Seattle's Department of Construction and Inspections piloted the CivCheck AI tool from March to October 2025, cutting average intake review time by about 50% and correction cycles by 35% for small residential permits, while flagging where AI compliance checks still fall short.
~50% %
Reduction in intake review time (Mar-Oct 2025 pilot)
~35% %
Reduction in correction cycles (Mar-Oct 2025 pilot)
87% %
Completeness-check accuracy
Details
Maturity
Pilot
Promoter
Seattle Department of Construction and Inspections (SDCI) & Innovation & Performance
Period
March 2025 – October 2025 (pilot); findings published June 2026
Keywords
urban planning, permitting, digital government, AI pilot evaluation
Context
Seattle's Department of Construction and Inspections (SDCI), with the city's Innovation & Performance team, ran a proof-of-value study of CivCheck, an AI pre-screening tool, from March to October 2025, evaluating its use for small residential building permits.
Activities
Applicants who ran their plans through CivCheck before submission were compared against those who did not. The tool's completeness checks were tested for accuracy separately from its design-compliance checks.
Results
Average intake review time fell by about 50% and correction cycles dropped by roughly 35%; completeness checks were 87% accurate and design-compliance checks 92% accurate, and 86% of pilot users reported that the tool added business value.
Conclusions
SDCI's own assessment cautioned that accuracy alone does not guarantee time savings — what matters most is whether the tool catches all needed corrections in one pass — and recommended a further production pilot focused on completeness pre-screening rather than a full rollout.
Implementation
Indicative cost
Medium (€50k–€500k) — Vendor-partnered pilot; a specific contract value is not published in the available sources.
Time to results
Short (< 1 year) — Seven-month pilot, March to October 2025; findings published June 2026.
Staffing & skills
Seattle Department of Construction and Inspections (SDCI), Seattle Innovation & Performance team
Conditions for success
Structured evaluation methodology combining accuracy testing and user surveys
Prior deployment experience from Honolulu informing the Seattle pilot
Common failure modes
Compliance checks proved harder to automate reliably than completeness checks, given the complexity of Seattle's building code
Where it fits
Governance type
municipal government department
Scale
small residential permits, single city
Income level
high income
Commonly funded by
National / regional programmes
Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.
Do you run this practice?
Claim it —
verified implementers get a public contact pathway and can propose corrections.
Data sources
Where this practice's information was retrieved from, and when.
Montevideo built a generative-AI assistant, with Microsoft's AI Co-Innovation Lab, that answers architects' questions on the city's 4,000+-article construction-permit code …
