Michele Fraga Spadini
Create Your First Project
Start adding your projects to your portfolio. Click on "Manage Projects" to get started
UX Writing | Visual Arts for the Department of Transportation
Project type
Conversational Design
Date
Since August 2025
Location
Brasilia
Project Overview:
Multichannel chatbot (web and WhatsApp) developed for the Brasília Department of Traffic, focused on automating high-demand services, reducing ambiguity around public services, and providing clear, guided journeys for citizens and accredited companies.
Target Audience:
- Citizens of Brasília who own vehicles, as well as companies accredited by DETRAN-DF.
Objectives:
- Reduce and eliminate the need for human support;
- Increase user satisfaction;
- Minimize ambiguity in institutional and legal information;
- Guide users across services, documents, and accredited companies;
- Facilitate consultations and actions such as traffic fine payments, issuance of a duplicate driver’s license, and password reset for accredited companies (Dircrep).
Conversational Strategy:
- Design of structured and closed flows, including guided journeys supported by micro-agents;
- Segmentation of responses to avoid information overload;
- Use of contextualization for queries related to documentation, appeals, and specific services.
Metrics Tracked:
- Volume of automated interactions;
- User satisfaction;
- Most frequently accessed topics;
- Distribution of interactions by journey type.
Results and Learnings:
The adoption of clear and guided conversational flows enabled the chatbot to reach more than 60,000 monthly users, with 82.4% positive evaluations, highlighting the importance of clarity and disambiguation in digital public services. Step-by-step guidance across journeys significantly reduced reliance on human support and enabled efficient resolution of recurring demands, such as traffic fine inquiries, duplicate driver’s license issuance, and support for accredited companies.
The project demonstrated that, in large-scale governmental contexts, strong results are more closely related to information organization, context control, and conversational design decisions than to the technical complexity of the solution itself.

