Jon McNeill, CEO of DVx Ventures and former President of Tesla, opened HIMSS26 with a deceptively simple provocation: automation always comes last. In a conference hall full of people who flew to Las Vegas to talk about AI, his statement landed with force.
McNeill's talk drew from the operational framework he used to help grow Tesla's revenue from $2 billion to $20 billion in 18-30 months. He calls it "the algorithm":
- Question every requirement
- Delete every unnecessary step
- Simplify what remains
- Accelerate cycle time
- Automate
The implication for healthcare is pointed. Most organizations are trying to skip to step No. 5. They’re automating broken processes and calling it transformation.
McNeill said clear-cut goals are required. Cut diagnosis time in half. Reduce documentation time by 30%.
At Tesla, it was getting to 10 clicks to buy a car. Tesla got close at 13 clicks, largely by transforming a 12-page loan document into a single paragraph. "Question every requirement" was the driving philosophy: if it wasn't an actual law or regulation, get rid of it.
We thought about that a lot on the flight home, especially considering what our data shows every day.
What the data shows
Across one network we support — with thousands of providers, multiple EHR source systems, and millions of attributed patients — we track quality performance year-over-year across core preventive care and chronic disease measures. When we reviewed the most recent annual results, the directional story was strong.
Breast cancer screening improved 10-15 percentage points. Colorectal cancer screening and depression screening posted gains in a similar range. The share of diabetic patients with uncontrolled A1c dropped by close to 10 points. These aren’t rounding errors. They reflect real shifts in care delivery at scale.
But the same data also shows where the work is unfinished, and where a care coordinator or quality director can act right now.
Providers who more recently joined the network, a cohort representing 40-45% year-over-year growth, are performing 10-15 percentage points below established providers on depression screening. That isn’t a data artifact. It’s a specific, actionable gap: a defined cohort, a known measure, a performance differential with a dollar value attached. Targeted onboarding support and workflow guidance for PHQ-2 administration closes that gap. No new AI is required; it’s just the right question asked of the right data.
The blood pressure story is sharper still. The network average for controlling high blood pressure looks respectable in isolation. But the Medicare-age population sits 25-30 percentage points below that average. That gap affects tens of thousands of patients.
Those patients aren’t a statistical abstraction. They’re attributed to specific providers and visible in the data today. The intervention isn’t complicated: a follow-up visit, a medication adjustment, a documented reading. What was missing was the signal being surfaced clearly enough to act on.
Where cost and quality intersect
Healthcare organizations moving beyond isolated proofs of concept into a phase where AI, interoperability, and cybersecurity are being treated as one interconnected strategic agenda was the dominant message of HIMSS26. People are talking about a transition from stalled pilots to innovation at scale. But scale without specificity is just noise.
The cost story in value-based care is the same as the quality story: the opportunity is visible in the data long before it’s acted on. Excess SNF spend is traceable to discharge destination decisions made in the last 48 hours of a hospitalization. Carrier excess maps directly to referral patterns, specifically which specialists a provider habitually sends patients to and whether those specialists practice high-cost care. Inpatient excess concentrates in a handful of providers whose admission rates run above peer benchmarks for the same patient population and acuity mix.
None of this requires prediction. It requires connecting data that already exists — claims, quality scores, provider attribution, and utilization patterns. It requires asking the same question McNeill asked at Tesla about every step in the manufacturing line: why does this cost exist, and is it producing value?
The biggest challenge in healthcare innovation is adoption. "If they don't use it, you'll never find the value. You've got to match that solution with the workflow."
To drive adoption, match a cost insight to a workflow so it’s concrete enough for care coordinators to act on a Monday morning. Show them which patients are driving excess SNF spend, which discharges are going to the wrong facility tier, and which specialist referral patterns are generating outlier costs for clinically equivalent cases.
The actual lesson
Both McNeill and Dr. Halamka of Mayo Clinic Platform barely mentioned AI in their talks. Not because AI is unimportant. Because the organizations getting the most from AI are the ones simplifying the underlying process first. They asked the hard questions about what the data was saying before they reached for automation.
The tens of thousands of Medicare-age patients with uncontrolled blood pressure in that network don’t need new AI model to identify them. They need someone to look at the existing data, surface the gap clearly, and make it easy for a provider to act on it at the next visit.
That’s what good quality analytics do. McNeill's algorithm is a useful reminder of the order of operations: simplify first, then automate. The data has been telling us where to look. The job is to make sure the right people can see it.
Want to know how you can succeed in value-based care?
Ask our clients. Or better yet, ask us. We’ll show you what we’ve done, what we’re doing, and what we can do with you.
Healthmonix helps provider organizations and ACOs improve quality and cost performance across MIPS, MVP, ASM, and MSSP through the Healthmonix Prism™ platform. healthmonix.com
