EV Conversion Revival

EV Conversion Revival

Resurrected a stalled EV build by embedding AI into a 3-person workshop team

AI-Enabled Team Leadership Complete Claude Projects Claude Cowork
EV Conversion Revival project

Published:

Context

Multi-year EV conversion project (ICE to electric) had stalled. 3-person workshop team with deep mechanical expertise but limited electrical integration knowledge. Months of blocked progress on a critical wiring integration issue — the kind of problem that sits at the intersection of multiple specialties where no single team member has the full picture.

Approach

Rather than bringing in external expertise or attempting to solve it myself, I embedded AI tools directly into the team’s existing workflow:

  • Loaded Claude Projects with vehicle specs, wiring diagrams, and full project history
  • Introduced Claude Cowork for real-time collaboration during workshop sessions
  • Positioned AI as a shared resource for the team, not a replacement for their expertise

The critical design choice: meet the team where they work, not where the tool works. No new apps to learn. No workflow to change. Just a new participant in their existing problem-solving conversations.

Outcome

The blocking electrical integration issue — unresolved for months — was solved in a single 30-minute collaborative session. Team members now independently use AI tools for problem-solving without my presence, applying the same approach to subsequent mechanical and electrical challenges.

Key Insight

AI adoption succeeds when embedded in existing workflows, not when requiring users to adapt to new processes. The PM's job is designing the introduction — not operating the tool.

Portfolio Signal

  • Team enablement for non-technical stakeholders
  • Change management in practical, physical-world settings
  • AI tool deployment beyond individual productivity — team-level impact
  • The Context-Station pattern applied to humans: load the environment, let the agents work

Corporate Translation

For hiring managers: this is the 'how do you get a team of 50 engineers to actually use the AI tools we bought' problem, solved at small scale first.