Intelligent Pooled Testing
For Efficient COVID-19 Screening

Redefining The Rules For Testing

Request Info

Our Approach

wePool AI provides a computational testing strategy that leverages Artificial Intelligence to predict a subject’s probability of testing positive for COVID-19, and uses it to segment test populations into distinct pools.

Informed by clinical data, our segmentation intelligence allows laboratories to reliably clear multiple subjects with a single test by minimizing the probability of encountering a positive result within an otherwise negative pool.

Based on wePool AI’s Prevalence Assessment, our software can recommend group screening strategies to maximize test kit savings, and drastically increase sample analysis capacity.

Test Kits Savings

wePool AI enables test kit savings even in high-prevalence environments. Our simulations show we can save between 40% and 60% of test kits utilized.

Increase In Capacity

Test more subjects with less total kits used. We expect to deliver a 150% to 300% increase in sample analysis capacity.

Reasonable Pool Sizes

Appropriate pool sizing per your target prevalence. Using pool sizes between 3 and 5 samples, we enable the same savings you would expect from larger and riskier pool sizes.

Privacy At Our Core

Our algorithms work with and are trained to expect de-identified data to ensure data privacy and HIPAA compliance.

Maintaining Sensitivity & Specificity

Because our model utilizes small pool sizes, we ensure minor to no loss in testing specificity and sensitivity.

Model Flexibility

We understand different facilities have different testing procedures. Our model is designed to work with and integrate into existing data collection & sample extraction protocols.

What is Pool Testing and How Does it Work?

Pool testing, or pooled sampling, is when samples from multiple subjects are combined and tested with a single test kit. A negative result confirms that all samples are negative; a positive result would require individual testing, or breaking pools further into smaller subsets.

Although the methodology has been used for Malaria & HIV, countries like Germany and Israel have been employing it for COVID-19. On July 2020, the FDA issued an Emergency Use Authorization to allow sample pooling of up to four individuals. However, high disease prevalence⁠—a high rate of positives⁠—prevents effective use of the methodology.

Read More

wePool AI in Action

Powered by Artificial Intelligence, our solution seeks to mitigate high prevalence concerns. Our model has been tested with synthetic precisionFDA datasets containing sociodemographic and clinical testing data.

Our results confirm that pool testing alone is not viable at higher disease prevalence, wherein test kit utilization savings diminish rapidly. By contrast, the wePool AI solution delivers consistent decreases in tests used, and substantial increases in sample analysis capacity—even in high prevalence environments. Our simulations were performed with pool sizes constrained to 4 samples, and our results are expected to improve as more pool testing options become authorized.

If wePool AI were applied to Quest Diagnostics’ and Labcorp’s combined test per day throughput, our solution will have enabled the clearing of an estimated additional 80 million subjects by the end of 2020.

About Us

wePool is an interdisciplinary group that came together during the MIT COVID-19 CHALLENGE in April 2020.

The 48-hour hackathon selected 1,500 participants out of over 4,500 applications and challenged us to tackle several issues related to COVID-19.

We partnered with mentors and experts in the field. Our team focused on testing efficiency in the face of test shortages around the globe.

wePool successfully emerged as Hackathon Winners during Beat The Pandemic Round I, and again during Beat The Pandemic Round II in Early June, 2020.

Learn More

wePool in the Press

"One such team, wePool, sought to address the challenge of enabling mass testing in a cost-effective and rapid manner…

wePool sought to take the method a step further by implementing machine learning techniques to intelligently segment subjects and reduce the chances of getting a positive hit within a pool of otherwise negative tests.”

Read Full Article

Meet The Founders

Guillermo Siman

MIT MBA 2022
Economics, Business & Technology

Yash Patil

Florida Institute of Technology
Operations Research & Data Analytics

Smrithi Sunil

Boston University PhD Candidate
Biomedical Engineering

Contact Us

We are looking for Partner Organizations and Collaborators.

Company Based In


Email Us