Concept Miner takes your project verbatim and clusters them into the most talked about topics. It uses machine learning to analyze the themes according to how the respondents talk about them. The topics are not linear, as they would be using Boolean, so you might find that the topic talks about staff friendliness and efficiency in the same topic. You can select the topics that are appropriate and make them into tags, which can be used in your categorization model alongside your normal Boolean queries. Tags understand a lot more than keywords, so you will be able to cover more of your verbatim with less work.
Note: You can have up to five cores in your Model Builder account, but currently you can only use one core per model.