“Am I okay, and what should I do next?”
People worried about memory and thinking wanted to understand their results and what to do next. Raw scores and medical language could leave them with more questions.
The product needed to make cognitive assessment easier to access and understand, with clear boundaries between educational guidance and clinical judgment.
Boundaries for the product
- Keep diagnosis and treatment decisions with clinicians
- Make the basis for each result understandable
- Explain results without alarming or overwhelming the reader
- Work for older adults with limited technical confidence
- Support a viable direct-to-consumer service
Assessment first, AI second
For the original at-home kit, I grounded results in cognitive tasks and used AI to explain and summarize them. Task scores determined the result; AI helped make it understandable.
That approach took more work than asking a model for an assessment, but gave us a clearer basis for reviewing the explanation. The model was not making a diagnosis.
Insight framing over raw scores
In V1, I organized results by cognitive domain—memory, attention and executive function—with plain-language summaries and relative comparisons. I left out traffic-light risk labels and terms such as “decline” or “impairment.” The aim was to explain what the tasks showed without turning a score into a verdict.
Intentional cut lines for the MVP
I kept these out of the original MVP because we did not yet have the evidence or clinical workflows to support them:
- Longitudinal prediction
- Risk labeling
- Automated care plans
- Clinical escalation
From kit to clinician-reviewed service
The next version replaced the physical kit with licensed BrainCheck Assess and added a 1:1 results consultation with a neurologist. That changed the product from an assessment report into a service with a clinical conversation about the results and possible next steps.
I designed the $299 cash-pay offer and built the acquisition, onboarding, payment and clinician-review workflows using Claude Code and Codex.
Designing for the person arranging care, too
The person paying for an assessment may be arranging it for a parent or another family member. I separated the purchaser from the participant in onboarding, captured their relationship, and carried that context into the invitation and clinician review.
The invitation explains who arranged the assessment, while each person gets the next steps relevant to them. This made the handoff between a family member, the participant and the clinician an explicit part of the product.
Making an interrupted checkout recoverable
I worked through the less visible parts of the experience: validating phone numbers, restoring an unfinished checkout, making consent explicit, and recognizing payments already completed or in progress. A returning customer should reach the right next step without being asked to pay again.
Measuring clarity and completion
Early product evaluation focused on whether people could complete the assessment, understand the results and know what to do next. These measures guide usability improvements; they do not establish clinical accuracy or long-term health outcomes.
- Assessment completion rate
- Drop-off points within tasks
- Time-to-insight
- Qualitative feedback on confusion vs. reassurance
- User questions after results, as a signal of trust gaps