Back in the early days of Maisha Meds, our malaria program ran off a simple spreadsheet. We were managing about 100 reimbursements each week to 20 community pharmacies in Kisumu. Our mission was simple: improve access to affordable care by supporting private pharmacies.
But as we began expanding, we realized we were going to need more than a spreadsheet to keep up. Processing claims manually could become a bottleneck not just for us, but for the local providers we served.
Too Many Claims, Too Little Time
Coming up with a new way to handle reimbursements wasn’t just about efficiency. It was about speed, too.
Imagine managing a local drug shop or pharmacy where every shilling counts. Every day spent waiting for a reimbursement means fewer medicines on your shelves, fewer patients served, and fewer opportunities to grow your business.
These reimbursements are essential to improve the quality of care, and they’re backed by research. In the private sector, malaria tests often go unused because they come at an additional cost to patients, and because pharmacies may lose money on negative tests compared to simply selling malaria medicines over the counter. Our approach realigns these incentives, making the most effective path also the most affordable one.
Doing this at the last mile, and ensuring resources reach patients who need them most, requires a more nuanced level of oversight. In our spreadsheet days, auditors would cross-check entries weekly, call a subset of patients to verify they’d received our services, and calculate reimbursement amounts themselves. This worked just fine when our reimbursements were in the hundreds per week—but how could we grow to a program that has now directly supported over a million patients?
There was only one way forward: We needed to build something that didn’t yet exist. We needed to innovate.
Reimagining Reimbursements from the Ground Up
This led us to build a claims management system called Madai, the Swahili word for “claims.”
Running quietly on the backend, Madai integrates with our core pharmacy software without disrupting providers’ day-to-day operations. It gives our team a dedicated portal for auditing, with greater visibility and control over the entire verification process.
The first iteration of Madai—version 1.0—focused on digitizing our manual processes. This made it easier to view claims, review patient data, and issue reimbursements. But as hundreds of claims grew into thousands, we needed more automation because we couldn’t hire more auditors every time the volume of claims increased.
We sat down with our auditors and studied how they worked. What patterns did they look for? What raised red flags? These insights birthed Madai 2.0, a system that thinks like an auditor and can start making informed decisions on their behalf.
Madai can now automatically flag irregularities, such as patient information being used across multiple visits, or care administered in less time than it takes to run a malaria test. The system also sends a unique code to patients’ phones, ensuring they’re present during visits.
And thanks to a colleague who came from the banking world, we introduced compliance scores modeled after financial risk assessments. These scores consider factors like historical rejection rates, regional malaria prevalence, and unique patient identifiers—allowing lower-risk pharmacies to receive faster approvals.
You Can’t Spell Madai Without AI
Another breakthrough came when we integrated artificial intelligence into Madai.
By partnering with Audere, we began automatically scanning malaria test images using computer vision. The AI confirms each test matches what the pharmacy reported, ensuring accuracy without slowing the process down.
Today, Madai is a robust platform that integrates real-time compliance checks, automated flags and approvals, photo verification, and AI-assisted review. Together, these systems quickly identify which claims still require manual review by a human auditor.
Over the course of the year we built the latest iteration of Madai, from January 2023 to January 2024, the average claim processing time dropped from six days to less than a day—even while the total number of claims was growing rapidly. The vast majority are now approved automatically, allowing us to handle over 80,000 claims per month without growing our audit team.
In the end, building Madai wasn’t just a tech challenge. It was about understanding the human side of the process, sitting with auditors, and learning from them. It was about tapping into both the local healthcare context and our own team’s diverse strengths.
And there’s no reason Madai should stop here or be limited only to Maisha Meds. We’ve built a platform that can one day help other organizations deliver and audit healthcare effectively across the private sector. When each claim represents a real person in need of care, the systems behind them need to be just as human-centered.
Elizabeth is a Senior Product Manager at Maisha Meds where she leads product development for our loyalty programs.
