Multiple members of our team currently work at our school's Library and Information Services department. We found a way to make it easier for students and staff to contact the help desk with accurate and detailed information. We wanted to create a program that helped users walk through the difficult and boring steps of creating a ticket in a fun and quick manner. It also will create the ticket without them having to ever log in to Request Tracker.

What it does

The Library Digital Assistant (LDA) helps the user through the steps of creating a connectivity ticket. The user messages a Natural Language Processing (NLP) service in Microsoft's "Cognitive Services" suite called Language Understanding Intelligence Service (LUIS) by using Microsoft Bot Framework as a front end. LUIS will generate a keyword from the user's responses. With the keyword generated, we can predict the next steps of creating a ticket. For example, if the user inputs, "I am having a problem connecting to the Wi-Fi," LUIS will interpret this as a wireless problem, which will lead to the next logical question of, "What device seems to be giving you trouble?" After collecting all the information needed to create a ticket (problem, wired or wireless, what type of device, MAC Address, residence hall, and user's email), the LDA will create a ticket in the request tracker system.

How we built it

We started by deciding we would utilize the chat bot functionality of Microsoft Bot Framework with Azure hosting. We split up our task into training LUIS, setting up the Request Tracker (RT) API, and writing the JavaScript for the chat bot that will also integrate LUIS and RT. For the Request Tracker API, our group utilized the Azure cloud service to host an Ubuntu Linux virtual machine, which then serves our personal Request Tracker.

Challenges we ran into

  • The responses from LUIS cannot be stacked, which makes the flow of the conversation difficult to control.

  • The integration into Azure and LUIS was difficult as none of us had previous experience with them.

  • Creating a Request Tracker server

Accomplishments that we're proud of

We're proud that we were able to learn new technologies and languages in an enthusiastic and responsive manner. We had never done something like this before and we feel like it has very real applicability at our jobs and in healthcare. Another point that we are proud of is converting our full URL into concatenated environmental variables to make our service more secure.

What we learned

We learned new Microsoft technologies such as Azure and LUIS. We learned some basics of Node.js. We all became more experienced in Git.

What's next for Library Digital Assistant

Scalable goals include implementing a Bing API to help confused users look up what they need to know. This project could be modified for more healthcare applications such as teaching LUIS key medical terms that can be used to make check-in for a patient easier and faster.

Built With

Share this project: