Sohum Patel graduated first in his cohort in Psychological and Behavioral Sciences at Cambridge, completed his medical degree with Distinction, and later became one of the university’s youngest lecturers.

Mattijs studied Information Engineering at Cambridge after growing up across Belgium, Switzerland and Singapore.

Today, the two founders are building Penelope Health, a company helping healthcare teams understand the insurance rules that determine what will and won’t be covered and paid for.

The problem is not simply that healthcare payments are expensive or inefficient. It is that the knowledge required to process them is fragmented across payer portals, PDFs, emails and the heads of billing specialists.

For decades, that knowledge has been largely handcrafted: interpreted manually, passed from person to person and recreated inside disconnected systems. As AI begins to automate more healthcare workflows, another problem becomes clear. AI can only work reliably if it has structured, trustworthy information to work from.

Penelope is building that infrastructure.

The doctor who saw the system behind the patient

Sohum’s interest in improving the system began during medical training.

During residency, he saw how administrative chaos directly contributed to patient harm. A substantial amount of clinical time was consumed by documentation, navigating fragmented systems, chasing information and dealing with the processes surrounding payment. 

That’s time taken from patient care. At times, the administrative burden felt as demanding as the clinical work itself.

The drive to improve the system eventually led Sohum to a fact-finding mission to the US. 

Learning the language of healthcare administration

A friend told Sohum there was only one way to understand the industry: “You’ve got to get on a flight.”

So he did.

To save money, he flew into Canada, took a five-hour FlixBus and crossed the Rainbow Bridge on foot into the United States at two in the morning with an overpacked suitcase. He was heading to a healthcare billing conference near Niagara Falls.

The trip was a deliberate immersion. Sohum wanted to understand how providers, billing companies and insurers handled claims, denials and payer policies every day.

“I was learning in conversation,” he says. “Whenever I came across a term I didn’t know, I’d look it up between meetings and use it in the next conversation.”

Over the next two days, he spoke to anyone who would listen. Forty-eight hours later, he had a burgeoning network, early customer commitments and the foundations of what would become Penelope.

Back in London, his co-founder Mattijs was waiting to find out whether the company they had committed themselves to would actually have a reason to exist.

The trip did not prove that Penelope had a business. But it showed that the problem was real, widespread and painful enough for customers to want a better solution.

“At that point, we had a hypothesis, not a company,” Sohum says. “Startups have no fundamental right to exist. We had to earn the right to keep going.”

The six-year-old taking GCSEs

Long before startups, Sohum had developed an unusual relationship with ambition.

At six years old, he was one of the youngest people in Britain to pass a GCSE, the national exam normally taken by sixteen-year-olds.

His older sister was preparing for a computing exam two years early, covering the family whiteboard with programming exercises every weekend.

When she left the room, six-year-old Sohum would wander over and try to solve the problems himself. His parents eventually realized this was more than childhood curiosity and enrolled him in the same classes.

That instinct to push against expectations followed him to Cambridge. While studying medicine, Sohum wanted to spend his research year studying philosophy of neuroscience instead of a more traditional scientific discipline. The university initially refused, arguing the subject was too different.

Rather than accept the decision, he appealed until the university agreed to make an exception.

Before long, Sohum returned to teach students in the department that had once fought to join. He later became one of Cambridge’s youngest lecturers.

Thrown into the deep end

Mattijs’ childhood involved a different kind of education.

Born in Belgium, he moved countries every few years, living in Geneva twice and spending three years in Singapore before eventually studying Information Engineering at Cambridge.

Each move meant a new school, a new culture and often a new language.

When he first arrived in Singapore, he did not speak English.

“The first week, I was just sitting there having no clue what was happening,” he says.

He learned quickly because he had to.

Outside school, Mattijs trained as a competitive swimmer at a local club and competed at nationals. For several years, his coach only spoke Mandarin, a language Mattijs did not understand, relying on his friends to translate.

It was an early lesson in moving forward without having every piece of information.

“Even without the language, you can make yourself understood,” he says. “You just have to adapt quite quickly.”

That ability to enter unfamiliar environments, learn fast and keep going before things make sense would become useful years later, when he and Sohum began navigating healthcare administration.

Building the infrastructure for healthcare AI

US healthcare processes trillions of dollars in payments every year, yet much of the infrastructure behind those transactions remains fragmented, manual, and opaque. Claims pass between providers, insurers and billing companies through a web of disconnected systems and manual hands, creating complexity that costs the industry billions of dollars annually.

Penelope helps healthcare teams understand what each insurer will and won’t pay for. It brings together rules that are scattered across payer websites and policy documents, making them easier to find and apply. That helps teams check claims before submitting them, prepare requests for approval and appeals, and avoid denials caused by missing information or misunderstood coverage rules. As such, Penelope is creating the knowledge infrastructure required for real-time adjudication.

Sohum and Mattijs’s ambition is for Penelope to become the connective tissue between payers and providers. 

Building the infrastructure for healthcare is a highly ambitious goal. Then again, it is not the first time either of these founders have chosen the harder path - and succeeded.