He survived a misdiagnosis. Then he built an AI platform for clinical decisions.

He survived a misdiagnosis. Then he built an AI platform for clinical decisions.

On March 20, 2017, Clement Okoh walked into a Lagos hospital with what previous doctors he consulted believed was a muscle strain and routine pain. Hours later, he said he could no longer walk.

He later learnt the muscle strain diagnosis was incorrect. What had been dismissed as routine pain was later diagnosed as aggressive multiple myeloma—blood cancer that develops in plasma cells—eating away at his spine. By the time the error became clear, the damage was severe. The tumour had weakened his vertebrae —the bones that form the human spinal column—so much that a minor fall was enough to fracture his spine and leave him paralysed.

Within hours, Okoh said he was flown to the United States. The doctors at John Hopkins Hospital, in Baltimore, Maryland, USA, he said, gave him four to five years to live, with a range of immediate risks: stroke, pulmonary embolism, deep vein thrombosis, blood poisoning, and internal bleeding. Surgeons removed the tumour and fused his spine. Okoh recalled his neurosurgeon once telling him that he would never walk again. But he did.

That recovery did more than save his life; it shaped his direction afterwards. During his time in intensive care and rehabilitation, he resolved that if he survived, he would return to Nigeria and work on building systems that could reduce the chances of similar outcomes in the future.

That promise became Monte Sereno Health, an artificial intelligence-powered platform designed to deliver proactive primary care and continuous health management, founded in 2021.

The company is attempting to address a deeper structural problem in Africa’s healthcare systems: fragmentation. Patients often move between informal providers, under-resourced clinics, pharmacies, and labs that rarely share data, while overstretched doctors make decisions with limited information.

A 2021 World Health Organisation (WHO) report on health information systems found that 30 of 47 African countries lacked the capacity to accurately register births and deaths, with cause-of-death data largely unavailable. The absence of common data standards further limits the ability to integrate and compare health information across systems.

Okoh’s misdiagnosis, he said, was not simply incompetence. It was the predictable outcome of a fragmented system, where doctors operate with limited data, patients carry paper records, and there is little real-time verification or support during clinical decisions.

In many cases, diagnosis depends on a single doctor’s judgment, often without access to full patient history or decision support tools. A study by Mayo Clinic, a non-profit academic medical centre, shows that up to 20% of serious conditions are misdiagnosed during initial visits globally. Telehealth, which has expanded access in recent years, does not fully solve the problem. It connects patients to doctors, but offers little oversight or quality control during consultations.

“You have no idea who you’re talking to, and there’s no real-time quality check,” Okoh said.

Monte Sereno’s answer, Okoh stressed, is not another telemedicine app. It is what he describes as a healthcare operating system: a full-stack digital infrastructure designed to sit above and connect every part of the care journey.

Instead of isolated consultations, the platform works by embedding artificial intelligence (AI) into every interaction. During a medical session, Monte’s AI agent, called StarPilot, sits alongside the doctor and patient, analysing symptoms in real time, pulling medical records, and querying global research databases.

If a patient reports a fever and headache, the system does not stop at common assumptions. It asks where the patient has travelled, cross-references disease prevalence, and suggests follow-up questions or tests. A visit to Lagos, for instance, would trigger prompts to rule out malaria or typhoid and not just flu.

The goal for Monte is not to replace doctors but to reduce the margin of error, according to Okoh. A 2025 cross-sectional study of Nigerian medical practitioners found that prevalence rates for medical errors range from 42.8% to as high as 89.8%.

“The AI can challenge both the doctor and the patient in real time,” Okoh said. “But the doctor still makes the final call.”

One of the platform’s central features is a portable electronic health record that follows the patient across providers and geographies.

Monte Sereno’s system digitises records, even from paper, using uploads. Once integrated, the data becomes part of a continuously updated profile that informs every interaction on the platform.

The system not only stores information; it interprets it. If a medication becomes unsafe due to new research, the platform flags it automatically. If a doctor prescribes a conflicting drug, it alerts both parties.

Monte Sereno is being built with Africa’s constraints in mind, according to Okoh. The continent faces a deepening health workforce crisis, with a projected shortfall of 6.1 million workers by 2030. At the same time, data from the Africa Centres for Disease Control and Prevention and UNECA shows a $66 billion annual gap in health financing.

In Nigeria, doctor-to-patient ratios can reach one to 10,000, according to the Nigeria Medical Association. In some rural areas, patients travel more than 30 kilometres from their homes to get medical attention where available.

Through built-in translation tools, Monte allows doctors in other countries like India, Egypt, and Latin America to consult with patients in Nigeria without language barriers. In pilot tests, multilingual consultations were conducted seamlessly, with each participant seeing responses in their preferred language.

It also supports shared consultations, where multiple patients can be assessed using a single device. Inspired by trials in India, this model helps extend care to communities with limited access to smartphones and reliable internet, where a single phone can serve thousands of patients.

But this approach raises privacy concerns. When multiple patients use the same device, sensitive data, such as medical histories, diagnoses, and personal details, can be exposed if safeguards are weak.

In low-connectivity settings, where devices are reused, and security is harder to enforce, the risks of data leaks or unauthorised access increase. Without strong encryption, user authentication, and clear data separation, patient confidentiality could be compromised, especially in communities where health-related stigma is high.

Okoh said the company has taken steps to address these risks.

“We opted out of letting LLM providers use any data from our platform to train their models,” he said. “We also have strict privacy agreements in place, and all data is encrypted and anonymised. Our serverless infrastructure on AWS is HIPAA-compliant.”