Sentiment analysis for specific program/services we offer

Hi! I am new to this community (as well as ML), so thanks in advance for your help and patience!

My organization wants to implement sentiment analysis to know how are clients feel about our programs and services (positive/neutral/negative). So far, we can determine our clients’ overall sentiment about working with us. But, we would like to drill even deeper into their sentiment around specific aspects of our programs and services.

For example, we might receive feedback like this: ‘Thanks for the great degree program you made available! I really like working with you. The application process was easy but it was hard to contact the school I want to attend.’

In that feedback, we would probably find that the overall sentiment was positive, as was the application process but more negative when it came to contacting their school of choice.

Any ideas on how to start building an ML algorithm to create this kind of sentiment analysis?