Glide — Adaptive Case Management Engine

The problem

A repair shop takes in a damaged car. Over the next six weeks the job touches an estimate, an insurer’s assessor, a parts order, the work itself, delivery, an electronic invoice, and a bank transfer that arrives — or does not. Two of those steps live in a billing package, two more in an accounting system, and the rest live in someone’s head.

Ask where the job stands and there is no answer, because there is no single place the answer lives. Every system holds a fragment; none of them holds the case.

What I built

Glide is the engine that holds the case: one record spanning every system involved, with one state anybody can read. A workflow is described as data — what may happen, and what must be true in order to close — rather than as a numbered sequence. A company that starts invoicing before delivery is a configuration change, not a rewrite.

Connectors watch the systems that already exist and report what happened there. The engine decides, and nothing counts as done because an API returned success — only because the change was seen in the other system. The status a user reads is derived from what actually happened, so there is no field left to disagree with reality.

Where it stands

Built and tested end to end: a workflow imported and published, two systems connected, a case carried through to close. The rules that must never break are written down beside the code that enforces them.