loading
Menu
Let's Talk

Business Impact

  • 1.5s

    Average Response Time

  • 134K

    Queries Answered per day

  • 24/7

    Live Chat Availability

  • 94%

    Accuracy

Customer Key Facts

  • Size : 1,000+ employees
  • Location : Chicago, IL
  • Industry : Government

Problem Context

During the COVID-19 pandemic, the Illinois Department of Employment Security (IDES) started experiencing an unprecedented surge in inbound unemployment claims from those seeking help for available Federal and State pandemic assistance, such as FPUC (Federal Pandemic Unemployment Compensation). To effectively respond to the large number of incoming queries surrounding unemployment claims, the Illinois Department of Employment Security wanted a chatbot and a telephony bot that could handle their basic 37 queries related to benefits and compensation for unemployed individuals.

Challenges

 

  • Intents clashing due to similarity in phrases
  • Dialogflow’s integration to CMS to receive real-time responses
  • CMS to handle real-time changes in link and its labels
  • Building accelerators to automate the changes from development environment to production environment of CMS and Dialogflow agent
  • Integration to third-party telephony operator, Cisco, for handoff to live agent
Challenges

Technologies

Dialogflow
Google Cloud Platform
Node Js
Google Sheets
Cisco

COVID-19 Question-Answering Virtual Agent for State and Federal Unemployment Benefits

Solution

Quantiphi developed and deployed a Rapid Response Virtual Agent for the IDES website and call center that effectively assisted citizens with 3.2 million inquiries in its first two weeks of going live. The virtual agent was trained on 100+ custom intents in just four weeks for responding to unemployment claim-related inquiries. Today, it is capable of handling over 120 different type of user queries. The virtual agent was also integrated with a content management system for easy update and modification of intent responses by IDES staff and Subject Matter Experts (SMEs).

Result

  • Over 5.8 million conversations have been handled by the virtual agent
  • 5.6 million website conversations
  • 286,866 voice conversations handled after hours
  • And 40,000 per night, whereas before there was no help after hours

Looking for similar project?

Let's Talk

Let’s get your Data + AI transformation started!

Let's Talk