ActionDesk

• 

  pipeline

• 

  email_content_ai

• 

  ai_priority_assesment

• 

  backend-uvicorn running

Implementation Notes

For this prototype, the frontend and backend were developed in parallel to prioritize speed and clarity of the user experience.

The project is fully integrated it used mock data in backend instead of using API calls to get data from Lofty.

The backend is fully implemented, including: action prioritization logic explainability (“why this matters”) draft generation approval and simulated send workflows

This separation allowed us to focus on delivering a strong AI-native experience while still building a functional backend system.

The full application can be run locally using Docker:

docker compose up --build

This spins up the backend services and environment required to support the system.

In a production setting, the frontend would connect directly to the backend APIs and real CRM data sources (e.g., Lofty), replacing the current mock data layer.

Built With

• chatgpt
• fastapi
• figma
• python
• react