The ideas, frameworks, and instincts leaders are betting on right now
Komodo Summit opened not with slides or statistics but with a story of two real families navigating the crushing uncertainty of rare disease diagnosis. What followed was a morning of sessions that kept returning to one question: Are we using our data to see, or are we still operating blind?
The healthcare landscape is fundamentally shifting toward a future of real-time, trustworthy intelligence. Executives from Eli Lilly, GeneDx, and more shared a collective vision for an industry redefined by AI.
1. Moving Beyond the “Reporting Cadence”
Arif Nathoo, Co-Founder and CEO of Komodo Health, challenged leaders to stop relying on weekly or monthly reports to drive critical business decisions. Most Life Sciences organizations today spend weeks on custom analyses, receive a 50-page deck that’s already stale on arrival, and make decisions on data that’s months old.
“By the time you assemble all of that, the picture’s already changed. You’re doing archaeology. You’re excavating the past.”
—Arif Nathoo, Co-Founder & CEO, Komodo Health
The real AI revolution is organizational: eliminating the layers that summarize, delay, and dilute insights before they reach decision-makers. When built on the right data foundation, AI eliminates that chain, giving every leader direct, real-time awareness to drive market dominance. Nathoo left the audience with three direct challenges:
- Stop running your business on a reporting cadence. Weekly calls and quarterly decks were the best anyone could do when humans had to assemble intelligence. That’s no longer the case.
- Ask questions you’ve never been able to ask. Move from “What was our market share last month?” to “What’s changing in my market this week that I haven’t noticed yet?” These strategic questions can now be answered in real time.
- Kill something that’s “working fine.” Functional but blind workflows are the enemy of real awareness. If your dashboards work “well enough,” that’s exactly why they should be replaced.
The formula for AI adoption at scale is clear: 10% algorithm performance, 20% data and infrastructure, and 70% process change. Most organizations obsess over the first 30% and leave the rest untouched.
2. Architecting Trust: Plausible Is Not Trustworthy
Amit Sangani, Komodo’s Chief Technology Officer, emphasized that in healthcare, plausibility is not enough. Any large language model can produce a confident answer. That’s the dangerous — not the impressive — part.
“The failure mode isn’t a crash. It’s a well-formatted lie that nobody catches until it’s too late… In healthcare, if you get a wrong insight, you make wrong decisions. That has a direct impact on patients.”
— Amit Sangit, Head of Engineering, Komodo Health
Sangani’s “Five Pillars of Trust” are the essential foundation for reliable AI.
AI must be:
- Grounded. Every answer traces back to a verifiable source of truth, not a model’s opinion.
- Explainable. The model “shows the work” behind the result, making it easy to trace exactly how it reached its conclusion.
- Reproducible. If you run the same analysis today and again next week, you’ll get the same consistent answer.
- Verified. The model acts as an “immune system” against error, automatically checking every output for accuracy before delivering it to the user.
- Governed. Data is kept entirely private and secure. It is never used to train the AI and stays strictly within your control.
The competitive landscape is transforming rapidly. Model capabilities are converging fast. The raw models are commoditizing. The differentiator is the infrastructure supporting the model: the data that it sits on and the governance that makes its outputs defensible to regulators and boards. Every insight needs a “receipt’ that shows exactly how a number was derived.
“The winners in healthcare AI won’t be the ones who adopted AI first. They’ll be the ones who adopted AI they can trust.”
3. Scaling Beyond the Pilot: Cultural Buy-In
Diogo Rau, EVP and Chief Information & Digital Officer at Eli Lilly, shared how the organization moved from AI pilots to full-scale production by winning over the toughest stakeholders first.
He offered a field guide to navigating organizational resistance. These often masquerade as technical concerns and signal that a team is afraid to rethink processes. Objections come in three forms, and none of them are really about AI:
The three excuses:
- “We need to get our data foundation right first.” This most common excuse is rarely a real precondition and almost always a delay tactic.
- “We don’t have the resources.” This is the inverse of the actual solution. Free up capacity by automating; don’t wait to acquire more capacity first.
- “Regulations don’t allow it.” This is usually untrue and rarely provable when challenged. Ask for the specific citation.
Leaders cannot advocate for technology they’ve never touched. Build an app. Find a side project. Get your hands on the tools you’re asking others to adopt.
“You can’t go in and make a case for anything AI-related without being a practitioner… It will shape your perspective and give you a personal connection to whatever kind of case you’re trying to make.”
— Diogo Rau, EVP & Chief Information & Digital Officer, Eli Lilly & Co.
4. The Power of the “Flow State”
Paul Gurney, Head of Product at Komodo, presented the possibility of a world where AI handles repetitive, mundane tasks, allowing humans to focus on high-value cognitive work: asking incisive questions, interpreting results, and making decisions that improve patient outcomes. Three factors make that possible:
Principles of AI-Enabled Work:
- Reliability. AI must navigate the inherent messiness of healthcare data and still deliver trustworthy answers.
- Immediacy. Speed isn’t a convenience feature. It’s a prerequisite for peak human performance. Slow answers break focus.
- Flexibility. Open APIs. Users apply their own logic. No rigid, off-the-shelf experience.
The Flow State Argument for Speed
Achieving peak human productivity requires a “flow state” — operating at the top of one’s license on challenging problems without distraction. Make something faster and cheaper, and total utilization multiplies. AI users who once ran one analysis now run dozens — because the cost of curiosity has dropped.
“AI is going to do the bulk of the unpleasant work for us, and we get to do the fun stuff: the asking of incisive questions and turning those results into decisions… ultimately in service of making life better for patients.”
— Paul Gurney, Head of Product, Komodo Health
Every session underscored the same truth: The winners won’t be the ones who adopted AI first. They’ll be the ones who built it right. That means replacing the reporting cadences and blind workflows that slow intelligence down with systems that are fast enough to keep decision-makers in flow. It means governance that makes every output traceable and a defensible data foundation. Because at the end of the chain isn’t a dashboard — it’s a patient’s life.
Dive deeper into the full Summit session recordings here.
