Another option is to use extract instead of replace, it's hard to know if this approach is going to generalize well with the rest of your data since you are just providing one text string as sample.
library(stringr)
x="Book FCP1 is missing in BOOKS setup. Book may be new."
str_extract(x, "(?<=Book\\s)[A-Z]+\\d")
#> [1] "FCP1"