Rebuilding an inspection platform with AI

Avitas started as a GE venture, later acquired by Baker Hughes, with a big goal: use drones, robotics, and AI to make industrial inspections safer and faster across oil & gas, power, and energy.

I was the solo designer for the first 1.5 years. The product had been built by engineers and product, and inspectors struggled to use it. I defined the high-level flows, aligned the team on goals, and shaped it into one consistent, scalable platform. Instead of spending long hours building custom demos for every customer, the team could now focus on one unified product that powered every demo too. Along the way, sales prospects grew 35%.

Stakeholder interviews

When I started, the six PMs weren’t really talking to each other. Each managed a different industry and set of customers, so everyone saw only their slice of the product. I interviewed them one by one to piece together the bigger picture: what the platform was for, who it was really serving, and how each feature connected to the others.

Using this feedback, I created a concept map that aligned business goals at the company level, defined target personas, and visualized how existing features and workflows were interconnected.

Customer interviews and site visit

During site visits, I spoke with subsea and land inspectors to understand their workflow, needs, and pain points. Here’s what I found:

Journey map

I turned what I learned into a journey map of the full inspection, from setting up an asset to the follow-up visit. For every need along the way, I checked it against the roadmap: already built, missing from the MVP, or beyond the short-term plan. Shared with customers and the team, the map became how we agreed on priorities and what to build next.

Design components & specification

My design partner and I built a design system and component library.

Prototype