AI-Powered HVAC Selection System

AI-Powered HVAC Selection System

In this project, I designed an AI-powered website aimed at helping users discover eco-friendly and energy-efficient HVAC (Heating, Ventilation, and Air Conditioning) system recommendations for their homes. Instead of going through complex consultations, users can simply upload a short video of their home and get HVAC recommendations that best fit their needs, budget, and sustainability goals.

In this project, I designed an AI-powered website aimed at helping users discover eco-friendly and energy-efficient HVAC (Heating, Ventilation, and Air Conditioning) system recommendations for their homes. Instead of going through complex consultations, users can simply upload a short video of their home and get HVAC recommendations that best fit their needs, budget, and sustainability goals.

Shape Image

Team

1 Product Designer, 1 Product Manager, 3 Developers

Role

Product Designer

Duration

Jun. 2024 – Dec. 2024

Tools

Figma

CONTEXT

Problem

The city of Ithaca, NY has committed to electrifying 6,000 buildings by 2030 to reduce carbon emissions. However, nearly half of local homes still rely on natural gas heating. Despite the long-term efficiency and sustainability benefits of electric HVAC systems, adoption remains lower than expected. So our design challenge is:

How might we help people confidently adopt eco-friendly HVAC systems for their homes?

USER RESEARCH

Understanding the Pain Points

Here are the pain points I gathered after:


  • Interviewed 10 homeowners and renters at different stages of HVAC use and planning.

  • Conducted competitive analysis of online HVAC retailers and installation services.

  • Mapped the end-to-end user journey, identifying major friction points. Based on these pain points,


I then identified corresponding opportunity areas that our product could work towards.

Shape Image

Complex Specs and Biased Guide

HVAC specs (BTU, SEER ratings, etc.) are hard to understand. Many felt that reviews or store reps were influenced by promotions, leading to uncertainties.

Card Image
Shape Image

Time-Consuming Process

HVAC planning process involves site visits, manual measurements, and consultations with vendors, which is very fragmented and time-consuming.

Card Image
Shape Image

Uncertain Cost and Sustainability

Users struggled to look beyond the upfront price when comparing HVAC options and were unsure whether sustainable models would deliver long-term savings.

Card Image
Shape Image
Shape Image
Shape Image
Shape Image

Plain, Unbiased Guidance

Help users easily understand HVAC information through plain, transparent insights that feel trustworthy and objective.

Shape Image

One-Stop HVAC Planning

Make the HVAC journey feel seamless by connecting every step and removing the manual work and repetition.

Shape Image

Transparent Cost & Eco Insights

Show both upfront and lifetime costs alongside energy efficiency to help users make informed, sustainable choices.

IDEATION

Mapping the New User Flow

Based on user research, technical capabilities, and product requirements, I sketched initial user flows in low-fidelity wireframes to visualize the experience.

DESIGN SYSTEM

Establishing a Clean, Modern, and Sustainable Visual Identity

As the design moved into mid- and high-fidelity, I established a clean and modern visual identity. Soft forms and a green-blue palette create an approachable, sustainable feel, while black and gray ensure clarity and readability.

ITERATION

Mode of Interaction

After defining the user journey and design system, I explored different interaction models to make the 3D planning experience intuitive and approachable.


I chose the Guided Flow because HVAC planning is unfamiliar to most users and typically a one-time or occasional task. Step-by-step guidance and clear instructions help users feel supported and confident when using the technology and making purchase decisions. This approach compensates for the lack of familiarity compared to the flexible canvas.

Shape Image
Card Image

☺ Step-by-step Guided Flow

The model sits on the left and recommendations on the right. The process is divided into small, guided steps with clear navigation and a progress indicator. All components are organized into structured bento-style sections.

✓ Structured flow for easier first-time navigation
✓ Clear inline instructions guide users at each step
✓ Lower cognitive load and increased confidence through smaller and simpler steps

Shape Image
Card Image
Card Image

Flexible Canvas

The model sits on a grid-based canvas, with the product list overlaid on the left. This layout follows the pattern of interior design tools, allowing users to freely zoom, rotate, and drag-and-drop products into the space.

✓ More flexibility in exploration and precision in placement
✓ Familiar interaction pattern makes it intuitive to interact
✗ May overwhelm users unfamiliar with HVAC planning, increasing hesitation and decision anxiety

Giving Users a Clear Sense of Progress

After choosing the Guided Flow, I needed a progress indicator to help users understand where they are in the process and navigate more easily. I chose Iteration 2 for its clarity, effectiveness, and seamless integration with the overall flow.

Shape Image

Iteration 1

✓ Clearly separates each step
✗ Visually distracting & breaks layout harmony

Card Image
Shape Image

☺ Iteration 2

✓ Clean and minimal
✓ Provides clarity without distraction

Card Image

Making Product Comparison Easier

To make product comparison easier, I refined the recommendation list design through iteration and user testing.


Initially, each product card in the list only displayed price, efficiency, and electricity cost, which are core decision drivers, to ensure a clean experience. However, user testing revealed that even with these key specs, users still felt uncertain. To address this, I added a short “reason for recommendation”: a simple, one-line explanation as a trust signal to make the recommendations feel more transparent and personal.

Shape Image

Initial Version

✓ Highlights key product data clearly
✗ Lacks explanation behind recs, leading to hesitation

Card Image
Shape Image

☺ Revised Version

✓ Builds trust with clear reasoning for each rec
✗ Slightly more text-heavy than the first version

Card Image

I also reduced image size and restructured the layout to show more products at once, improving browsing efficiency and reducing the need for scrolling, especially important since users typically compare multiple options.

Shape Image

Initial Version

✗ Too much scrolling and low browsing efficiency

Card Image
Shape Image

☺ Revised Version

✓ Fits more products per view, reducing scroll effort

Card Image
Shape Image
Shape Image

FINAL DESIGN

Step 1: Initial Setup

Users start from the landing page and click “Get Started” to begin. They upload a short video of their home and select their region, after which the system generates a 3D floor plan model of their space.

Phone Shape Image
Shape Image
Shape Image

Step 2: Indoor Unit

Based on the floor plan, region, and selected room, the system recommends a list of indoor units. Users can add products to their room model and adjust placement, or drag them directly from the list onto the model. Once they finish configuring all rooms, they can move on to the next step.

Phone Shape Image
Shape Image
Shape Image

Step 3: Outdoor Unit

Using the selected indoor units as input, the system recommends an outdoor unit and suggests the best connection layout. Users can also explore other products or adjust placements as needed.

Phone Shape Image
Shape Image
Shape Image

Step 4: Checkout + Download

After finalizing their selections, users can complete their purchase directly on the platform and download an installation plan for future reference.

Phone Shape Image

REFLECTIONS

Looking Forward & Final Thoughts

Shape Image

Expand Key Features

Add a personal center and shopping cart that let users track orders, save preferences, and manage purchases, making the platform more flexible and personalized.

Shape Image

Explore E-commerce Integration

If we partner with major e-commerce platforms, I would explore how to integrate their systems into ours to support seamless in-platform purchasing.

This project was my first experience building a complete design system from the ground up, teaching me how to create something consistent, scalable, and aligned with the overall brand. It also challenged me to turn a highly technical process like HVAC planning into an intuitive, user-friendly experience. Through this journey, I grew not only as a designer but also as a problem-solver, learning to balance usability, clarity, and impact in a product built to make technology feel delightful and empowering.