AI products · Product strategy · Systems design
Designing systems that strengthen judgment
I work on AI-enabled products and workflows, using research, prototyping, and systems thinking to decide what the product should do, what AI can do reliably, and where human judgment should remain.
Designing with AI
Decide where judgment belongs, then draw the boundary deliberately
Every AI feature changes the division of work between people and the system. I use four questions to determine what AI can do independently, what requires human decision-making, and where AI should support human judgment.
- ReliabilityCan the system do this consistently well?
- RiskWhat happens if it gets it wrong?
- ReversibilityCan a mistake be detected and undone?
- Human judgmentDoes the outcome depend on context, empathy, ethics, or accountability?
- ReliabilityCan the system do this consistently well?
- RiskWhat happens if it gets it wrong?
- ReversibilityCan a mistake be detected and undone?
- Human judgmentDoes the outcome depend on context, empathy, ethics, or accountability?
Selected work
Where research changed the product
Each project began with a different assumption. Research changed what the product needed to do.
Sensory Sprout · Conversational AI
Preserve meal reliability before expanding choice
Research showed that families first needed to trust a small set of meals everyone could reliably eat. That shifted the product from generating more options to establishing meal reliability first, with AI expanding choice only after that foundation was in place.
Read the Sensory Sprout case study
Prototypes as discovery · Applied AI
Delay the intervention until the point of leverage is clear
Two working AI prototypes tested different assumptions about where intervention would actually help. The comparison revealed which behavior was worth designing for before committing to a product direction.
Read the AI Diagnostic case study
RefWorks · Enterprise SaaS
Earn trust through transparency and control
Research showed that users struggled to trust automated work they couldn’t inspect. That shifted the product toward visible system status, reversible actions, and clearer user control.
Read the RefWorks case studyMore work
StoryJam · Collaborative product
Protect independent thinking before the group begins to converge
Read case study
Research · Finance · Healthcare
Make complex workflows visible across people, systems, and pain points
Explore earlier systems workThe best AI systems increase capability without diminishing judgment.
Let’s talk
I’m interested in product roles where AI, complex systems, and consequential human decisions intersect.