgvNWJ6cEvrToZJGROP5fus4xL8mW3Y/edit?usp=sharing

Presentation Slides: https://docs.google.com/presentation/d/1BYnk59MUmmtOzam41e9zDzLuHWJ73Jfg_A3bCXrxz18/edit?usp=sharing

Presentation Slides

Tech Stack

Layer	Technology
Frontend	React 19, React Router, Axios, Vite
Backend	FastAPI (Python 3.11), SQLAlchemy ORM, Pydantic v2
Database	PostgreSQL 15
AI	Google Gemini 2.5 Flash (`google-genai`)
Auth	JWT via `python-jose`, bcrypt via `passlib`
Container	Docker + Docker Compose

Features

Lead ingestion — add leads manually or bulk import via CSV/TXT
Automated enrichment — 8 data sources called per lead on trigger
Lead scoring — rule-based 0–100 score with full breakdown
AI outreach generation — draft email, talk track, sales insights via Gemini
On-demand AI SDR Assessment — written fit recommendation grounded in local market data
Pipeline management — Working On, Completed, and per-SDR assignment tracking
View filters — Key Info / Market Data / All views on each lead detail page

Data Sources & API Usage

U.S. Census Bureau API

Endpoint: https://api.census.gov/data/2022/acs/acs5
Pulls: total rental units, population (2017 and 2022 for growth calc), renter ratio, median income
Used for: Rental Market Size score (30 pts), Population Growth score (20 pts), Renter Ratio score (15 pts)
Key: Required — free at https://api.census.gov/data/key_signup.html

FRED — Federal Reserve Economic Data

Endpoint: https://api.stlouisfed.org/fred
Pulls: state-level unemployment rate, GDP growth rate
Used for: Economic Health score (20 pts), market stability signals in insights
Key: Required — free at https://fred.stlouisfed.org/docs/api/api_key.html

HUD Fair Market Rents API

Endpoint: https://www.huduser.gov/hudapi/public/fmr
Pulls: 1BR, 2BR, 3BR fair market rent for the lead's metro area
Used for: Market Insights card (2BR FMR display), market tier classification
Key: Required — free at https://www.huduser.gov/portal/dataset/fmr-api.html
Note: Small Area FMR metros nest data differently — handled in hud_service.py

NewsAPI

Endpoint: https://newsapi.org/v2
Pulls: Recent headlines about the company, recent rental market news for the city
Used for: Company Signals news items, Market News card
Key: Required — free tier at https://newsapi.org/register (100 requests/day)

Wikipedia API

Endpoint: https://en.wikipedia.org/api/rest_v1
Pulls: Introductory paragraph for the lead's city
Used for: City Overview card on lead detail page
Key: None required — public API

OpenCorporates API

Endpoint: https://api.opencorporates.com/v0.4
Pulls: Company registration status, incorporation year, jurisdiction, entity type
Used for: Company Signals (legitimacy verification), Company Signals score
Key: None required for basic usage — public API

Tavily (AI-powered web search)

SDK: tavily-python
Pulls: Company summary from web, Walk Score for the property address, supplemental market news
Used for: Company Signals summary text, Walk Score display, market news fallback
Key: Required — at https://app.tavily.com

Google Gemini 2.5 Flash

SDK: google-genai
Model: gemini-2.5-flash-lite
Key: Required — at https://aistudio.google.com/app/apikey

Gemini AI — Prompts & Usage

Gemini is called in two places during the enrichment pipeline.

1. `generate_insights()` — called during enrichment run

A single Gemini call that returns all four AI-generated outputs at once as a JSON object.

System instruction:

You are a senior B2B sales intelligence analyst specializing in property management software.

Prompt structure:

You are analyzing a lead for EliseAI's sales team. EliseAI sells AI-powered property management software.

LEAD: company, location, score
MARKET DATA: rental units, population, population growth, renter ratio, unemployment, GDP, 2BR FMR
COMPANY: summary, status, incorporated, recent news
CITY: overview, market news, rental news

Return ONLY a JSON object:
{
  "sales_insights": {
    "Market Overview": "2-3 sentences on the rental market opportunity",
    "Economic Health": "2-3 sentences on economic indicators",
    "Property Context": "2-3 sentences on the property management landscape",
    "Company Signals": "2-3 sentences on portfolio scale, recent activity, and pain points",
    "Market News": "2-3 sentences summarizing relevant recent news"
  },
  "talk_track": [
    "Opening hook referencing a specific local market stat",
    "Value proposition tied to their market context",
    "Call to action with urgency driver"
  ],
  "email_subject": "Compelling subject line under 60 chars",
  "draft_email": "Professional 3-paragraph cold email..."
}

Outputs produced:

sales_insights → displayed across the Market Data view cards
talk_track → SDR Talk Track card (Key Info view)
email_subject + draft_email → Draft Outreach Email card (Key Info view)

2. `generate_ai_summary()` — on-demand, triggered by SDR clicking "AI Summary"

A separate Gemini call that writes a personalized SDR fit assessment grounded in local market data.

Prompt structure:

You are a senior SDR at EliseAI...

Your analysis must be grounded in the specific market data for {city}, {state} — not just
the company's national reputation. If the company is large but this location's market data
is weak, say so plainly and reflect it in your prioritization recommendation.

LEAD DATA: company, market, score, rental units (local, not portfolio), population growth,
renter ratio, unemployment, 2BR FMR, company info, recent news, AI insights

Cover in order:
1. Brief note on the company's scale
2. Honest assessment of the local market using the data above
3. Whether this location is a strong or secondary outreach target
4. Clear prioritization recommendation

Return plain text only, no bullet points, no headers.

Output: AI SDR Assessment card (Key Info view), persisted to the database.

Important: The numeric lead score (0–100) is entirely rule-based. Gemini is used for text generation only — no AI involvement in scoring.

Running Locally with Docker (Recommended)

Prerequisites: Docker Desktop

1. Clone the repo

git clone <repo-url>
cd elise-gtm

2. Set up your environment variables

cp .env.example .env

Open .env and fill in your API keys. See the API Keys section below. The DATABASE_URL line can be left as-is — Docker sets it automatically.

Also set a JWT secret:

# Run this to generate a secure secret and paste it into .env
python -c "import secrets; print(secrets.token_hex(32))"

3. Start everything

docker-compose up --build

This starts three containers:

PostgreSQL on port 5432 — database, auto-creates tables on first run
Backend on port 8000 — FastAPI server
Frontend on port 5173 — React app served by nginx

4. Open the app

http://localhost:5173

Register an account and start adding leads.

5. Stopping

docker-compose down          # stop containers, keep database
docker-compose down -v       # stop containers + wipe database

Running Locally Without Docker

Prerequisites

Python 3.11+
Node.js 20+
PostgreSQL 15+ running locally

Backend

cd backend
python -m venv venv
source venv/bin/activate      # Windows: venv\Scripts\activate
pip install -r requirements.txt

cp ../.env.example .env       # fill in your API keys and DATABASE_URL

Create the database in PostgreSQL:

CREATE DATABASE eliseai;

Start the server (tables are created automatically on startup):

uvicorn main:app --reload

Backend runs at http://localhost:8000. API docs at http://localhost:8000/docs.

Frontend

cd frontend
npm install
npm run dev

Frontend runs at http://localhost:5173.

Getting API Keys

Service	Free?	Link
Google Gemini	Yes (generous free tier)	https://aistudio.google.com/app/apikey
U.S. Census	Yes	https://api.census.gov/data/key_signup.html
FRED	Yes	https://fred.stlouisfed.org/docs/api/api_key.html
HUD	Yes	https://www.huduser.gov/portal/dataset/fmr-api.html
NewsAPI	Yes (100 req/day free)	https://newsapi.org/register
Tavily	Yes (small free tier)	https://app.tavily.com
Wikipedia	No key needed	—
OpenCorporates	No key needed	—

Project Structure

elise-gtm/
├── backend/
│   ├── main.py                  # FastAPI app, CORS, router registration
│   ├── database.py              # SQLAlchemy engine + session
│   ├── schemas.py               # Pydantic request/response models
│   ├── scoring.py               # Rule-based lead scoring logic
│   ├── models/
│   │   ├── lead.py              # Lead + LeadEnrichment ORM models
│   │   └── user.py              # User ORM model
│   ├── routers/
│   │   ├── auth.py              # Register, login, JWT auth dependency
│   │   └── leads.py            # All lead CRUD + enrichment endpoints
│   └── services/
│       ├── pipeline.py          # Orchestrates all enrichment services per lead
│       ├── gemini_service.py    # Gemini AI — insights, email, talk track, summary
│       ├── census_service.py    # U.S. Census ACS data
│       ├── fred_service.py      # FRED economic data
│       ├── hud_service.py       # HUD Fair Market Rents
│       ├── news_service.py      # NewsAPI headlines
│       ├── tavily_service.py    # Tavily web search + company summary
│       ├── walkscore_service.py # Walk Score via Tavily
│       └── opencorporates_service.py  # Company registration data
├── frontend/
│   ├── src/
│   │   ├── api/client.js        # Axios instance with JWT interceptor
│   │   ├── pages/
│   │   │   ├── Dashboard.jsx    # Main lead list, pipeline trigger
│   │   │   ├── LeadDetail.jsx   # Enriched lead profile + AI outputs
│   │   │   ├── WorkingOn.jsx    # In-progress leads
│   │   │   ├── Completed.jsx    # Completed leads
│   │   │   ├── Login.jsx
│   │   │   └── Register.jsx
│   │   └── components/
│   │       ├── Sidebar.jsx
│   │       ├── StatusBadge.jsx
│   │       ├── ScoreBadge.jsx
│   │       ├── AddLeadModal.jsx
│   │       ├── ConfirmModal.jsx
│   │       └── Toast.jsx
│   ├── Dockerfile
│   └── nginx.conf
├── backend/Dockerfile
├── docker-compose.yml
├── .env.example
└── .gitignore

Lead Scoring Logic

Scores are calculated in backend/scoring.py using only rule-based logic — no AI.

Category	Max	Source	Tiers
Rental Market Size	30	Census	200k+ → 30, 100k+ → 22, 50k+ → 15, else → 5
Population Growth (5yr)	20	Census	10%+ → 20, 5%+ → 15, 0%+ → 8, declining → 0
Economic Health	20	FRED	<4% unemployment → 20, 4–6% → 13, >6% → 5
Renter Ratio	15	Census	50%+ → 15, 40–50% → 10, <40% → 5
Company Signals	15	Tavily + News + OpenCorporates	Signals found → 13, none → 5

Priority tiers: 80–100 = High · 60–79 = Medium · 0–59 = Low

API Endpoints

Method	Path	Description
POST	`/auth/register`	Create account
POST	`/auth/login`	Login, returns JWT
GET	`/leads`	List all leads
POST	`/leads`	Create a lead
GET	`/leads/{id}`	Get lead + enrichment detail
PATCH	`/leads/{id}/status`	Update SDR status
PATCH	`/leads/{id}/assign`	Assign lead to current SDR
PATCH	`/leads/{id}/email`	Save edited draft email
POST	`/leads/{id}/analyze`	Run enrichment on a single lead
POST	`/leads/{id}/ai-summary`	Generate on-demand AI SDR assessment
DELETE	`/leads/{id}`	Delete lead and all enrichment data

Interactive API docs available at http://localhost:8000/docs when the backend is running.

Name		Name	Last commit message	Last commit date
Latest commit History 18 Commits
.claude		.claude
elise-gtm		elise-gtm
.gitattributes		.gitattributes
README.md		README.md
claude.md		claude.md
sampleData.txt		sampleData.txt

Folders and files

Latest commit

History

Repository files navigation

EliseAI GTM Lead Enrichment Platform

Presentation Slides

Tech Stack

Features

Data Sources & API Usage

U.S. Census Bureau API

FRED — Federal Reserve Economic Data

HUD Fair Market Rents API

NewsAPI

Wikipedia API

OpenCorporates API

Tavily (AI-powered web search)

Google Gemini 2.5 Flash

Gemini AI — Prompts & Usage

1. generate_insights() — called during enrichment run

2. generate_ai_summary() — on-demand, triggered by SDR clicking "AI Summary"

Running Locally with Docker (Recommended)

1. Clone the repo

2. Set up your environment variables

3. Start everything

4. Open the app

5. Stopping

Running Locally Without Docker

Prerequisites

Backend

Frontend

Getting API Keys

Project Structure

Lead Scoring Logic

API Endpoints

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

1. `generate_insights()` — called during enrichment run

2. `generate_ai_summary()` — on-demand, triggered by SDR clicking "AI Summary"

Packages