HITL - Human in the Loop

When the automated data extraction or classification is not confident, or when the data is known to be wrong because business rules fail on it, a human operator needs to be involved. They make corrections or confirm uncertain data. Oftentimes, they inadvisedly train the underlying AI by doing this. A human is added into the loop of the end-to.end process. The idea is to minimize the human touch, hence they don’t play the main role anymore.