First, we need to understand about Markov model. Consider an example, you want to forecast the weather will be sunny, cloudy or raining. You go out of our home and observe weather for few days and calculate the transition probability, calculated as the number of transitions of that type we observed divided by number of days. In the Hidden Markov model, we do not have windows or doors to go out and check the weather. So, we calculate probability based on how many people are bringing umbrellas, wearing shorts and sunglasses.
In Artificial Intelligence, Hidden Markov models are used to build applications for speech recognition, gesture recognition, language recognition, motion sensing, protein folding, etc.