Digitizing Hospital Records with AI: What I Built for LaserEye Centre
A real-world case study: how I built an AI agent for a Nairobi hospital that automatically processes uploaded medical record images, extracts patient data, and stores it in a structured database.
Insights
The Challenge: Paper Records in a Digital World
LaserEye Centre, a specialist eye clinic in Westlands, Nairobi, was managing a large volume of patient records in physical or scanned image form. Retrieving records was slow, error-prone, and dependent on manual data entry. The ask was simple: can we make this smarter? The answer was an AI agent that does it automatically.
How the Agent Works
Staff upload a scanned image of a patient record through a simple interface. The agent receives the image via n8n, passes it to an AI vision model that performs OCR and data extraction, identifies key fields — patient name, date of birth, diagnosis, prescription, attending doctor — and structures them into a clean JSON object. That object is then written into the clinic's database automatically, with no manual typing required.
The Technical Stack
The entire system runs on n8n. Image upload triggers a webhook, which passes the image to an AI vision node for extraction. Extracted fields are validated, formatted, and sent to the database via API. Error handling is built in — if the agent can't confidently extract a field, it flags the record for human review rather than guessing. Accuracy and reliability were non-negotiable in a medical context.
Impact and What This Means for Healthcare in Kenya
The agent reduced data entry time for each record from several minutes to near zero. More importantly, it standardised the data structure, making records searchable and retrievable in seconds. This kind of AI deployment isn't just a productivity win — it's a step toward the kind of digital health infrastructure that Kenya's healthcare sector needs. And it's built with tools that are already accessible to any developer willing to learn them.
More to Discover
Digitizing Hospital Records with AI: What I Built for LaserEye Centre
A real-world case study: how I built an AI agent for a Nairobi hospital that automatically processes uploaded medical record images, extracts patient data, and stores it in a structured database.
Insights
The Challenge: Paper Records in a Digital World
LaserEye Centre, a specialist eye clinic in Westlands, Nairobi, was managing a large volume of patient records in physical or scanned image form. Retrieving records was slow, error-prone, and dependent on manual data entry. The ask was simple: can we make this smarter? The answer was an AI agent that does it automatically.
How the Agent Works
Staff upload a scanned image of a patient record through a simple interface. The agent receives the image via n8n, passes it to an AI vision model that performs OCR and data extraction, identifies key fields — patient name, date of birth, diagnosis, prescription, attending doctor — and structures them into a clean JSON object. That object is then written into the clinic's database automatically, with no manual typing required.
The Technical Stack
The entire system runs on n8n. Image upload triggers a webhook, which passes the image to an AI vision node for extraction. Extracted fields are validated, formatted, and sent to the database via API. Error handling is built in — if the agent can't confidently extract a field, it flags the record for human review rather than guessing. Accuracy and reliability were non-negotiable in a medical context.
Impact and What This Means for Healthcare in Kenya
The agent reduced data entry time for each record from several minutes to near zero. More importantly, it standardised the data structure, making records searchable and retrievable in seconds. This kind of AI deployment isn't just a productivity win — it's a step toward the kind of digital health infrastructure that Kenya's healthcare sector needs. And it's built with tools that are already accessible to any developer willing to learn them.
More to Discover
Digitizing Hospital Records with AI: What I Built for LaserEye Centre
A real-world case study: how I built an AI agent for a Nairobi hospital that automatically processes uploaded medical record images, extracts patient data, and stores it in a structured database.
Insights
The Challenge: Paper Records in a Digital World
LaserEye Centre, a specialist eye clinic in Westlands, Nairobi, was managing a large volume of patient records in physical or scanned image form. Retrieving records was slow, error-prone, and dependent on manual data entry. The ask was simple: can we make this smarter? The answer was an AI agent that does it automatically.
How the Agent Works
Staff upload a scanned image of a patient record through a simple interface. The agent receives the image via n8n, passes it to an AI vision model that performs OCR and data extraction, identifies key fields — patient name, date of birth, diagnosis, prescription, attending doctor — and structures them into a clean JSON object. That object is then written into the clinic's database automatically, with no manual typing required.
The Technical Stack
The entire system runs on n8n. Image upload triggers a webhook, which passes the image to an AI vision node for extraction. Extracted fields are validated, formatted, and sent to the database via API. Error handling is built in — if the agent can't confidently extract a field, it flags the record for human review rather than guessing. Accuracy and reliability were non-negotiable in a medical context.
Impact and What This Means for Healthcare in Kenya
The agent reduced data entry time for each record from several minutes to near zero. More importantly, it standardised the data structure, making records searchable and retrievable in seconds. This kind of AI deployment isn't just a productivity win — it's a step toward the kind of digital health infrastructure that Kenya's healthcare sector needs. And it's built with tools that are already accessible to any developer willing to learn them.