In this episode of Concept to Care, we sat down with Ben Hackett, Founder and CEO of Basalt Health, to talk about what it really takes to bring large language models into clinical workflows — not as demos or copilots, but as production software inside real health systems.
Ben has spent his career building healthcare products at scale, from research tooling to patient advocacy platforms, and that experience shows up clearly in how he thinks about AI: cautiously optimistic, deeply empathetic to clinicians, and grounded in the realities of governance, workflow, and trust.
This conversation is less about hype — and more about what happens when AI meets the messy, human, high-stakes reality of care delivery.
In our conversation, Ben shares insights on:
How patient satisfaction data at Accolade led directly to the founding insight behind Basalt
Why patients returning for care is more about context than speed or access
How Basalt balances structured clinical protocols with LLM-powered flexibility
What they learned from early product designs that “looked perfect” but failed in testing
Why healthcare AI must be auditable, repeatable, and human-validated
How operators — not just clinicians — quietly shape patient outcomes
The realities of selling into health systems and earning trust over time
