Cities in Action: Affordable Housing: Artificial Intelligence and Public Portals Bring Focus

Overview

City Name: Walnut Creek, California

Partner: Accela

Problem: Like all cities, Walnut Creek maintains an archive of Planning Commission and City Council resolutions, staff reports, letters, etc. Among those countless pages and paragraphs are the details of approved building projects going back ten years and more. These are the details that bring the city’s affordable housing efforts into focus. Nested in these documents are legally defined specifics such as:

• Total Housing Units
• Affordable Housing Units
• Density Bonus
• Numbers of Units Dedicated to Extremely Low, Very Low, Low, and Moderately Low-income residents

Solution: To achieve its reporting requirements, the city dutifully added data fields to Accela forms, its system of record. For previously approved building projects, however, the new fields could only be populated by reading through the archives going back more than a decade. As with many understaffed city departments, prioritizing this laborious work was not practical. Instead, the team enabled Artificial Intelligence to read the documents, extracting the important data points from the long and tedious narrative. Once the historical data was captured in Accela, the team turned its attention to making the data available in digestible, meaningful ways.

Outcome: The project pursued an overlooked feature of Generative AI services, reading and extracting needed data elements from large document archives. The city’s Accela database now reflects key housing data dating back to 2014. The success benefits future reporting, but also makes clear that cities should explore other ways in which AI can save money and serve the community.