Even a single critical prior art missed by a searcher may come to haunt the patent applicant/patentee later as it may hamper the patent prosecution or become a ground for invalidation or infringement. Patents are difficult to read, and many are intentionally written in a vague manner which is tough to understand and search. A drafter can also use novel terminology when drafting a patent application which is not easily known at the time of search. Semantic search works over contextual meaning of the input sentence or paragraph and so can help uncover missed results. You can easily NOT the results of the Semantic search with your Boolean query to see the delta results and scan them for any relevant record.

For prior art or invalidity searches with a short turnaround time, you can also directly start with a semantic search. In many cases you will find the matching set of prior art with just this approach. Both these types of searches might not need 100% precision and if you can make your case with a portion of the relevant prior art then further searches become redundant.