Clerk Assist Chatbot

Senior capstone project sponsored by Novobi 

CPS 491s20 - Team 3

Clerk Assist


Team Members:

Picture of Ben Bratton

Ben Bratton

CIS 2020 - Seeking a UI/UX development job after May.

Email: benjbratton@gmail.com 

Picture of Luke Hengehold

Luke Hengehold

CIS 2020 - Seeking a software development job after May.

Email: hengeholdl1@udayton.edu

Picture of Noah Roswell

Noah Roswell

CIS 2020 - Seeking a software development job after May.

Email: roswelln1@udayton.edu 

Picture of Ryan Hetman

Ryan Hetman

CIS 2020 - Seeking a software development job after May.

Email: hetmanr1@udayton.edu


Meet our project owners:
Phu Phung
Phu Phung
Phu served as our advisor and project structural lead at the Univeristy of Dayton.

Hai Dinh
Hai Dinh
Hai served as our product owner at Novobi in Vietnam. He provided us with the requirements for the project and we had weekly meetings to check in on our progress.

About the Project:

The project was to create a chatbot that would help in the operation of a retail store.
Hai provided us with some resources such as DialogFlow to start on, and we were tasked with meeting the following requirements:


1) The chatbot itself

2) Webhook

3) Database

4) Deployment on accessible platforms


With these guidelines we proceeded to make our bot with the primary function of querying inventory of a store's databse to, as the name suggests, assist the sales associates of the store.

Schedule:

Our team used Trello to stay organized and on schedule using scrum policies.

See our simplified schedule below:

Sprint 1: 1/23 - 2/12
· Design and implement conversational flow
· Research Dialogflow API
· Research best ways to test and deploy

Sprint 2: 2/13 - 3/4
· Connect database (Firebase)
· Work on intents and entities in Dialogflow
· Implement bot with messenger apps and google assistant

Sprint 3: 3/5 - 4/24
· Finish intents
· Write fulfillment functions
· Make the bot more user friendly


Design

We used all google products to streamline the development process.
The user interacts with one of our three connected integrations which speaks to the server running the code on DialogFlow. The code then queries the database housed in Firebase to retrieve as well as input data.

Clerk Assist Architecture Diagram

Integrations 

To provide flexibility we designed the bot to work with both the Google Assistant, Hangouts Chat, and as an embedable iFrame element that can be added to any webpage.      

Embedable UI
Try the Demo



Google Assistant
You can sign up to be a beta tester of our bot on your google account and simply ask the google assistant: "Talk to Clerk Assist"
Sign up Here
*You will need to use a non-UD email, so the Google Assistant isn't blocked.

HangoutsChat
Unfortunately the bot is only accessible in a closed alpha for HangoutsChat, but check out our demo video to see how it works.
 


Copyright © 2020 · Clerk Assist Team