Healthcare AI Platform | Jobs To Be Done Research | 2021
Strategic Direction Changed: Shifted product roadmap from consulting-heavy model to self-service platform, reallocating engineering resources toward data pipeline automation and model explainability features that users actually trusted.
Risk Mitigated: Identified critical trust barriers that would have prevented self-service adoption—preventing an estimated $400K in wasted development on advanced AI features users wouldn't adopt without data confidence and model transparency.
Opportunity Sized: Validated that actuaries and data scientists prioritize data trust (spending 70% of their time on data quality issues) over sophisticated AI capabilities, informing a phased $1.5M product investment strategy that addressed real adoption blockers.
The company's healthcare AI platform delivered real-time predictive analytics to hospital networks and insurance carriers. But there was a fundamental business model problem: the system required significant hands-on consulting support, making every customer interaction resource-intensive and limiting scalability.
Leadership faced a critical question:
How do we transition from a high-touch consulting model to a self-service product without losing customers or launching features they won't use?
The stakes were high. The wrong product decisions could mean investing millions in sophisticated AI capabilities that users would never adopt—or worse, stripping away support before users were ready, triggering churn.
I partnered with the VP of Product, engineering leadership, and the data science team to answer this question through research that would inform our entire strategic roadmap.
When I joined the stakeholder alignment sessions, the prevailing assumption was clear: "Our AI models need to be more sophisticated to reduce consulting dependency. If we build smarter predictions, users won't need our help."
The proposed roadmap reflected this belief—advanced machine learning features, more complex algorithms, cutting-edge predictive capabilities. The team believed the path to self-service was through AI innovation.
I conducted in-depth interviews and observational studies with actuaries and data scientists—the two primary user groups who would need to adopt a self-service model. I used the Jobs To Be Done framework to understand not just what they did, but why they did it and what motivated their decisions.