Maestro

• 

  Light Mode

• 

  Dark Mode

• 

  Color Theme

Inspiration

We recognized the automotive retail industry's key challenges: inconsistent service quality, limited availability, outdated information, and high training costs. Our vision was to create an AI assistant that delivers consistent, professional sales service 24/7 while maintaining real-time inventory accuracy.

What it does

Our system combines RAG technology with professional sales expertise to provide 24/7 car sales consultation, offering inventory-based recommendations, maintaining professional interactions, handling special cases (like $0 prices), and guiding customers through the sales process using proven techniques.

How we built it

We developed a system integrating advanced RAG technology with a five-layer prompt architecture. Our solution utilizes sentence transformers for intelligent inventory search, incorporates professional sales knowledge through structured prompts, and maintains conversation context for natural interactions.

Challenges we ran into

• Balancing similarity thresholds for relevant recommendations
• Managing conversation context effectively
• Natural and fluent conversation
• Model illusion issue

Accomplishments that we're proud of

• Successful integration of RAG with professional sales knowledge
• Innovative five-layer prompt architecture
• Interactive design

What we learned

• The importance of balancing technical capabilities with business needs
• The value of structured prompt engineering in professional contexts
• The critical role of context management in natural conversations
• The significance of handling edge cases professionally

What's next for Maestro

• Multimodal support for visual car comparisons
• Enhanced personalization capabilities
• Real-time inventory synchronization
• Multi-language support

Built With

• flet
• llm
• natural-language-processing
• openai
• python
• rag
• semanticsearch
• sentencetransformer