Johannesburg clinical data company BioInformatiCo has placed second in the 2026 CDISC AI Innovation Challenge with Leova, an AI platform that builds and tests the databases used to collect data in clinical trials.

It was runner-up in its category to a joint entry from drug maker Novartis and Amazon Web Services. "Placing second in the CDISC AI Innovation Challenge is an important endorsement of what we have built," said founder and chief executive Shirley Collie.

The company plans to open a seed funding round in the next two to three months.

Every clinical trial needs a database before the first participant can join, built to match the study's protocol: what information is collected, who qualifies, and how results will be judged. Specialists build the forms, set the rules that catch bad data and test the whole system, and errors are costly to fix later.

"Before the first participant can enrol, study teams may spend 10 to 20 weeks building and testing the database," Collie said, adding that the cost falls hard on grant-funded African researchers working on vaccines for diseases such as HIV, TB and Ebola. Leova brings the steps into one workflow. It reads the study protocol, builds the database and then tests it using synthetic participants, made-up records that expose gaps before real patient data goes in.

Collie says that in one study, testing and validation fell from five days of specialist work to one. The challenge was run by CDISC, the non-profit body that sets data standards used in drug research and regulatory submissions, and drew 50 submissions from 30 organisations. Leova placed second of 19 entries in the category for using AI-generated synthetic data to test software, and was presented at CDISC's US conference in Denver this month.

BioInformatiCo does not sell Leova as software for customers to run themselves. It delivers finished, verified databases with its own data management staff, and bills by milestone rather than by the hour. It has not said how much it plans to raise.

CDISC's own results page lists Leova as runner-up, describing it as turning a protocol into a standard digital format and generating repeatable synthetic-data scenarios to test study builds. ITWeb reported the result. TechFinancials carried details of the planned seed round and a comment from Professor Glenda Gray, director of the Infectious Disease and Oncology Research Institute at Wits and former head of the South African Medical Research Council: "African researchers need to be able to own, control and evaluate their own data."

The detail worth noting for founders is where Leova came from. BioInformatiCo did not set out to build an AI product and then look for buyers. It built databases for clinical trials, found the slow, repetitive parts of its own work, and automated them. That means it already knows its customers, and the challenge result shows the product holds up against a global drug company and the world's largest cloud provider.

Its business model is also a sign of where a lot of AI is heading: a service company using AI to deliver faster and at a predictable price, rather than handing customers software to run themselves. That suits clinical research, where accuracy matters, and budgets are tight.

It fits a wider push to build medical evidence in Africa rather than borrow it from elsewhere, the same gap behind UCT's Google-funded medicine research. The seed round will be the next test of whether investors see a scalable business, not just a well-run specialist firm.

You might also like our piece on the new physiotherapy regulations, a Grade 9 learner's lung cancer research, and Endeavor Catalyst's new fund.

Get more SA tech and business news and subscribe to The Open Letter.