AI for understanding, code for execution
The model interprets conversational requests and produces structured decisions. Deterministic services enforce business rules and perform operations.
An AI scheduling assistant that turns conversations into confirmed appointments.
Illustrative conversation and workflow — not a live product capture.
People request vague times, change their minds, reschedule existing appointments, and expect immediate responses.
A language model can understand these requests — but understanding isn't enough. An actual booking requires availability checks, business rules, persistent state, and reliable execution.
I built TamiTor, a conversational scheduling system that connects WhatsApp, AI-powered intent understanding, and calendar operations through structured backend workflows.
Customers interact naturally. Behind the scenes, the application manages the process from request to confirmation.
The model interprets conversational requests and produces structured decisions. Deterministic services enforce business rules and perform operations.
Temporal manages multi-step processes, including waiting for user responses, retries, and long-running workflows.
WhatsApp messaging, Google Calendar OAuth, PostgreSQL, Redis, and Python backend services work together to maintain application state.
Booking flows incorporate validation, error handling, state tracking, and controls against duplicate operations.