Core needs
Both sides of the platform
What patients and clinicians need, measured and ranked.
Case study · Verily Lightpath · 2021 to 2025 Draft · highlighted text is placeholder
Helping care teams act 35% faster in chronic care
Lightpath is a two-sided platform for chronic disease management. I aligned senior leaders on its three-year product vision, then helped define and evaluate an LLM capability that summarizes patient records and recommends next-best clinical actions.
- Role
- Staff UX researcher
- Partners
- PM, UX design, clinical, product analytics
- Users
- Physicians, nurse practitioners, health coaches, patients
- Methods
- Interviews, survey, vision workshops, iterative evaluation
Situation
Care teams had more patients than hours. Every visit started with reading the chart.
Context
- The platform served two sides: people managing chronic conditions, and the clinicians and coaches who care for them.
- Leaders across product, commercial and clinical had different views of where the platform should go next.
- Physicians, nurse practitioners and health coaches spent much of their prep time reading records to decide what to do next.
Constraints
- A product vision had to hold for three years while the market for AI in care moved quickly.
- Any AI recommendation had to be safe, traceable and easy for a clinician to overrule.
- Clinicians could give research only short sessions between patients.
- Faster provider time to action
- 35%
- Years in the product vision
- 3
- Clinicians and coaches interviewed
- 40
Task
Agree where the platform goes. Then make care teams faster.
Establish core patient and clinician needs and the openings to stand apart, align senior stakeholders on a three-year vision, and evaluate the first AI-enabled clinical workflow.
A shared vision
Three years, one direction
Product director, head of commercial and clinical leadership behind one plan.
A faster workflow
From chart to next action
An LLM summary that shortens visit prep and points to the next step.
Action
Two phases. Set the vision, then prove one workflow.
Phase 1 · Needs and vision
What should the platform become?
Interviews and a survey with patients and clinicians, then vision workshops with the product director, head of commercial and clinical leadership.
Phase 2 · Iterative evaluation
Does the summary help clinicians act?
Physicians, nurse practitioners and health coaches used the record summary on real case types, round by round, with product analytics tracking time to action.
Findings
Clinicians need the next action. Not more data.
- Prep time crowds out care
Clinicians described spending a third of each visit's time reading notes to rebuild the patient's story.
- Trust needs a source
A recommendation was used only when the clinician could see which notes and results it came from.
- One summary doesn't fit every role
Health coaches wanted goals and habits; physicians wanted changes in medication, labs and risk.
“If it tells me what changed since last time and where it read that, I'll use it every day.”
Recommendations
What we built toward.
| Recommendation | Why | What it looks like |
|---|---|---|
| Summary by role | Coaches and physicians look for different things. | Separate views for coaching goals and clinical changes. |
| Next-best action | Prep time crowds out care. | One suggested next step per patient, with the reason in a line. |
| Source on demand | Trust needs a source. | Every statement links to the note or result behind it. |
Result
A vision leaders signed, and a faster workflow. Launched as Verily Lightpath.
Strategy
- Senior stakeholders aligned on a three-year product vision and strategy, including AI-enabled clinical workflows.
- The platform launched as Verily Lightpath.
Product
- The LLM record summary cut average provider time to action by 35% within three months of launch.
- The role-based summary pattern was reused across later care team tools.