Using Dialogflow to Quickly and Inexpensively Build NLP Voice Prototypes
Problem/Background:
The GSA helps government agencies with the acquisition of $50 Billion of goods and services from over 20,000 industry partners each year. They also have 87,000 pages of information on how to do business with the GSA. The complexity of going through this information can be a friction point for vendors, partners, and employees. This friction could reduce opportunities for millions of small business partners.
Introducing the GSA FAQ App
This app will transform GSA FAQs into an easily accessible virtual agent that allow users to access this information. This agent will be accessed through Google Home speakers and Smart Displays.
Technology Approach:
This app will be built using Dialogflow. Dialogflow is a Google NLP product that allows users to create Intent and Response communication flows. The information that the GSA is in charge of is massive. There are 7 areas that the GSA provides goods and services to customers. It was important that all the areas were accessible while not making the app difficult for customers to use. To ensure that the intents launch properly I would create Parent/Child Virtual Agent Structure. I simulated the Parent/Child Virtual Agent Structure by using the follow-on intent feature.
Ensuring that Citizens and Government Employees can easily access public information by using AI solutions is a quick-win AI use case. By using no-code AI building applications, Government Innovators can make a real difference for American Citizens.