Car Buying Assistant

Buying a car is one of the most complex consumer decisions people make. Buyers must compare dozens of models, evaluate reliability and performance, understand pricing differences, and navigate dealership information - often across many scattered sources.

Most people research vehicles through a mix of forums, review sites, dealership listings, and YouTube videos. This creates information overload and makes it difficult to confidently identify which vehicles actually fit their needs.

I saw an opportunity to simplify this process by creating a tool that translates user preferences into personalized vehicle recommendations.

While exploring how people research vehicles, I noticed that most buyers don't start with a specific model in mind. Instead, they think in terms of preferences:

• budget
• performance vs reliability
• body style
• fuel efficiency
• practicality

Existing car marketplaces are built around browsing listings, not helping users discover what they should buy in the first place.

This suggested an opportunity for a pre-search discovery tool that guides users toward vehicles that match their priorities before they ever start browsing listings.

I built the Car Buying Assistant, a prototype web application that guides users through the car discovery process and generates personalized vehicle recommendations.

Instead of requiring users to manually compare hundreds of vehicles, the assistant asks users about their priorities - such as budget, performance preference, reliability expectations, and lifestyle needs - and then recommends vehicles that best match those criteria.

The goal was to explore how recommendation logic and conversational interfaces could simplify early-stage car buying decisions.

• Preference-Based Recommendations → Users input preferences such as budget, body style, reliability, and performance priorities to generate tailored vehicle suggestions.
• Structured Vehicle Knowledge Base → Vehicle information such as horsepower, drivetrain, fuel economy, and reliability considerations are organized into a structured dataset to support recommendations.
• Conversational Interface → Users can interact with the assistant conversationally to explore options and ask questions about vehicles.
• Explorable Results → Recommended vehicles include key specs and insights to help users quickly compare options.

I approached this project as an exploration of how AI-assisted development tools can accelerate product prototyping.

First, I mapped the core user journey: how someone moves from "I need a car" to a short list of realistic options. From there, I defined the inputs that would most strongly influence recommendations, such as budget, reliability priorities, and performance preferences.

I then built a working prototype using v0, which allowed me to quickly design the interface and test the user flow.

In parallel, I experimented with building a backend using Next.js and Supabase to support structured vehicle data and recommendation logic. This helped me better understand the technical requirements for scaling a product like this beyond a prototype.

This project helped me explore how recommendation systems can simplify complex consumer decisions.

One key challenge was translating subjective preferences like "fun to drive" or "reliable" into structured attributes that a recommendation system could evaluate.

I also learned how quickly AI-assisted tools can accelerate early-stage product development. Using v0 allowed me to move from concept to working prototype much faster than traditional development workflows.

If I continued developing this project, the next steps would include expanding the vehicle dataset, refining the recommendation logic, and integrating real-time market data.

In the long term, a system like this could evolve into a full car discovery platform that integrates real-time inventory, pricing insights, and user feedback to continuously improve recommendations.