AI PropTech ·
Should AI flag housing permit deficiencies before filing in Korea?
On November 12, Global News reported Nova Scotia municipal reactions to a federal budget proposal for AI-enabled permitting [1]. Budget 2026 proposes C$100 million over three years from 2027–28 to support municipalities [2].
Problem definition and research question
A Korean city building office wants fewer rounds of missing drawings and inconsistent application fields. If an automated prompt looks like an official review decision, designers may rework valid documents or disregard a real request. Which pre-filing deficiencies can be shown reliably?
Evidence and causal analysis
The Canadian budget proposes adoption support [2]. Neither the report nor the budget gives measured time savings from Nova Scotia deployments [1][2]. The Korea Real Estate Board describes Seumteo as the national online system for building and housing permit workflows [3]. A total turnaround figure mixes filing and departmental consultation, so it cannot show missing documents are the bottleneck. Drawing discrepancies and code interpretation also require different review.
| Dimension | International evidence | Korean evidence and application |
|---|---|---|
| Stage | Federal budget proposes municipal support [2] | Seumteo handles online building and housing applications [3] |
| Decision | Nova Scotia officials discuss implementation boundaries [1] | Korean pilot would distinguish document consistency from official judgment |
Comparison of alternatives
Option A uses rules to check required attachments and matching document identifiers. It is inexpensive to maintain, but staff still search inside plans. Option B compares selected fields across plans and forms and flags their locations. It can find harder discrepancies, but legibility and version mismatches create false flags. The building office owns the checklist, correction notices and permit decisions.
Implementation and evaluation
Blind-review comparable completed applications by use and complexity. Compare the current process, rules and AI preview against staff adjudication. Measure applicant effort, working days to the first deficiency request, false requests and staff recheck time. Pause outward notices if the system suggests compliance decisions or repeats false alerts on one drawing type. Reclassify that sample before proceeding.
Conclusion and open questions
Where missing materials cause the delay, a simple intake check can be tested first. Cities with long cross-department consultations may see little benefit. Can a municipality link waiting times to the reasons for first-round deficiencies?