CAVAS AI Assistant

CAVAS AI Assistant is a configurable AI workflow system designed for business processes that require structured intake, analysis, routing, review, and controlled handoff.

The system combines AI-assisted interpretation with deterministic business rules, explicit workflow states, persistence, auditability, and human approval before important actions.

It can be adapted to workflows such as lead qualification, service routing, operational triage, opportunity review, internal decision support, and other processes where unstructured inputs need to be converted into structured, reviewable actions.

The architecture is designed to keep AI within clear boundaries: deterministic logic handles hard constraints and workflow control, while AI supports interpretation, classification, summarisation, and structured recommendations.

The current implementation uses Python, OpenAI Platform, SQLite, Telegram, pytest, and GitHub Actions, with a strong focus on reliability, testing, and fail-safe behaviour.