Enacting StatGPT

Cities strive for tools and processes that leverage data and technology to identify in real time the needs and challenges of their residents, while actively engaging with them and optimizing budgets and services. This resource guide outlines the initial steps for transitioning from traditional performance management toward the artificial intelligence (AI)-enabled “StatGPT” for city management. It draws from decades of city government experience, academic research from Harvard University, and insights from leading local government officials and private technology partners. It builds on the Transforming City Operations with StatGPT paper, written by Professor and Director Stephen Goldsmith and published in July 2025, and offers steps tested in a series of virtual and in-person workshops held by the Data-Smart City Solutions program at the Bloomberg Center for Cities at Harvard University.

Bringing together representatives from 14 large US cities and private partners from leading technology companies, the workshops were designed for exploring the transformative potential of AI and generative AI (GenAI) in city performance management. Through a structured, collaborative process involving virtual presessions and an in-person design-thinking workshop, city stakeholders, technology partners, and subject matter experts identified critical opportunities for AI integration that can fundamentally reshape how cities measure, monitor, and improve their service delivery.

Cities are experiencing a paradigm shift in performance management, moving beyond traditional metrics-based approaches toward more dynamic, insight-driven systems that leverage non-traditional data sources and advanced analytics. By documenting and sharing the process through this workbook, the Data-Smart City Solutions team hopes to encourage responsible experimentation of AI tools and agents within city organizations, led by the performance teams and the chief data and information officers. It provides a template and approach for experimenting and developing prototypes, bots, or agents that can propel the organization toward StatGPT. This model is similar to Peak Academy, the city of Denver’s process improvement and innovation training and other models emerging from the open data era.

This guide serves to help city leaders and public servants create an AI-enabled StatGPT tool for city management, moving beyond theoretical discussions by developing and testing AI performance management tools.

It focuses on three primary areas for AI intervention, based on the aforementioned research and workshops: performance diagnostics and key performance indicator (KPI) measurement; standardizing work; and visualization, communication and community engagement. 

By following the process laid out in this workbook, users will: 

• Review current processes to identify a clear and actionable path for integrating AI
  to improve service delivery, operational efficiency, and community trust, and
• Learn how to develop specific, high-potential prototypes.

Download and get started with the workbook here.