MUNICIPALITY OF PANAMÁ
Generative AI-Powered Access to Construction Regulations
Executive Summary
Grupo TX successfully implemented a specialized Generative AI solution for the Municipality focused on the Construction and Works Department's complex regulatory framework. The Department operates under numerous regulations, procedures, and technical requirements that govern construction activities throughout the capital. Previously, municipal staff struggled to provide immediate and accurate responses to inquiries from citizens, engineers, architects, and other stakeholders due to the extensive and complex nature of these regulations.
The solution deployed on an Amazon Web Services (AWS) serverless cloud infrastructure uses a Retrieval-Augmented Generation (RAG) architecture to create a conversational AI chatbot interface that provides instant access to the department's comprehensive knowledge base. Municipal employees now leverage this AI assistant to retrieve precise regulatory information through natural language conversations. Rather than searching through extensive documentation, staff members can simply ask questions and receive accurate, context-specifi c answers drawn directly from authoritative sources.
Challenge
The Municipality of Panama City faced signifi cant challenges in its Construction and Works Department. The department operates under numerous complex regulations, procedures, and technical requirements governing construction activities throughout the city. Municipal staff struggled to provide immediate and accurate responses to inquiries from citizens, engineers, architects, and other stakeholders due to the extensive and intricate nature of these regulations.
This created several problems:
● Delays in responding to stakeholder inquiries
● Inconsistent information being provided across diff erent staff members
● Difficulty in accessing specifi c regulatory details when needed
● Inefficient use of staff time spent searching through documentation
● Risk of regulatory non-compliance due to information gaps
Solution
Grupo TX developed and implemented a specialized Generative AI solution using a Retrieval-Augmented Generation (RAG) architecture to create a conversational AI chatbot interface that provides instant access to the department's comprehensive regulatory knowledge base. The solution was designed specifi cally in Spanish and trained on the complete corpus of construction-related procedures, regulations, and requirements maintained by the municipality.
The AWS-based architecture features:
● A web application interface embedded in the Tributary Management System for user access.
● Amazon Lex for processing user queries with a multi-tier response system:
-
First checking frequently asked questions with traditional AI
-
Then accessing a broader knowledge base with Amazon Bedrock if no match is found
● AWS Bedrock for leveraging pre-trained generative AI models for advanced language understanding.
● S3 bucket storage containing all Works and Construction Procedures Documents.
● Comprehensive security features including fi ne-grained IAM roles, KMS encryption, and AWS WAF protection.
The serverless architecture ensures the solution is secure, scalable, and reliable while adhering to best practices according to the AWS Well-Architected Framework.
Results
The implementation of the AI-powered regulatory chatbot has delivered signifi cant benefi ts, among them:
● Response Time Reduction: Average time to retrieve regulatory information decreased fi vefold from approximately 15 minutes to under 3 minutes.
● Staff Productivity: This in turn led to more than doubling the number of inquiries handled per staff member per day.
● Knowledge Democratization: Access to regulatory expertise extended beyond specialist staff, enabling all front-line employees to confi dently address basic regulatory questions.
“It used to be that only a few people could fi nd information about specifi c regulations sometimes, now thanks to the newly implemented chatbot it has become much easier.”
Luis Carballeda, Director of Works and Constructions, Municipality of Panamá
Lessons Learned
-
Successful deployment on existing infrastructure demonstrated the importance of designing AI solutions that complement rather than replace current systems. The integration with the Tributary Management Solution minimized disruption and accelerated adoption.
-
Tiered Response Approach: The multi-tier approach (checking FAQs before accessing the broader knowledge base) optimized costs while ensuring comprehensive coverage, a strategy that could be applied to other municipal departments.
-
Knowledge Management: The project highlighted the value of organizing and structuring existing regulatory documentation before implementation, suggesting that future AI deployments should include a preliminary knowledge organization phase.