Jeff Elton.

In a glass building on Massachusetts Avenue in Cambridge sits a company most patients will never hear of, holding a record of how cancer actually behaves in the real world - not in the tidy fiction of a clinical trial, but in the EMRs, lab systems, and imaging files of roughly a hundred health systems. Jeff Elton built that. He calls it ConcertAI, and for eight years he ran it as founding CEO.

Most healthcare founders pitch a moonshot. Elton did something quieter and harder: he assembled data nobody else could assemble, then pointed generative and predictive AI at it. Oncology was the target because oncology is where the stakes, the complexity, and the data sprawl all collide at once. The result was a business that scaled from a 2017 startup to a roughly $1.9 billion valuation, with about 1,300 people and customers spanning biopharma, providers, regulators, and medical societies across the US, Europe, and Japan.

What makes Elton unusual isn't the title. It's the sequence. He wrote the book on healthcare disruption - literally, in 2016, with Anne O'Riordan - before he built a company to act on it. Most people write the memoir after. He published the thesis, then spent the next decade proving it with a balance sheet.

In May 2025 he handed the CEO chair to Eron Kelly and slid one seat over to Vice Chairman. Not a goodbye. A repositioning. He stayed on as the strategist and the institutional memory while the company entered its next phase. Founders rarely make that move gracefully. Elton treated it like a chess opening rather than an exit.

To understand why the dataset matters, start with the problem he spent a career circling. Clinical trials are controlled experiments. They tell you what happens to a carefully chosen group of patients under carefully managed conditions. They do not tell you what happens to the patient who shows up with three other conditions, the wrong insurance, and a tumor that does not read the textbook. That gap - between the trial and the clinic - is where real-world evidence lives. Elton's wager was that if you could capture the clinic at scale, in all its mess, you could close the gap. The catch is that the clinic does not hand over clean data. It hands over EMRs, lab information systems, and imaging archives that were never designed to talk to each other.

ConcertAI's answer was to treat those provider systems as the source of truth and to build the connective tissue - and later the AI - on top. The acquisitions tell that story plainly. TeraRecon brought imaging and the radiology pipeline. CancerLinQ, which began life inside the American Society of Clinical Oncology, brought a network of oncology practices and the longitudinal records that come with them. Bolted together, they became something more valuable than the sum: a multi-modal view of cancer that pairs the scan with the genomics with the treatment with the outcome.

Elton's career reads backwards from how most healthcare founders are built. He did not start in a lab or a garage. He started in the room where strategy gets argued - first as a partner at McKinsey, then as Global Chief Operating Officer and SVP Strategy at the Novartis Institutes for BioMedical Research, then as a Managing Director at Accenture leading North America life sciences strategy and global predictive health intelligence. Three vantage points on the same industry: the consultant who advises, the pharma operator who builds drugs, the technologist who predicts. ConcertAI is what happens when one person carries all three perspectives into a single company.

That background shows up in how he talks about the technology. He is not a hype merchant. When he describes generative AI, he reaches for the word "context" - the machine's talent for understanding what a question is actually asking and where it sits in a process. In oncology that is the whole game. A patient's record is not a row in a spreadsheet; it is a story with gaps, contradictions, and shorthand. Reading it the way a clinician reads it is precisely the task that defeated earlier generations of software. By 2024, at the ASCO Annual Meeting, ConcertAI was unveiling generative and predictive solutions, including its CARA AI line, aimed squarely at that comprehension problem - trial design, patient matching, decision support.

There is a tell in one of his lines about scale. The company can now operate "almost in real-time," he has said, and then he adds the honest part: this was "something we did not expect or anticipate in 2018." Founders love to claim they saw it all coming. Elton admits the pace of the field outran his own forecast. It is a small admission, but it is the kind that separates an operator from a salesman.

The move to Vice Chairman in 2025 fits the same temperament. The company framed it as a transition to accelerate AI innovation and growth, with a new CEO at the wheel and the founder kept close as strategist and thought leader. He had taken the company from nothing to a roughly $1.9 billion valuation, about 1,300 employees, and a customer list that runs through biopharma, providers, regulators, and medical societies. Handing that over while staying in the building is its own kind of discipline. So is chairing the board of MassBio, where he now helps steer the broader Massachusetts life sciences cluster that ConcertAI grew up inside.

Add it up and a pattern emerges. The 2016 book argued that "no action is not an option" - that incumbents who waited would lose to those who reorganized around data and value. The 2022 fundraise, $150 million from Sixth Street, was a market verdict that the thesis had legs. The 2024 product launches were the thesis turned into software a clinician could use. And the 2025 transition was the founder betting that the idea was now bigger than any one operator. The through-line is not a single breakthrough. It is patience - a decade of compounding a data asset that competitors cannot simply buy or copy, because much of it was assembled one health system, one acquisition, one integration at a time.