Ariel Katz’s first useful database began with an empty search result. He had spent a summer working in a Columbia University research lab, then transferred to Binghamton University as a psychology student. When he tried to find another lab, the opportunities seemed to vanish into departmental pages, private inboxes, and the general fog that collects around information nobody has bothered to organize. Katz was qualified. The work existed. The route between the two was broken.
His response was not especially academic. He started building. ResearchConnection gathered professor profiles and research opportunities into a searchable system, turning a junior-year frustration into a company before Katz had quite adjusted to the idea that he was running one. He and three co-founders mapped faculty across American universities, raised early capital, and discovered that a database is only useful when people know it exists.
Their solution arrived glazed.
The team bought 10,000 Krispy Kreme doughnuts and offered them to Binghamton students who signed up. Roughly half the campus joined in under 24 hours. The stunt also caused trouble because the team had not secured permission to table. It was equal parts acquisition campaign and undergraduate farce, but the result contained a durable lesson: information products need distribution as badly as they need information.
Katz had already taken an indirect route to that table. He attended the University of Pittsburgh, spent time at the Hebrew University of Jerusalem, and transferred to Binghamton, three universities in three years. At Binghamton he joined two laboratories, completed an honors thesis, and graduated in 2015. His training was in psychology rather than computer science. The habits still transferred neatly: frame a question, gather evidence, test an explanation, revise when reality objects.
He also worked as a research assistant in a Columbia lab studying consciousness, visual perception, and metacognition. Academic credits from that period include work on social-cue recognition and cognitive control. ResearchConnection was therefore more than a directory conceived by an outsider. It grew from someone who had experienced both sides of the doorway, first as a student searching for a lab and then as a researcher inside one.
The second map was hiding inside the first
ResearchConnection grew to hundreds of thousands of users and a team of about 20 before it was sold in 2016. The most revealing customers, however, were not students. Companies used the platform to find people and generate leads. Katz initially found that behavior irritating. Later, he recognized the business tucked inside it. Structured knowledge about expertise could help institutions make expensive decisions.
After the sale, Katz took time off, traveled, and met Ian Sax in India through a salesperson who had worked with both men. They stayed in touch. When Sax later called on a Saturday and offered to acquire a small project Katz was building, Katz said no. The refusal became a friendship, then a working relationship. Near the end of 2017, Katz presented Sax with a deck for H1. Sax committed within the first few slides.
One problem, increasing scale
H1 began with a larger, more consequential version of the old problem. Doctor information was everywhere and useful context was nowhere in particular. Publications sat in one system. Clinical-trial experience lived in another. Affiliations, specialties, network participation, referrals, and patient-population data changed on different schedules. A spreadsheet could hold names. It could not easily answer which investigator fit a particular trial, which expert understood a particular field, or which provider belonged in a particular network.
“Culture is what a person does when nobody else is looking.”Ariel Katz
A company built around the next question
Katz describes H1 as a source of truth for doctor information. The phrase is tidy; the work is not. H1’s platform integrates fragmented data and applies AI to the repeated act of matching a doctor to a need. Pharmaceutical teams use it in drug development and medical affairs. Clinical teams evaluate investigators and sites. Health plans work on provider networks and directory accuracy. Digital-health services use provider information to help people navigate care.
This is less a directory than a machine for asking better questions. Who has worked on this kind of trial? Where do they practice? Which patient populations do they serve? What have they published? The right doctor changes with the question, so the data must carry context rather than merely contact details.
The business model followed the architecture. Doctors could maintain their own profiles, while organizations licensed the platform and its data. In the early years, life-sciences companies formed the center of gravity. The platform helped medical-affairs teams identify specialists and helped clinical teams compare possible investigators. As H1 added health plans and digital-health customers, provider directories became another expression of the same matching problem. A misspelled address looks mundane beside a clinical-trial decision, but both failures begin with unreliable identity and context.
Katz’s language around AI is similarly practical. He focuses on adoption inside workflows, particularly the administrative steps surrounding trial planning, data analysis, and provider information. The promise is not a theatrical robot doctor. It is fewer hours spent reconciling records and more reliable inputs when a person must choose. H1’s Doctor Graph, as the company now calls its central model, is designed to connect identity, expertise, and professional relationships rather than leave each fact in its original silo.
H1 entered Y Combinator’s Winter 2020 batch after it already had a Series A term sheet. Katz has said the accelerator widened the company’s network. During the same period, H1 moved at a pace that sounds mildly indecent even by venture standards: seed, Series A, and Series B in a single year. Katz recalled that the Series B process took six days once the data room opened.
Fast money can make a company look inevitable in retrospect. Katz’s own account is less polished. He talks about emailing people, failing a thousand times, getting one response, and then watching most scheduled meetings produce nothing. Persistence, in his telling, is not a heroic mood. It is arithmetic.
The awkward business of growing up in public
In 2021, Forbes placed Katz on its 30 Under 30 Healthcare list. He has called the recognition embarrassing, which is a more interesting reaction than another framed certificate. His public conversations tend to return to the less decorative parts of leadership: hiring people older and more experienced than himself, firing friends when roles no longer fit, and keeping standards legible as a team spreads across continents.
His definition of culture is useful because it is behavioral. What does a person do when nobody is watching? Do they choose the easy patch or the durable one? Do they help a user without waiting to be prompted? A small founding team can share urgency by osmosis. Hundreds of employees require language, systems, and managers who can translate the original instinct without turning it into corporate wallpaper.
Katz has also argued that some periods of life can be intentionally imbalanced. He spent early stretches working through nights, then learned to delegate and protect time for family. It is not a universal prescription. It is his attempt to make sense of the founder’s bargain after living both sides of it.
“The best team is a team that can celebrate wins together and laugh at losses together.”Ariel Katz
Expansion that rhymes
By 2025, H1 was widening its reach through acquisitions. Ribbon Health added provider data used by digital-health companies and health plans. Veda added automation for the same stubborn provider-data problem. Former FDA Commissioner Stephen Hahn joined the board. Each move pushed H1 beyond its original life-sciences foothold while staying close to the company’s recurring question: how can an organization identify and work with the right doctor?
In May 2026, H1 announced a $40 million investment round led by CVS Health Ventures. The announcement followed joint projects focused on improving provider-directory accuracy. H1 said it served 85 percent of the top 20 pharmaceutical companies and nine of the ten largest health plans, and described the business as profitable. The figures matter because they show the map becoming infrastructure, mostly out of sight of the people whose searches depend on it.
Katz’s stated ambition remains plain: connect the world to the right doctor. He thinks AI can remove administrative work and improve how clinical trials and provider networks are assembled. He also has a more distant idea called nature in space, built around the day humans need to grow fruits and vegetables away from Earth. Backers have not arrived for that one yet. Katz appears content to wait for the calendar to catch up.
The space project sounds like a charming detour until you notice the familiar pattern. Find a vital resource. Observe that access is poorly organized. Build the missing route. A research lab, a doctor, a tomato beyond the atmosphere: Katz keeps returning to the distance between what exists and who can reach it.
The practical lesson in his career is smaller than the ambition and more portable. Frustration can be evidence. The blank search result may be a market map in disguise. Start by asking what information people repeatedly need, where it is trapped, and what decision becomes possible once the pieces meet. Doughnuts remain optional, although history suggests they help.