Transforming Local Government Services with Generative AI

The Challenge

A Local Council Government Body was grappling with significant challenges in its administrative operations, primarily stemming from high staff turnover and inefficient information management. Front-line employees had an average tenure of just 6 months, resulting in a constant drain of institutional knowledge and extended lead times for new hires to reach full productivity. This turnover severely impacted service quality and consistency.

The council's information retrieval process was equally problematic, with staff struggling to access accurate data from multiple source systems. On average, it took 12 minutes to find key knowledge insights, with only a 68% accuracy rate in first-time responses. These inefficiencies led to prolonged response times and frequent dissemination of incorrect information to citizens.

The Solution

To address these challenges, WeBuild-AI engineered a comprehensive, state-of-the-art Generative AI solution tailored to the Local Council's needs.

At the core of this solution was a lightweight deployment of our Pathway platform, which seamlessly integrated several key technologies. Streamlit was utilised to create an intuitive user interface, ensuring ease of use for council staff. AWS Bedrock was implemented to provide a robust LLM backbone, while Anthropic's Claude 3 served as the sophisticated core language model.

To support this advanced AI system, a vector database architecture was constructed using higher-level AWS services, enabling efficient data storage and retrieval. Security was paramount in the design, with the implementation of AWS Control Tower, establishing a secure Landing Zone for the customers data and operational Generative AI platform.

This multifaceted approach culminated in the development of a scalable, AI-based knowledge management system, capable of evolving with the council's future needs and use cases.

Outcome

The implementation of the Generative AI solution yielded transformative results for our customer, dramatically enhancing the efficiency and effectiveness of the Local Council's operations.

The most striking improvement was the reduction in mean time to discovery for key knowledge insights, plummeting from 12 minutes to a mere 30 seconds - a 96% decrease. This was coupled with a significant boost in the accuracy of first-time knowledge responses, soaring from 68% to an impressive 95%.

These enhancements translated into tangible improvements in service delivery, successfully managing an average of 200 daily enquiries over a seven-day work week. The focus on optimising both telephony channels and in-person enquiries at the councils head quarters ensured a comprehensive upgrade of citizen-facing services.

Beyond these immediate benefits, the implementation established a scalable architecture capable of supporting a wide array of use cases across the council's functions, including tourism and other knowledge-intensive citizen support services.

This successful deployment has positioned the Local Council at the vanguard of AI-driven public service delivery, enabling faster and more accurate responses to citizen inquiries while steadfastly maintaining a human-centric approach to governance. The result is a more responsive, efficient, and citizen-focused local government, setting a new standard for public service in the age of AI.