What do you like best?
The best thing about spaCy is that it is open-source still gives industry-ready state-of-art algorithms. Also their documentation is sound enough to let a learner independently understand the whole library with any experts advice. Even they provide tons of examples for a newbie to practice.
What do you dislike?
I have been using this library form a long time and I am truly satisfied by it. Have never faced any crucial issues, yes but I had some minor errors which I was easily able to solve using the public forums.
Recommendations to others considering the product
spaCy has a good detailed documentation which will help you all the way but still you may need some easy extra help so I suggest you to go through this simple tutorial by analyticsvidhva.com : https://www.analyticsvidhya.com/blog/2017/04/natural-language-processing-made-easy-using-spacy-%E2%80%8Bin-python/ , this will really boos your learning rate.
What business problems are you solving with the product? What benefits have you realized?
We at TechDynasty use spaCy to carry out Natural Language Processing and create algorithms over it to meet the needs of our customers and serve them the solutions.