01
Prove demand before building
Kano model surveys and fast-turnaround prototypes answered the hard question first: would caregivers actually embrace a digital tool? Only then did the real build start.
Vertical
Healthcare AI that earns trust.
Caregivers make high-stakes decisions with imperfect information. I spent 18 months at Microsoft Health Futures proving an AI platform could meet them there — three product waves, 50K+ users reached, 92% patient satisfaction.
The problem
One in five adults is a family caregiver — coordinating medications, appointments, and care plans across a fragmented system. They need the right information at the right moment, explained at a level they can act on. AI that can't earn that trust doesn't get used.
The pattern that worked.
From the Connected Care playbook — the moves that shipped, not theory.
02
Map AI to real pain points
A socio-ecological framework mapped AI capabilities to actual caregiver pain — from individual information needs out to community-level support — so the technology served the workflow, not the other way around.
03
Build the trust system
AI product work has a failure rate. Usability testing showed caregivers needed an unmistakable signal for AI-generated vs human content — so the product got source visibility, privacy controls, and clearly stated purpose for every section.
04
Anchor to a real engagement model
The strategy anchored to the KLAS patient engagement model — security, health, personal capabilities, and a supportive environment — giving providers tools to extend quality care into the home.
Fair questions.
- Do I need to know what to ask for?
- No. Bring the business problem. The first call is a conversation about what's stuck — if there's a fit, we define the smallest useful engagement together. Diagnosis is the job; you don't need the answer before you book.
- What kind of healthcare work fits the queue?
- An issue is a shipped thing: a patient-facing flow a stranger can finish, an AI-assisted workflow your team can run, an internal tool that removes a real bottleneck. If it doesn't ship, it's not an issue.
- How do you handle the trust question with AI in healthcare?
- The same way Connected Care did: failure modes as design inputs, visible sources, clear AI-vs-human attribution, and review paths — before launch, not as a risk review at the end.
Draft — Chris to supply
Compliance posture for healthcare work — how Chris handles PHI boundaries, BAAs, and regulated data in an engagement. Chris to supply; nothing here should imply a stance he hasn't stated.
Last updated September 23, 2026
Ready to start?
Bring the problem.
Ship the fix.
Tell me what is stuck. We will agree on the scope, budget, and delivery target before work starts.