How an AI platform is rebuilding patients record in Kenya

For years, conversations about Africa’s healthcare crisis have focused heavily on what is broken-underfunded hospitals, shortages of specialists, and delayed diagnoses that cost lives.

Less attention has been paid to practical solutions that work within existing constraints.

Yet a new generation of health innovators is shifting that narrative, building tools designed not for ideal systems, but for the realities patients and clinicians face every day.

One such effort is Tiba, an artificial intelligence-powered health platform founded by Flavian Simiyu, a graduate of Bachelors of Science Biomedical Engineering from the Technical University of Mombasa.

Simiyu’s entry into health technology was shaped by personal loss and a clear-eyed view of systemic failure.

After watching his uncle move between hospitals for months with persistent stomach pain, only to receive a terminal pancreatic cancer diagnosis when it was too late to treat, he became convinced that misdiagnosis and fragmented medical records were not just medical problems, but structural ones.

“My uncle did everything he was supposed to do. He went to hospital, took the medication, and followed instructions. What failed him was not effort or trust, but a system that could not see the full picture early enough,” Simiyu said.

.

Keep Reading

• Why KMPDC, patients must take medical misconduct seriously

Across much of African countries, that fragmentation remains the norm. Most health facilities still rely on paper records, while patients often seek care from multiple providers.

Simiyu noted that test results in hospitals are repeated, medical histories are incomplete, and clinicians make decisions without access to prior information.

“Even where electronic medical record systems exist, uptake has been slow,” he said.

Simiyu said that in Kenya, fewer than one in five healthcare workers actively use the national KeEMR system, reflecting broader challenges around cost, training and usability.

Across much of Africa, that fragmentation is structural. Most healthcare facilities still rely on paper records, while patients often seek care from multiple providers.

Each visit generates new handwritten notes, test results and prescriptions that remain locked in filing cabinets.

Clinicians rarely have access to prior records from other facilities, forcing them to make decisions based on incomplete histories.

Even where electronic medical record systems exist, adoption has been limited by high costs, long training periods and rigid workflows that do not reflect the pace of frontline care.

Tiba’s solution begins by addressing that fragmentation without attempting to erase existing practices.

“Rather than replacing paper, the platform is designed to absorb it. Clinicians continue writing notes, filling out charts or dictating observations in familiar languages,” Simiyu noted.

Using optical character recognition, speech-to-text and clinical language models, Tiba captures handwritten notes, scanned documents, lab reports and voice dictation, converting them into structured digital data in the background.

“This process happens without interrupting consultations or requiring clinicians to change how they work,” he explained.

Once captured, patient information is organised into a single longitudinal record that follows the patient across visits and facilities.

“Instead of scattered files, clinicians can see symptoms, investigations and treatments over time, regardless of where care was previously provided,” Simiyu said.

He added that the system is built to recognise inconsistencies and gaps, such as recurring symptoms with changing diagnoses, that often signal deeper underlying conditions.

Tiba then adds a clinical intelligence layer on top of this unified record.

“Clinicians can interact with patient data using natural language, asking questions such as which medications were previously prescribed, whether similar symptoms have appeared before, or which tests have already been conducted,” he said.

“The platform retrieves relevant information instantly, reducing repeated investigations and supporting faster, more informed decision-making,” Wafula added.

Beyond retrieval, the platform continuously analyses patient data in context.

By comparing symptoms, diagnoses, lab results and treatment responses, Tiba identifies patterns that may not be immediately visible in high-volume clinical settings.

The founder explains that when potential risks or inconsistencies are detected, the system generates evidence-based prompts to guide clinicians toward further investigation or alternative diagnoses.

“These prompts are advisory rather than prescriptive, preserving clinician autonomy while reducing the likelihood of oversight,” he said.

This model reflects growing evidence that AI, when deployed as decision support rather than automation, can improve quality of care.

In Kenya, similar AI-powered clinical support tools tested in outpatient settings have reduced errors in diagnosis, investigation ordering and treatment decisions across tens of thousands of patient visits.

For Simiyu, these results demonstrate how technology can function as a safety net in overstretched health systems.

“Clinicians are often working under extreme pressure, with limited time and information. AI can help connect the dots, not by replacing doctors, but by making sure nothing important is missed,” he said.

Patients are also integrated into the solution where Tiba includes a patient portal that allows individuals to upload paper records, access their digital medical history and share information across facilities.

“As patients move between providers, their records move with them, reducing repeated tests, lowering out-of-pocket costs and improving continuity of care. For patients, the platform shifts ownership of medical information away from institutions and back to individuals,” he noted.

According to Simiyu, affordability is central to the model.

He said that traditional electronic medical record systems can cost tens of thousands of dollars and require ongoing technical support.

Tiba, however, is designed to run on basic mobile devices with minimal infrastructure, lowering the barrier to adoption for small and mid-sized clinics.

“For facilities already using digital systems, the platform can integrate as an AI layer, adding decision-support capabilities without replacing existing software,” he said.

For Simiyu, the significance of Tiba lies not in its technical sophistication but in its practical impact.

“Misdiagnosis is often not about lack of knowledge. It is about missing information. When patient stories are fragmented, clinicians are forced to guess,” he said.

As African governments and donors continue to invest in digital health, Tiba offers a different lesson, that technology works best when it adapts to local realities rather than imposing new ones.

By quietly organising information, supporting clinical reasoning and preserving existing workflows, the platform aims to reduce preventable diagnostic errors and ensure fewer families hear the words “we found it too late."