The Solar University is an in-house Learning Management System (LMS) kicked off by our PM and a senior developer while I was still heads-down finishing the Volt 2.0 redesign. The business case was obvious: we were paying $20,000 a month for Docebo, a clunky third-party training platform that was prone to technical issues out of our control. It was also one of the first touch points reps had with our system—the final onboarding step required in order to fully access Volt and start selling. Building an in-house LMS would have been time prohibitive just a handful of months prior, but with AI we could do it ourselves in a fraction of the time.

This also meant the PM and developer could move fast with AI handling the early design scaffolding. I knew if I didn't get my design input in early, the product would get too far down a road without it. Online learning UX is something I'm genuinely interested in so I made sure to carve out some time. Over the course of one evening and into the next day, I conducted a compressed design process: a full UX audit with AI-assisted note-taking, an AI-research and brainstorm session that both validated my instincts and surfaced new ideas, and a working prototype delivered in roughly half a day.

First I went through our full onboarding flow the way a rep would—hitting the required online training and going through the full set of lessons. As I went, I opened a Claude chat in Slack as a running notepad, recording my observations via voice transcription as they came to me.

I noted what I thought the platform got right—the lessons were bite-size less than 10 minutes each with some interactivity, layout and format variety, with questions or quizzes to confirm comprehension. And I noted the UX problems—the bite-sized lessons encouraged me to keep going but they didn't automatically lead to the next required lesson. Instead I lost momentum figuring out where I was and what was next. And overall the mobile experience was not clean and responsive.

Ideas surfaced mid-audit: what if instead of a standard quiz, certain lessons asked reps to record themselves explaining a concept in their own words? For technical content—net metering, true-ups, utility rate structures—the kind of knowledge you need in the field, being able to articulate it out loud is a better test of understanding than picking an answer from a list.

With the audit feedback and research in hand, I moved into Figma and built a working interactive prototype in roughly half a day. The goal wasn't pixel perfection or realistic content— it was to show the flow, the structure, and the decisions I'd made— something the PM and developer could actually click through rather than interpret from static mockups.

The prototype is built on the Volt 2.0 design system, which meant I wasn't starting from zero on visual language. I focused my time on the architecture and the UX decisions that needed to be visible to be understood.

One of the clearest problems in the Docebo audit was the break in momentum between lessons. In the prototype, finishing a lesson immediately serves up the next one. If a rep is working through a track, the experience is continuous until the track is done.

What the prototype shows:

• Seamless lesson sequencing — completing a lesson flows directly into the next, no dead ends or unnecessary navigation
• Progress indicators at both the individual lesson and overall track level — always visible so reps know exactly where they are and how close they are to finishing
• Learning tracks with a required vs. optional distinction — clear from the start which content gates full app access and which is available to explore further
• Prerequisite unlocking — tracks open in sequence as required content is completed, giving structure without over-constraining
• Multiple assessment types — multiple choice, flip card review, and accordion-style content, with the groundwork for the recording format as a future layer
• Track completion animation — a small celebration moment when a rep finishes a full track, consistent with the gamification thinking running through the rest of the app

Solar University was another product just at the finish line that never fully shipped before the bankruptcy came. But this case study is worth documenting — because it demonstrates how AI changes the pace and texture of design work.

The audit, research, synthesis, and prototype would have taken days through a conventional process. Using AI at each stage—not to replace the design thinking, but to accelerate the parts that slow it down—compressed that into a single day's work.

The research-to-prototype path is the part I'd carry forward into any future work on learning products. Jumping straight into a working prototype with AI rather than going through rounds of static screens first meant the PM and developer had something to react to immediately.

The combination of genuine design expertise and AI-assisted speed is something I want to keep exploring — this was an early proof of what that can look like.