evidoria

← Back to browse

Good practice Imported

Montevideo's AI Regulatory Assistant Answers Architects' Construction-Permit Questions Without Phoning City Hall

Uruguay · Montevideo · See the Uruguay profile · See the Montevideo profile

Evidence: Descriptive / self-reported Top 75% 47/100 · Ask Evidence Copilot about this practice

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 online. AGESIC confirms production since October 2024, but flags automatic answers with no human review.

4000 articles
Regulatory articles covered
6000 architects
Architects using the tool
72 hours
Fast-track permit turnaround

Details

Maturity
Scaling
Promoter
Intendencia de Montevideo (Gerencia de Permisos de Implantación y Construcción), with AGESIC and Quanam / Microsoft AI Co-Innovation Lab
Period
2024–2026 (in production since October 2024)
Keywords
digital government, urban planning, permitting, generative AI

Context

The Intendencia de Montevideo partnered with local firm Quanam and Microsoft's AI Co-Innovation Lab to help architects and technicians navigate the city's construction- and land-use-permit rules.

Objectives

The assistant aims to let users query more than 4,000 articles of Montevideo's Digesto Departamental online, replacing phone queries to the permits office.

Activities

Development began in January 2024 using natural-language processing with a Sentence Window Retrieval technique over the regulatory text, and the tool has been in production since October 2024 according to Uruguay's e-government agency AGESIC. The wider permitting reform also introduced a 72-hour fast-track ('Gestion Express') for lower-complexity applications.

Results

Quanam and the Uruguayan Chamber of Information Technologies report more than 6,000 architects already using the tool, describing its answers as more precise than human staff and citing faster responses and fewer phone queries -- claims from the vendor and industry association rather than an independent municipal evaluation.

Conclusions

AGESIC's own documentation flags that the system answers automatically with no human review before responses reach users, an acknowledged transparency gap in an otherwise institutionally tracked, production system.

Implementation
Indicative cost
Medium (€50k–€500k)
Time to results
Medium (1–3 years)
Staffing & skills
Intendencia de Montevideo (Gerencia de Permisos de Implantacion y Construccion), AGESIC, Quanam, Microsoft AI Co-Innovation Lab

Conditions for success

  • Structured regulatory corpus (Digesto Departamental) available for retrieval
  • Public-private technical partnership combining municipal domain expertise with vendor AI engineering
  • National AI-use-case registry (AGESIC) providing institutional tracking

Common failure modes

  • Automated answers are published without human review before reaching users, an acknowledged gap in AGESIC's own documentation
  • Performance and usage claims come only from vendor/industry sources, not independent evaluation

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.

Similar practices you may find useful