AI SRM

The best sales representative matching tool of all time!

This project is an MVP for an AI-powered app that recommends the best sales representative for a given client and product. It uses real CRM data and machine learning to help sales teams match opportunities with the right agent, every time.

🚀 Features

• Predicts the best sales agent for a given client-product combination
• Top 3 recommendations with probability/confidence
• Explainability: Shows why an agent is recommended (past deals, similar features)
• Streamlit web app: Easy-to-use interface for demo and experimentation
• Automatic build and deploy to GCP: Github PRs trigger code quality checks and release tags trigger a new deployment.

⚡️ Getting Started

1. Clone the repository

   • Get the source code locally:

     git clone https://github.com/MatthieuScarset/aisrm.git
     cd aisrm

2. Set up your Python environment

   • Make sure you have pyenv and pyenv-virtualenv installed.

   • Create your environment file:

     cp .env.example .env

   • Create and activate a new virtual environment:

     make init

   • Run make help to learn more about all available commands

👨‍💻 Develop

1. Data pipeline

   • Convert raw to processed data

     # Get the raw data.
     # Clean and compiled into a new raw dataset.
     make data

2. Model training

   • Generate model for development purpose:

     # Build and save model files.
     make model

   • Commit a model version if you're happy with the results:

     cp -R models/dev-123456 models/v999
     git add modelds/v999
     git commit -m "feat: New model v999"

3. Application development

   • You need to start both the backend and frontend applications:

     # Start the backend application in dev mode:
     make api_dev
     # Start the frontend application in dev mode:
     make app_dev

🏗️ Deploy

1. Docker images

   • Alternatively, you can test containers separately:

     # Build and start services
     make up
     # Stop services.
     make down

   • Alternatively, you can test containers separately:

     # Rebuild the service
     make api_build
     # Start the service
     make api_start
     # Test the API routes in your browser:
     # http://localhost:8500
     make test_api
     # Stop when you are done with your work.
     make api_stop

     # Rebuild the service
     make app_build
     # Start the service
     make app_start
     # Test the frontend in your browser:
     # http://localhost:8501
     make test_app
     # Stop when you are done with your work.
     make app_stop

2. Cloud build

   • Option 1: Manual deployment

     # Build, push and publish the backend to GCP 
     make cloud_pipeline_api
     # Build, push and publish the fronted to GCP 
     make cloud_pipeline_app

   • Option 2: Create a release tag

     # Check latest tags
     git fetch --tags
     git tag
     # Tag must be named release-{app|api}-{0.0.x}
     # @see .github/workflows/deploy.yml -> on.push.tag
     git tag -a release-app-0.0.1 -m "Describe this version"
     git push --tags
     # Watch the Github Actions run
     # @see https://github.com/MatthieuScarset/aisrm/actions/workflows/deploy.yml

   • View your running applications:

     # Get the deployed API URL
     make cloud_get_api_url
     # Get the deployed APP URL
     make cloud_get_app_url

Any questions? Read the doc.

---

Made with ❤️ by the AISRM team