Sweet-Home

Inspiration

Finding a home is one of the biggest decisions people make, yet most property platforms still focus on filters like price, suburb, and the number of bedrooms.

We wanted to shift the focus from "Which house should I buy?" to "Will I actually enjoy living here?"

A home is more than a building. It shapes our daily routine, our commute, our neighbourhood, and our quality of life. We believe AI can help people make more confident housing decisions by understanding their lifestyle instead of simply matching search filters.

That's why we built SweetHome.

What it does

SweetHome is an AI-powered home discovery experience that recommends properties based on how people want to live.

Instead of starting with property listings, users first answer a short lifestyle questionnaire. The AI analyses their preferences and recommends homes that better fit their daily routines and priorities.

Users can then:

Explore AI-recommended homes Understand why each recommendation was made Visualise everyday life around a selected property Receive a personalised property report with transparent AI reasoning

Rather than overwhelming users with hundreds of listings, SweetHome focuses on helping people find homes that genuinely match their lifestyle.

How we built it

We built SweetHome as a modern web application using:

Next.js React TypeScript Tailwind CSS

The MVP uses curated mock data to demonstrate the complete end-to-end user experience while allowing us to rapidly validate the product concept.

Throughout development, we used OpenAI ChatGPT and Codex to accelerate implementation, iterate on the user experience, refine prompts, generate UI components, and improve the overall product design.

The current prototype demonstrates the complete AI-assisted workflow from lifestyle discovery to personalised property recommendations.

Challenges we ran into

One of our biggest challenges was defining the problem correctly.

Initially, we focused on recommending suburbs. Through multiple iterations, we realised the property itself should remain the centre of the experience, while lifestyle and neighbourhood provide the supporting context.

Another challenge was balancing AI recommendations with transparency. Rather than returning a simple score, we wanted users to understand why a home was recommended and what factors influenced the result.

We also intentionally limited the MVP to curated data so we could focus on validating the user journey before integrating live data sources.

Accomplishments that we're proud of

We're proud of building a complete end-to-end product experience within the hackathon timeframe.

Our MVP demonstrates:

AI-powered lifestyle profiling Personalised home recommendations Explainable AI recommendations Lifestyle exploration around each property A personalised property report

Most importantly, we created an experience that feels different from traditional property search websites by putting people before properties.

What we learned

This project taught us that designing a great AI product isn't only about building powerful models.

It's equally important to understand the user's real problem, create transparent AI experiences, and keep humans in control of important decisions.

We also learned how AI-assisted development with ChatGPT and Codex can significantly accelerate prototyping while allowing us to iterate quickly on both product design and implementation.

What's next for SweetHome

Our MVP demonstrates the complete concept using curated data.

Our next steps include:

Integrating real-time property listings Connecting public transport and neighbourhood data Providing live lifestyle insights using location-based services Personalising recommendations based on evolving user preferences Building a richer AI reasoning engine that explains recommendations in greater detail

Our long-term vision is to transform property search from comparing houses to understanding how people will actually live in them.

SweetHome helps people experience life before choosing a home.