AI-GTM

Turn market signals into approved GTM work your team can execute.

A strategy workbench for research, planning, approval and execution. It combines agent workflows with operational tools instead of an ungrounded chatbot, and nothing external happens without an explicit approval.

_{The project page, with the strategy workflow and its approval gate.}

Contents

1. Strategy workflow
2. Components
3. Setup
4. Run
5. Design principles
6. Security
7. Contributing
8. License

Strategy workflow

flowchart LR
  A[Market signal] --> B[Research brief] --> C[Strategist agent]
  C --> D[Approved campaign plan] --> E[Tool execution] --> F[Outcome feedback]
  F --> B

Loading

Components

Path	What it is
strategist_agent/	Strategy workflows and tools
hunter_agent/	Lead-hunting Flask app (Exa search, Supabase storage)
furniture-agent/	Furniture lead agent: Streamlit UI, Telegram bot, lead scoring
invoice_system/	Invoice and proforma automation (Flask, PDF generation, scheduled e-mail)
dashboard/	Combined dashboard over the hunter and invoice agents
unified_agent.py	One agent (ARIA) that combines the strategist, hunter and invoice skills
telegram_app.py	Telegram front end for the unified agent
plans/, docs/	Design plans and specs

Setup

Requires Python 3.10 or newer.

git clone https://github.com/boss477/AI-GTM.git
cd AI-GTM
python -m venv .venv
.venv/Scripts/activate        # Windows; on macOS/Linux: source .venv/bin/activate
pip install -r hunter_agent/requirements.txt -r invoice_system/requirements.txt -r furniture-agent/requirements.txt
cp .env.example .env          # then fill in your own keys

All credentials come from the environment. .env.example lists every variable; .env is git-ignored.

Run

python dashboard/app.py          # dashboard on http://127.0.0.1:5000
python hunter_agent/app.py       # lead hunter (Flask)
python invoice_system/app.py     # invoice system (Flask)
streamlit run furniture-agent/app.py
python telegram_app.py           # Telegram bot for the unified agent

The Flask debugger is off by default. Set FLASK_DEBUG=1 locally if you need it, and never on a reachable host.

Design principles

• Agent reasoning is visible and reviewable.
• Approval gates are explicit before external actions.
• Tool boundaries and credentials are documented.
• Research stays connected to an executable next step.

See DESIGN.md for the product design.

Security

Never commit credentials. See SECURITY.md for how secrets are handled and how to report a vulnerability.

Contributing

Read CONTRIBUTING.md and the Code of Conduct first.

License

MIT © boss477. Agno and LangGraph are installed from PyPI under their own licenses; this repository no longer bundles them.