DuranGo AI

AI-First Local Commerce Intelligence Platform
Connecting users with local businesses through conversational AI

Overview

DuranGo AI is an AI-driven platform that redefines how users discover and interact with local commerce.

Instead of relying on traditional interfaces (maps, filters, directories), the system uses a conversational agent that understands intent, processes context, and generates personalized consumption routes in real time.

The platform is designed for Durango, Mexico, focusing on improving digital inclusion and economic visibility for local businesses.

Problem

Local commerce ecosystems face critical limitations:

• Lack of digital presence among small businesses
• Fragmented and unstructured information across social platforms
• High technical barriers for business registration
• Inefficient discovery experience for users

This results in a disconnect between available supply and potential demand.

Solution

DuranGo AI introduces an AI-First system where:

• The agent is the core system, not an auxiliary feature
• Users interact through natural language
• The system dynamically generates context-aware recommendations
• Routes are constructed based on location, time, and preferences

Core Features

Intelligent Business Discovery

Contextual recommendations based on user intent, location, and preferences.

Voice-Based Business Registration

Enables merchants to register their business in under 2 minutes using natural speech.

AI-Generated Consumption Routes

Dynamic route generation optimized by time, budget, and user context.

Conversational Interface

Eliminates the need for menus and filters through natural interaction.

Social Content Integration

References external social content without duplicating or storing it.

Architecture

The system follows an AI-First architecture, where all interactions are processed through a central agent.

Components

• Central AI Agent Handles intent detection, recommendation generation, and route construction.

• Business Agent Processes merchant registration and structures business data.

• User Agent Interprets user queries and generates responses.

• Frontend Layer (Next.js) Renders agent decisions (maps, routes, results).

• Backend (Supabase) Manages persistence of users, businesses, and sessions.

• Rendering Engine (CopilotKit) Translates agent outputs into UI actions.

Tech Stack

Layer| Technology| Purpose AI| Gemini API| NLP, reasoning, generation Frontend| Next.js| UI & rendering Backend| Supabase| Database & auth Agent Render| CopilotKit| Action execution Maps| Leaflet / Google Maps| Visualization Dev Accel| Antigravity| Code generation

Design Principles

• AI-First Architecture
• Conversational Interaction Model
• Minimal Friction UX
• Digital Inclusion Focus
• Real-Time Context Awareness

Impact

Economic

Increases visibility and transactions for local businesses.

Social

Reduces digital barriers for non-technical merchants.

Tourism

Enhances discovery of culturally relevant local experiences.

Differentiation

Feature| DuranGo AI| Google Maps| Yelp Voice Registration| ✅| ❌| ❌ Conversational Agent| ✅| Limited| ❌ AI Route Generation| ✅| ❌| ❌ Local-first Focus| ✅| ❌| ❌ Social Content Integration| ✅| ❌| ❌

Project Status

«MVP (Hackathon Prototype)»

The system is currently in the prototyping phase with a focus on validating the AI-first interaction model.

Future Work

• Real-time data enrichment
• Advanced recommendation models
• Multi-city scalability
• Merchant analytics dashboard

License

This project is developed for research and prototyping purposes.

Author

Developed as part of a hackathon project focused on AI-driven local commerce systems.