Ben Newton has spent a career making machines explain themselves. Before the phrase “AI infrastructure” acquired its present conference-hall sheen, he was thinking about logs, metrics, metadata, and the stubborn distance between a signal and a useful decision. His path ran through mathematics and physics, military and enterprise technology, Silicon Valley software companies, a podcast about data, three patents, and three IPOs. Then he took that education to Asheville and started again.
In 2021, Newton and his longtime friend John Clark cofounded Pluritem Health. Their target was not a glamorous consumer interface. It was the jumble underneath: records in incompatible formats, information scattered across systems, and useful context stranded where the person making a decision could not readily see it. They built a clinical data platform around aggregation, normalization, and analysis. Milliman acquired the company in August 2024. Newton stayed, became a Milliman principal, and now leads the practice behind CareFlowIQ.
The arc looks tidy in retrospect. Careers rarely feel tidy while being lived. Newton’s is better understood as a recurring fascination with translation. Raw material enters. Structure is added. A human receives something legible. The setting changed from application operations to clinical information, but the intellectual machinery remained recognizable.
01 / First principlesA physicist learns to love the plumbing
Newton studied mathematics and physics at Appalachian State University, graduating summa cum laude in 1997. He then went to Cornell, where his studies included physics and computer science and where he was a John and Fannie Hertz Foundation Fellow. The degrees matter less as decorative credentials than as a clue to his habits. Physics trains the eye toward systems, relationships, and the rules beneath visible events. Enterprise software supplies many visible events and even more unruly rules.
His early and middle career touched Northrop Grumman, EDS, Loudcloud and Opsware, BladeLogic, BMC, Sumo Logic, and Amazon Web Services. These were not identical businesses, but they lived close to a common anxiety: modern systems produce torrents of information while the people responsible for them need a comprehensible view. In 2013, writing about his move to Sumo Logic and Silicon Valley, Newton argued that enterprise startups were where the real action was. Consumer products had the attention; enterprise systems had the knotty work.
At Sumo Logic, Newton moved across product management, product marketing, and field evangelism. He explained observability before it became a standard item on every cloud diagram. In one presentation, he put the case plainly: “Everything you have produces data.” The next task was coverage and context, including the metadata that tells a team what the numbers belong to. A metric without identity is merely a number dressed for work.
“Everything you have produces data. It’s important to ensure you have all of the components covered.”Ben Newton, on modern application analytics
Three patents make the theme unusually literal. Milliman lists Newton as an inventor on “Key name synthesis,” “Data enrichment and augmentation,” and “Logs to Metrics Synthesis.” The titles will not trouble the bestseller list. They are, however, variations on the same useful obsession: take machine output, attach meaning, and turn it into a form that supports understanding.
The Newton translation loop
02 / The microphoneData needed a host, not another oracle
From 2016 through 2020, Newton hosted Sumo Logic’s Masters of Data podcast. The show ranged across data science, privacy, ethics, innovation, and technology. Its guest list gave him a front-row seat to a question that would later become central to his company: how should machine intelligence fit around human judgment?
A 2019 conversation with Stitch Fix’s Brad Klingenberg examined a business where algorithms touched inventory, marketing, operations, and the pairing of stylists with clients. The interesting part was cooperation. The machines could consider quantities of information beyond any individual’s reach, while people supplied judgment. The point was not to hold a coronation for the algorithm. It was to design a partnership that made the work more effective.
Newton’s public writing carries the same translator’s instinct. He used the frustrations of speaking with a two-year-old to explain why analysts and business teams need a shared language. He opened another essay with Lou Reed’s rule about never listening to your old stuff, then applied it to outdated analytics habits. A webinar framed logs and metrics as peanut butter and jelly: useful apart, more revealing together. The metaphors are slightly mischievous, which is why they work. Jargon tends to preserve the status of the speaker. An analogy transfers the idea to the listener.
03 / The founder turnA longtime friendship meets a long-neglected mess
Pluritem Health began in 2021 with Newton and Clark, two longtime friends whose backgrounds were notably complementary. Newton had spent decades around security, automation, product, sales, and analytics. Clark brought deep experience in data engineering and machine learning, including work connected to the technology that became Apple Health. They chose to work on fragmented clinical records: many files, many formats, many systems, and no guarantee that the context would arrive together.
Newton’s description of the problem was characteristically concrete. At the time of the Milliman acquisition, he said that 97 percent of medical data was locked in silos and legacy formats. The company’s proposed “key” combined aggregation with analytics, organizing information into a longitudinal view. The percentage became the headline statistic, but the founder’s deeper bet concerned sequence. Useful analytics come after reliable extraction and normalization. The order is unromantic and unforgiving.
“What the industry has needed for some time is a key to the lock, and Pluritem Health is that key.”Ben Newton, August 2024
There is an appealing consistency here. At Sumo Logic, Newton argued that complex applications became understandable through broad data coverage and meaningful metadata. At Pluritem, the source material changed, but context was again the scarce resource. His company was not asking a model to perform magic over a broken pile. It was trying to make the pile coherent first.
Completes mathematics and physics degree at Appalachian State University; begins Hertz Fellowship period.
Joins Sumo Logic after years across IT operations and enterprise software.
Begins hosting the Masters of Data podcast.
Cofounds Pluritem Health with John Clark in Asheville.
Milliman acquires Pluritem Health; Newton continues to lead the practice.
Milliman launches CareFlowIQ as its clinical intelligence platform.
04 / After the dealThe exit became a larger operating job
Acquisition stories often stop at the transaction because the neat ending is irresistible. Newton’s continued. Milliman bought Pluritem on August 1, 2024, positioning the platform alongside its healthcare intelligence work. Newton became CEO of the Pluritem practice and a Milliman principal. In October 2025, Milliman launched CareFlowIQ, presenting it as a platform that could convert disconnected records into searchable, context-rich information.
Newton said his team had developed the system using real records to handle real-world complexity. That phrase matters. Demonstrations reward the clean example. Operating environments supply the eccentric PDF, the old format, the missing field, and the note whose meaning depends on where it came from. A product becomes infrastructure when it survives the untidy cases without pretending they are tidy.
The post-acquisition work continued through new connections and customers. CareFlowIQ announced an expanded medication-history integration with Surescripts in November 2025. In January 2026, Heuro Health selected the platform for its clinical intelligence work. Each announcement moved the company further from a clever technical claim and toward an operating network. Integrations are where a data platform discovers whether its nouns agree with everyone else’s nouns.
05 / The stealable lessonBuild the context before selling the intelligence
Newton’s career offers a practical founder lesson: expertise can be a pattern rather than a sector. He moved through defense, IT operations, cloud software, observability, and healthcare technology. The durable skill was not allegiance to one category. It was the ability to recognize when data lacked the structure, context, or interface required for a person to use it.
His other lesson is about people. CareFlowIQ’s biography of Newton says he values customers and team members, prioritizing trust, collaboration, and outcomes in which both sides benefit. It sounds almost quaint beside the machinery. It is also the necessary conclusion. Data products do not become trusted by declaring themselves intelligent. Trust accumulates when a user can see where an answer came from, when the workflow respects judgment, and when the company behaves as though adoption is a relationship rather than a download.
The public record shows a man fond of systems and sufficiently suspicious of abstraction to reach for toddlers, sandwiches, and Lou Reed. He has spent years explaining that a machine’s output needs context. CareFlowIQ is the founder-scale expression of that argument. The fashionable layer may change every season. Newton keeps working on the layer beneath it, where names match, records connect, and the answer finally has somewhere trustworthy to stand.