Lately Bridget, Thuy, and I have been working on web support questions from the years 2018-2019 as well as current incoming questions. These questions involve editing sites on Cascade and using the COLA web editor. Our job is to filter through these emails that have been sent and determine what is being asked. We then look up if it is information that has already been documented on our web-support site or is not documented but needs to be. We are doing this because recently, in the last few months, we have added a chat bot to our web editing site to help making searching for answers easier for staff, admins, faculty, and grad students.
Below is an image of the chat bot that we have, a screenshot of the type of document that has already been added to the bot to link people too, and lastly what the training for the bot looks like. We include the title of the webpage with the instructions, and the hyperlink, and included various types of phrases that could possibly be asked about the topics answered in that instruction guide. As we go through more emails, we can identify a larger quantity of phrasing of questions that can be added to links we already have in the chat bot to make it more intelligent and accurate.
Underneath is part of the excel spreadsheet Bridget has made for keeping track of the type of documentation that needs to be made and added to our instructional site. These are what we have found so far during our first round of filtering through emails. Once, we start making these after the first round look throughs are complete, we can make the documentation needed and train the chat bot all at once. Then we will go through the second round of emails we previously filtered and update their status that the chat bot has been trained.