The challenge: skilled scientists stuck in admin
The lab specialises in ELISA testing, a standard diagnostic method used worldwide. Its scientists are highly trained, but their days were going on manual work rather than science: hours of data entry from handwritten collection sheets, manual validation of results, and tracking sample locations by hand. The expertise was there; the time to use it was not.
What we built
Rather than a full system overhaul, we ran a focused pilot on the three worst bottlenecks, built and deployed over a few days.
Smart data entry
Optical character recognition paired with AI validation now reads sample data automatically and corrects common errors as it goes. Data entry dropped from five minutes a sample to thirty seconds - four samples processed in under two minutes where one used to take five.
Automated test processing
A second tool validates results against control samples, generates the analysis curves, and calculates sample-pair averages. Processing a test plate fell from an hour to five minutes.
Sample location tracking
Handwritten location data is converted into precise coordinates and mapped to each sample, so positive results can be tracked for epidemiological analysis in real time.
Why it worked
The pilot stayed small and aimed at the work the scientists least wanted to do. None of their judgement was automated; the routine around it was. With the admin handled, the team got its capacity back - around twelve times the processing throughput, without adding staff - and the scientists returned to the analysis only they can do.
What's next
The lab is now extending the same approach to other diagnostic tests, deeper integration with its lab management software, and earlier predictive checks on results.