A visual way to move data between factory systems

A manufacturing consultancy kept hitting the same wall: every client ran a different mix of systems — an ERP here, a database there, a warehouse system that only spoke its own dialect — and getting them to share data meant writing custom integration code each time. It was slow, brittle, and didn't scale past a handful of engagements.

The product I helped build replaced that with something a consultant could operate without writing code: a visual canvas where you drag systems onto a board, draw connections between them, and schedule the data to flow on its own. Each client gets their own isolated workspace; connections that used to take an engineer a week to wire up can be assembled and running in an afternoon.

The interesting design choice was where to put the AI. Rather than letting it touch the data, it helps with the part humans find tedious — suggesting how two systems should map to each other when you connect them, so you start from a sensible draft instead of a blank board. The data movement itself stays in plain, reliable plumbing. It's the same instinct that runs through most of what I build: let the model handle the judgement and the suggestions, and keep the parts that have to be exactly right out of its hands.