InsideTracker
Conversational AI Health Coach
From static insights to an ongoing health companion
Platforms
iOS, Android, responsive web
Timeline
2024-2025
Teams
Product | Engineering | Science | Marketing
My role
In a 0→1, three-phase initiative, I led end-to-end design of a more engaging, personalized, and guided health experience- reframing static insights into an ongoing companion. Backed by research and ongoing validation, I identified a conversational AI model as the right solution to help users better understand their data and take action, leading to the introduction of a conversational AI health coach. I partnered cross-functionally to define the vision, interaction model, and UX, driving prototyping and alignment through launch: resulting in increased engagement and more sustained user action.
Impact
+26%
Monthly return rate
+24%
Weekly return rate
4.4/5
Satisfaction rate
Context
InsideTracker is a personalized nutrition platform that provides health insights based on blood, physiological, and genetic data.
The core of the InsideTracker experience is a blood results analysis and an Action plan with a list of tailored recommendations for user on how to improve their health.
The problem
The problem: ONLY 34% of users returned within a month after viewing their initial results. Which means that the majority of users didn’t really engage with the core experience of the product.
Blood results page
Action plan page
Grounding the problem with user data
I approached research from both qualitative and quantitative angles. I conducted user interviews and small focus groups to understand how people experienced their results, and a clear pattern was that they felt overwhelmed and unsure how to prioritize next steps. I paired that with analyzing usage data and CX tickets, where we saw repeated requests for more guidance.
79%
Felt Overwhelm &
Confused
65%
Want dynamic & personal experience
82%
Want a guided health journey plan
Want a guided
health journey
Exploring solutions collaboratively
Remote workshop (FigJam) with cross-functional team
Conversational AI that evolves over time.
Solution: a conversational AI that evolves over time.
Solution:
01
Phase 1: foundational AI
02
Phase 2: adding personalization
03
Phase 3: always-on coach (still work in progress)
Phase 1: building the foundation
Design solution
Phase 1 focused on introducing the InsideTracker's AI chat experience for the first time. The goal was to keep it simple, trustworthy, and easy to use- reduce overwhelm and validate interest.
Results of phase 1
45% returned: Within a month → Users stick around longer
29% returned: Users return rate within 7 days increased
3.9/5 satisfaction : measured via an in-product 1 question survey
User feedback validated the value of the new AI experience and enabled the team to move forward with deeper personalization in Phase 2 and beyond.
Phase 2: adding personalization
Design solution
Phase 2 focused on evolving the AI from a simple conversational assistant into a personalized guidance engine. The goal was to deliver answers grounded in the user’s actual biomarker data, habits, and recent activity.
We also renamed the AI tool and it is now: Terra.
Results of phase 2
58% returned: Within a month → Users stick around longer
41% returned: Users return rate within 7 days increased
4.4/5 satisfaction : measured via an in-product 1 question survey
Phase 3: always-on coach (in-progress)
AI evolved from a chat tool into a proactive, floating companion accessible from every screen.
Terra opens as an overlay so users can minimize it, continue exploring their results, and return exactly where they left off.
Note: This is part of a larger AI-Native Product Transformation initiative I'm currently leading
Design solution
Phase 1 focused on introducing the InsideTracker's AI chat experience for the first time. The goal was to keep it simple, trustworthy, and easy to use- reduce overwhelm and validate interest.
Results of phase 1
45% returned: Within a month → Users stick around longer
29% returned: Users return rate within 7 days increased
3.9/5 satisfaction : measured via an in-product 1 question survey
User feedback validated the value of the new AI experience and enabled the team to move forward with deeper personalization in Phase 2 and beyond.
Phase 1: building the foundation
Design solution
Phase 2 focused on evolving the AI from a simple conversational assistant into a personalized guidance engine. The goal was to deliver answers grounded in the user’s actual biomarker data, habits, and recent activity.
We also renamed the AI tool and it is now: Terra.
Results of phase 2
58% returned: Within a month → Users stick around longer
41% returned: Users return rate within 7 days increased
4.4/5 satisfaction : measured via an in-product 1 question survey
Phase 2: adding personalization
Project outcome
Consistent growth in both early engagement and long-term return
+26%
Increase in Return (within 1 month)
+24%
Increase in Return (within 7 days)
4.4/5
User satisfaction score
+26%
Increase in Return (within 1 month)
+24%
Increase in Return (within 7 days)
4.4/5
User satisfaction score
Conversational AI Health Coach
From Static Insights to an Ongoing Health Companion
InsideTracker
Impact
+26%
Monthly return
+24%
Weekly return
4.4/5
Satisfaction rate
My role
In a 0→1, three-phase initiative, I led end-to-end design of a more engaging, personalized, and guided health experience- reframing static insights into an ongoing companion. Backed by research and ongoing validation, I identified a conversational AI model as the right solution to help users better understand their data and take action, leading to the introduction of a conversational AI health coach. I partnered cross-functionally to define the vision, interaction model, and UX, driving prototyping and alignment through launch: resulting in increased engagement and more sustained user action.
Platforms
iOS, Android, responsive web
Timeline
2024-2025
Teams
Product | Engineering | Science | Marketing
About InsideTracker
InsideTracker is a personalized nutrition platform that provides health insights based on blood, physiological, and genetic data.
The core of the InsideTracker experience is a blood results analysis and an Action plan with a list of tailored recommendations for user on how to improve their health.
Current experience
However: engagement analysis showed that 68% of users returned less than once per month after viewing their initial results, often feeling overwhelmed and unsure what to do next.
Research
I approached research from both qualitative and quantitative angles. I conducted user interviews and small focus groups to understand how people experienced their results, and a clear pattern was that they felt overwhelmed and unsure how to prioritize next steps. I paired that with analyzing usage data and CX tickets, where we saw repeated requests for more guidance.
Solution: conversational AI that evolves over time.
01
Phase 1: Foundational AI
02
Phase 2: Adding personalization
03
Phase 3: Always-On Coach (still work in progress)
Exploring solutions
Remote workshop (FigJam) with cross-functional team
Phase 1: the foundation
Phase 1 focused on introducing the InsideTracker's AI chat experience for the first time. The goal was to keep it simple, trustworthy, and easy to use- reduce overwhelm and validate interest.
Phase 3: always-on coach (in- progress)
AI evolved from a chat tool into Terra - a proactive, floating companion accessible from every screen. Terra opens as an overlay so users can minimize it, continue exploring their results, and return exactly where they left off.
Note: This is part of a larger AI-Native Product Transformation initiative I'm currently leading.
Phase 2: personalization
Phase 2 focused on evolving the AI from a simple conversational assistant into a personalized guidance engine. The goal was to deliver answers grounded in the user’s actual biomarker data, habits, and recent activity.
We also renamed the AI tool and it is now: Terra.
+26%
Increase in Return
(within 1 month)
+24%
Increase in Return (within 7 days)
4.4/5
User satisfaction score
Outcomes
79%
Of users felt
Overwhelm & Confused.
65%
Of users wanted more dynamic & personal experience.
82%
Of users wanted more guided health journey plan.
Crafted with creativity, late-night iterations, Ctrl+Z reflexes, and a carefully curated playlist.
© Chen Segal 2025
© Chen Segal 2025
Crafted with creativity, late-night iterations, Ctrl+Z reflexes, and a carefully curated playlist.
© Chen Segal 2025