Schneider Electric: AI+Power

Context & problem

As the sole designer and AI product intern on this project, I was responsible for building a platform prototype from scratch: an integrated system pairing automated inspection with robot dispatch across the company's factories. A public demo date was set before the product existed, so my role centered on building out the frontend as the clearest way to make the product intent legible.

Almost all of my input was verbal, drawn from interviewing engineers, R&D colleagues, and on-duty factory workers, and from working directly alongside the team lead. Translating that unstructured domain expertise into a design driven by a product mindset was the key task. The problems identified were as follows:

• Maintenance should be preventive, not reactive

  Factory inspection runs on fixed intervals and human judgment, so early signs of degradation are easy to miss until they become failures.

• Domain knowledge is scattered

  Maintenance expertise sits in separate documents and in individual engineers' heads, never consolidated into structured, shared knowledge. Because none of it is a reusable asset, root cause diagnosis stays dependent on a few experienced people and cannot scale as equipment and data grow.

• Readily available field response is crucial

  Inspection tasks like photo capture, thermal measurement, and status confirmation all require staff on site, yet faults can occur at any hour. The gap between detecting a problem and reaching it is bounded by travel time, and it widens exactly when conditions are most hazardous.