From Data to Discovery: How AI Is Finding Hidden Cures — and Making Them Human

Drug development is one of the most expensive bets in science — and one with a 90% failure rate. But what if the answers to some of medicine's hardest problems are already out there, buried across decades of research, waiting to be connected?

That's the premise behind Rare Hopes, a self-funded nonprofit built on a remarkable collaboration between tellic LLC's 911-million-edge knowledge graph, AbbVie's applied biomedical research, and TXI Digital's human-centered product design. Together, these partners are using machine learning, Natural Language Processing, and biomedical semantic technology to mine over 53 million scientific documents — surfacing hidden connections that no human researcher could find alone.

In one striking case, the system linked 15 papers spanning 40 years to identify Ruxolitinib as a potential treatment for Carney Complex — a rare genetic disorder with no known therapy. The AI didn't just find a needle in a haystack. It found a needle no one knew to look for.

But the insight is only half the story. The other half is what happens when that discovery reaches a patient or a clinician. Raw data doesn't heal people — human experiences do. In this talk, we will explore how the team is translating the power of this knowledge graph into experiences that are intuitive, trustworthy, and genuinely useful for the doctors and patients who need it most. This session explores the full arc — from ontology-driven data architecture to conversational AI interfaces to patient-facing product design — and the hard-won lessons along the way: why domain-specific, curated data matters more than ever in an LLM world, why dashboards are giving way to knowledge systems, and why the most important question in AI-powered healthcare isn't "what can the model find?" but "can the person who needs it actually use what it found?"

The cure might already exist. The challenge is building the bridge.