When assisting a client with a workflow in Laserfiche Cloud, they needed to obtain the first (3) characters of their document types in order to use the first (3) letters in the document naming convention, so we tested the following regular expression: [a-z]{3} and the testing always returned the desired results. However, when running the actual workflow, the returned results shifted to the right by one character, where the 2nd, 3rd and 4th characters were displayed, instead of the first (3). (For example - the document type 'Resolution' returned Res in the test, but eso when the workflow ran.) To resolve the issue, we ended up using \S{3} for the regular expression instead. The 2nd regular expression returned the expected first (3) characters during the token testing, as well as when the actual workflow ran. I am wondering why the first regular expression tested correctly, but then provided different results when the actual workflow ran. Any thoughts?