Automated Government Call Centre

Project Info

Team Name

TeamDatakraft

Team Members

Andrew Goodman and 3 other members with unpublished profiles.

Project Description

Problem Statement
Government Departments are inundated with calls about their services, requiring significant resources. Callers often have to wait long times to speak to a representative to get the information they seek. Much of the information callers enquire about is on the website but they have difficulty accessing it for various reasons.

Solution Statement
Using cutting edge voice technology with AI/LLM's to create an automated phone service which answers callers questions, and has the option of sending the caller forms and links to additional resources mid conversation. It will also triage the call to a real person if it cannot adequately answer the users questions. Analytics are auto-extracted from all calls and aggregated into a dashboard to provide the department insights into the nature of calls and issues faced by callers to better direct future PR & communication efforts.

Key applications include:
Automating repetitive tasks and optimizing workflows to reduce administrative overhead
Providing real-time analytics and insights into government spending and procurement processes
Improving service delivery through AI-powered chatbots and personalized citizen engagement

Our team of three developed an AI solution—an Information Retrieval System—to enhance transparency and accountability by allowing citizens to easily access government information and services. We scrapped the information from the website, fed it into the knowledgebase in Vapi and ask questions regarding public information that are available in home affairs website. The app is pretty accurate and we have presented it in the demo video.

#automated #ai #voice agent #phone calls #administrative #customer service

Data Story

Example 1: Home affairs automated phone service - AI Governance challenge
We collected information from Home Affairs website, strored it in a structured format as CSV file, fed it to a knowledgebase and utilised LLM's with TTS and STT technology for the voicebot (vapi). The data used was statistics on immigration to Australia.
Demo Video - https://youtu.be/nL5qbpU_pjs

Challenge 2 - NT Location based Information Service
We used datasets from the NT website including crime stats, accommodation providers, tour providers, crocodile capture data, ANZAC story data, etc.
Demo Video - https://youtu.be/-gdXsqQknT8

Team DataSets

home_affairs_research

Description of Use We collect all the scrapped information, strored it in a structured format as CSV file, fed it to a knowledgebase (IR System)that we build using Vapi.

Data Set