Making Local Resources Accessible for All Communities

About SWIRLIE Chatbot

SWIRLIE is the first multilingual AI chatbot designed to provide hyper-localized information about community services, tailored to the user’s postal code. Created to address barriers faced by marginalized groups, it bridges accessibility gaps and improves access to vital community resources.

About the Organization:
The Smart Waterloo Region Innovation Lab (SWR) is a government-supported initiative that aims to improve the well-being of children and youth in the Waterloo Region through innovative, community-driven solutions.

Problem:

In the Waterloo Region, families often use the Family Compass website, managed by the Children and Youth Planning Table Organization, to search for services for children, youth, and their families. However, the website’s search bar relies heavily on keyword searches and is only available in English, limiting its accessibility and usability for many community members.

My Approach:

Research & Discovery:

  • Conducted focus groups and interviews with community partners to identify challenges

  • Led comprehensive research to understand nonprofit capacity limitations and data accuracy issues

  • Gathered feedback from diverse community groups to ensure inclusive solution design

Building the Solution:

  • Managed development across three major versions, adapting to emerging technologies

  • Designed SWIRLIE to pull data from 211 Ontario's API for localized service information

  • Pivoted to incorporate LLMs (Large Language Models) when GenAI became more cost-effective

  • Personally developed a web scraping tool to collect data from 300+ nonprofit websites

  • Led team of 3-6 developers through iterative development cycles

Stakeholder Collaboration:

  • Coordinated with nonprofits, government staff, and community organizations

  • Balanced diverse stakeholder priorities while maintaining project vision

  • Established partnerships with local organizations to ensure sustainable adoption

Design & Marketing:

  • Facilitated youth feedback sessions for prototype testing

  • Produced engaging short videos for community outreach

  • Implemented multilingual functionality based on community input

  • Created location-based features using postal code mapping

Results:

  • Successfully integrated GenAI capabilities, showcasing adaptability to emerging technologies

  • Implemented postal code-based localization, enhancing service relevance

  • Established multilingual support, breaking down language barriers in accessing community services

  • 95% improvement in data accuracy.

  • 90% increase in accessibility for underserved communities.

  • Secured contracts with Laurier University for PrincipalBot and 211 Ontario for broader data scraping.

Next

PrincipalBot