CPO Rising Series: Cartwheel CPO on Building AI into Healthcare Without Breaking Trust

What happens to the entire product development process when creating software becomes as easy as creating a Google Doc? In this episode of the CPO Rising Series hosted by Products That Count Resident CPO Jay Patel, Cartwheel CPO Sarah Turrin speaks on how the collapse of software costs is reshaping healthcare product leadership, why the era of CPOs as pure strategists is already over, and how she thinks about building AI into healthcare where a probabilistic model simply cannot be unleashed on patients without deterministic guardrails. Thirty days into her role at one of the nation’s largest K-12 mental health platforms, Sarah brings a rare combination of clinical sensitivity, technical fluency, and genuine urgency about what product leaders need to do right now.

Join us for new conversations with leading product executives every week. Roll through the highlights of this week’s event below, then head on over to our Events page to see which product leaders will be joining us next week.

Show Notes:

1. The era of CPOs as pure strategists is over. Every product leader at every level should be shipping production code somewhere, whether at their day job or a side project. As the lines between engineering, product, and design blur, you cannot lead teams through that transition unless you are embracing it yourself.
2. Software is now as easy as creating a Google Doc, and that changes everything about how you build. The entire PRD-then-design-then-build process was constructed around the scarcity of expensive engineering resources. That scarcity is gone. The process built around it needs to go too.
3. A prototype-first approach replaces the scarcity mindset. When code is no longer the precious resource, the right move is to prototype immediately when you have an idea rather than spending weeks on documentation and design before anyone touches a build. The learning comes faster, and so does the product.
4. Everyone is now a manager, not a doer. The shift is not just for executives. Product managers and designers are increasingly operating agents and systems rather than executing tasks themselves. The question for every role is how do you set up a system to do this work rather than doing it yourself.
5. Slow down to speed up on infrastructure. The organizations that do not invest now in making their technical infrastructure legible to machines, with the right abstractions for agents to reason about and act on, will pay an existential price. Not making room for those investments is the real risk.
6. Write your SOPs so agents can read them too. Making operational procedures interpretable by AI agents as well as humans is not a future project. It is a now project. Systems that are only human-readable will not be able to take advantage of automation as quickly as those that were designed for both.
7. In healthcare AI, deterministic guardrails and probabilistic models must work together. You cannot unleash a probabilistic model on patients and hope for the best. The right architecture pairs a deterministic rules engine built on clear clinical guidelines with a large language model that reasons and communicates within those guardrails. That combination creates safety and empathy at the same time.
8. AI in healthcare unlocks care that was previously too expensive to deliver at scale. The breast cancer screening assistant, the postpartum companion app, the school counselor billing guide: all of these use cases share the same architecture. AI can now provide personalized, grounded, safe guidance in clinical contexts that previously required expensive human time at every step.
9. Time to effective support is the north star metric in healthcare product. How quickly from seeking care does someone actually receive care that helps them? It captures operational efficiency, scheduling capacity, and user experience in a single number, and it forces every department to work together to move it.
10. What looks like a trade-off between quality and scalability is often a false choice. When shifting from synchronous to asynchronous care, the assumption was that patients would lose something. What the team discovered was that many patients actually preferred asynchronous care. The middle way, designed with user optionality, served both the business and the patient.
11. Universal design, combined with AI, is making accessibility affordable for the first time. An accessibility audit that used to cost thousands of dollars and take weeks to execute can now be done in minutes and implemented in another ten. The argument that accessibility features are too expensive to prioritize is no longer valid.
12. Language is no longer a barrier to healthcare access. A voice agent configured to speak 80 percent of the world’s languages is now a real thing. For a company serving students through public school districts across the country, that is the difference between a child getting mental health care and not getting it.
13. Decision velocity is a CPO’s most important cultural contribution. It is not just about the CPO making faster decisions. It is about creating enough strategic alignment and clarity that everyone in the organization can make good decisions quickly without escalating everything upward.
14. A culture of learning requires psychological safety to fail. CPOs tend to break down cultures where people are afraid to try things that do not work out. Creating the conditions where a team can say “it didn’t work” openly, and has the skills to evaluate why, is one of the most valuable things a product leader brings to an organization.
15. Having long conversations with Claude about your codebase is a legitimate onboarding strategy. In the AI era, a new CPO does not have to rely entirely on scheduled time with engineers who may not always explain things clearly. Claude never gets tired and never gets annoyed, and the depth of technical understanding it enables is genuinely transformative for the first 30 days.
16. Go to Gemba: be as close to where value is created as possible. Especially in a new role, the most important thing a CPO can do is shadow care coordinators, sit with clinical team members, and meet with customers directly. That has not changed. The speed at which you need to do it has.
17. Not every trade-off is as binary as it appears. Product leaders who see only two options, scale or quality, fast or careful, accessible or commercially viable, are not yet seeing all the variables. The creative work of product leadership is finding implementations that honor multiple constraints simultaneously.
18. The limiting factor in software development has shifted from engineering capacity to judgment. Building is no longer the hard part. Deciding what to build, maintaining system coherence while moving quickly, and evaluating whether what was built actually worked are the skills that determine outcomes now.
19. AI is fundamentally changing what is commercially viable in healthcare. Services that were previously too expensive to deliver at the margins healthcare companies operate under are now achievable. Multilingual support, accessibility compliance, asynchronous care personalization: the economics of all of these have shifted.
20. Building during maternity leave because you could not stop thinking about the problem is what real product obsession looks like. Not as a prescription for anyone else, but as a signal: when you genuinely understand a user’s pain because you are living it, and you have the tools to do something about it, the product almost builds itself.