This paper aims at analyzing how technologies, like AI, can enable us to take better measures against COVID-19 which has caused multi-disciplinary changes at the global level. Governments are seen trying to curb the spread of COVID-19. But one of the major problems they have been facing is the shortage of testing equipment. Considering this a strategy for finding alternative solutions is a must to ensure minimizing the number of tests that needed to be done. One such approach is: pool sampling, i.e. combined patient samples and testing the combine samples once. Pooling can succeed at a unitary cost, if all the samples taken are negative. But if a single sample comes out to be positive then infected patient does not mean failure. This paper describes how to optimally detect infected patients in pool samples, i.e. using a minimum number of tests to exactly recognize them, by making an assumption the a priori probabilities that every patient is healthy. Estimation of those probabilities using questionnaires, supervised machine learning or clinical examinations can be done. The algorithmic results achieved, are like informed divide-and-conquer methodologies and are efficient at performance.