Defining Where AI Could Improve the Furniture Shopping Experience
Rooms To Go wanted to understand where AI could meaningfully improve furniture shopping, not simply add AI features to the experience. I led product design across research, concept development, interaction design, and user validation to explore how search, personalization, and shoppable rooms could help customers move from inspiration to purchase with greater confidence.
Project Details
Lead Product Designer
Led product strategy, competitive research, storyboarding, user flows, interaction design, prototyping, and research synthesis from early exploration through validation.
Cross-functional Team
Product Manager, Product Designer, Retail Stakeholders, AI/ML Engineers
6 months
Product discovery, AI opportunity framing, concept development, interaction design, prototyping, and iterative user validation.
The Challenge
Furniture shopping creates a unique decision problem. Customers are choosing expensive products they may keep for years, but they often have to imagine how individual pieces will look together, fit their space, and work with what they already own.
The existing digital experience gave shoppers access to thousands of products but limited support for making those decisions. Working with product and retail stakeholders, we identified three recurring barriers:
Too much choice: Large catalogs made it difficult to know where to begin or narrow options with confidence.
Limited context: Individual product imagery did not help customers understand how pieces would work together in a room.
A disconnect between inspiration and purchase: Customers could find ideas online without knowing whether the products were available, affordable, or nearby.
The opportunity was not simply to introduce AI. It was to determine where intelligent assistance could reduce decision-making effort while keeping the shopping experience understandable and connected to real inventory.
Looking Beyond Furniture for Proven AI Patterns
Furniture retail had relatively few established AI interaction patterns, so I expanded the research beyond direct competitors. I studied products across fashion, fitness, search, and interior design to understand how other industries were using AI to reduce choice, personalize discovery, and translate vague intent into useful recommendations.
I evaluated the products against three questions:
How does the user communicate what they want?
What does AI contribute that traditional filters or navigation cannot?
How does the product maintain user control when the recommendation is imperfect?
Three opportunity areas consistently emerged: visual search, preference-based personalization, and conversational discovery.
Products such as Google Lens demonstrated how imagery could become an input rather than simply an output. For furniture, this suggested an opportunity for customers to begin with something they already liked instead of knowing a product name or category.
Visual search
Stitch Fix and Nike demonstrated different ways of learning customer preferences over time. The key takeaway was that personalization should reduce future effort, not require customers to repeatedly configure the same preferences.
Personalization through style preference quizzes
Conversational products showed how natural language could capture needs that traditional filters struggle to express, such as “a cozy sofa for a small living room” or “a dining table that works with dark walnut.
Conversational AI
Product Decision:
Where should AI actually appear?
The research surfaced many possible applications for AI, but adding intelligence everywhere would increase complexity rather than reduce it.
We prioritized concepts where AI could solve a specific limitation of traditional ecommerce: understanding search intent, learning customer preferences, connecting inspiration to inventory, and helping customers evaluate products in context.
Mapping AI to Real Shopping Behaviors
This storyboard illustrates three distinct customer personas and how they engage with the home design experience, each with unique needs, concerns, and goals. The framework maps their journey from initial mindset to the features that best support them.
Turning Concepts Into End-to-End Product Flows
Once the opportunity areas were defined, I mapped the full end-to-end experience. This included integrating new entry points, handling edge cases, and extending the visual system to reflect the tool’s dynamic logic.
Defined user flows for first-time use, returning users, and fail states
Created modular screens that flexed based on user input and backend data
Updated visual system to accommodate AI-specific UI components and guidance cues
Designing AI as Part of the Shopping Experience
From early concepts, we moved into defining the core product flows that would bring AI to life in a shoppable experience. These flows explored how customers could search naturally, save favorites, and purchase entire room sets, while ensuring everything stayed aligned with retail inventory.
The personalized home screen acted as a central hub for exploration, adapting to each shopper’s taste and history. AI-driven modules surfaced inspiration, curated rooms, and saved favorites in one unified view.
AI-Powered Search
Search transformed product discovery from keyword matching into intent recognition. Customers could describe what they needed in their own words and see results organized into meaningful sets.
Shop by Room
We turned styled inspiration into an actionable shopping flow. Instead of browsing disconnected SKUs, customers explored complete rooms and tapped directly into product details, pricing, and availability.
User Favorites
We gave shoppers a way to pause decision-making without losing momentum, helping reduce abandonment and drive return visits. By tying saved items back to live inventory, customers always knew what was available when they came back.
User Testing & Design Validation
To validate usability and ensure alignment with user expectations, a series of moderated testing sessions were conducted, focusing on key features and flows. Testers navigated through a prototype, completing tasks while providing live feedback on functionality and design. Tester qualifications included age, region, familiarity with the Rooms To Go brand, and recent experience shopping for furniture.
In the first round, five testers evaluated the prototype, and their feedback was synthesized to inform design updates. The updated prototype was then tested by four additional users to validate the changes and identify any remaining concerns.
Customer Validation
User testing showed that shoppers felt more confident when AI suggested complete rooms rather than single items. Natural language search reduced frustration and helped customers find products faster.
Business Alignment
By tying AI flows to real inventory and promotions, we created a seamless bridge between online discovery and in-store sales. This alignment positioned AI as both a shopper benefit and a retail growth driver.
Launching AI search, shoppable rooms, and favorites delivered immediate impact on engagement and conversion. These flows laid the foundation for future enhancements like expanded personalization.
Quick Wins
This project reinforced that designing with AI is less about creating new interaction patterns and more about deciding when intelligence actually improves an existing behavior.
The strongest concepts did not ask customers to “use AI.” They helped them describe what they wanted more naturally, discover products in context, preserve progress across sessions, and move from inspiration to real inventory with less effort.
User testing was equally important in showing us where AI should step back. Removing generated imagery and simplifying prompt behavior made the experience more familiar and trustworthy without reducing the intelligence behind it.
Reflection
3
Distinct product vision tracks developed to meet key business goals: in-store integration, self-service utility, and inspiration-driven browsing.
6
Future-forward feature concepts ranging from AI decor quizzes to AR room builders, UGC-style galleries, and smart post-purchase delivery tracking.
156M+
Estimated app market opportunity based on analysis of top competitors’ download volume used to support ASO and product roadmap prioritization.