Jordan Campbell · Director of Product Design & AI Operations at Inbox Health

I design healthcare payment and AI products, and I ship them in code.

Sole product designer across a seven-app platform serving 2.5M+ patients. I start with research, then build in TypeScript and React with Claude Code.

  • +23% Checkout conversion
  • 17 Practices on a 0-to-1 card reader
  • 50 Automation flows saving about 30 hours a month
  • 138 Opportunities from on-site research

Selected outcomes

  • Checkout: Rebuilt on Stripe Payment Element from six or seven to three steps. Payment conversion up 23%, above a 20% target; AutoPay opt-in roughly doubled.
  • Card reader, 0 to 1: Relaunched on alpha fixes and grew to 17 practices. The card-on-file pilot cut post-visit statement costs 77%.
  • AI support: Designed, 0 to 1, the customization for AI support chat and AI phone calls, with human review before a response is sent.
  • Partner API docs: Turned a bare list of endpoints into documentation with field explanations, use cases, test calls, and a copy-for-LLM button. It opened partnership and integration conversations.
  • Design system: Built in my first two months, then consolidated several scattered systems and the engineering library into one.
  • Research in Guatemala: Shadowed 9 teammates across 14 sessions and logged 138 improvement opportunities, which led to forming the product operations team.
  • Carrier LYNX: Excursion evaluation from one to five days to 15 minutes, per Carrier’s own materials.

Selected work

  • Inbox Health: Checkout conversion up 23%, above a 20% target.
  • Inbox Health Card Reader: Relaunched after alpha fixes and grew to 17 practices.
  • Inbox Health AI Support: Designed, 0 to 1, the customization for AI support chat and AI phone calls, with human review before a response is sent.
  • Inbox Health Partner API Docs: Turned a bare list of endpoints into documentation partners can build on, with a copy-for-LLM button.
  • Carrier LYNX: Excursion evaluation: days → 15 minutes, per Carrier’s own materials.
  • The Home Depot: AR room scanning app that generates floor plans from a phone camera.
  • McDonald's SMART Request: Menu change request time: days → 17 minutes in early tests.
  • Sight: Dashboards for a machine-vision product that read the cameras a site already had.
  • Data Transaction Token: A token explorer designed for a wider audience than developers.
  • Hibernate Sleep: Complete brand refresh: identity, app, packaging, and pitch deck.

Machine-readable portfolio

Use the public-only LLM brief, Markdown portfolio, or knowledge JSON. Expanded and private records are excluded.