AI Home Furnishing

Home furnishing increasingly starts outside the retail journey. People collect inspiration from Pinterest, Instagram, and saved images, but still struggle to translate what they like into something that works in their own home.

IKEA's digital experience is strong at browsing and buying, but less present in the earlier, more uncertain stage of the journey: imagining possibilities, making decisions, and building confidence. This creates an opportunity to move IKEA from a transactional endpoint to a more active design partner.

With AI, the experience can shift from asking users to manually search, filter, and assemble a room, to generating a complete proposal from two simple inputs: a photo of their room and a source of inspiration.

The initial direction was to treat AI as a way to help users identify furniture from an inspiration image and match it with IKEA products. While technically useful, this framed the experience around product extraction rather than user confidence.

The stronger opportunity was to start from the user's room and make the first output a complete transformation. Instead of asking users to select objects, configure preferences, or choose products upfront, the system would generate a full proposal first.

This led to a clear interaction model:

Generate first, refine after.

The flow was shaped around three principles:

Start with the user's space The room photo becomes the anchor for the experience. The output should feel personal, not like a generic style transfer.

Let AI make the first proposal The system takes responsibility for interpreting the inspiration, matching IKEA products, and composing a complete room.

Keep refinement guided After generation, users can explore changes through simple controls like mood and functional furniture improvements, rather than complex object-level editing.

The design direction intentionally avoids turning the product into a professional design tool. The goal is not to give users unlimited control, but to create a low-friction way to see what is possible, build confidence, and move toward action.

The experience was shaped by what the AI system could reliably support, not by an idealized version of the technology.

Behind each generated room, the system needs to run several steps in sequence: checking the user's room photo, interpreting the inspiration image, matching similar IKEA products, building a furnishing bundle, generating the room image, and validating the output.

This introduced three important constraints:

Latency The generation process takes time, so the waiting state needed to feel purposeful rather than passive.

Imperfect matching The system may not always find an exact product, color, or material match, so the interface needed to communicate alternatives honestly.

Regeneration cost Every meaningful change can require a new generation, so refinement had to be deliberate rather than endlessly granular.

These constraints pushed the design away from open-ended editing and toward a more controlled interaction model. The system makes a strong first proposal, then gives users a small set of meaningful ways to explore variations.

The final flow is intentionally simple: users add their room, add inspiration, wait through a clear generation process, then see their redesigned space before making any product decisions.

Rather than splitting the experience into multiple configuration steps, the intake is designed as one moment. The user provides two inputs on the same screen: a photo of their room and, optionally, an inspiration image. If they do not have inspiration ready, they can choose from curated IKEA examples.

From there, the system takes over. It interprets the room, translates the inspiration, matches IKEA products, and generates a complete furnished proposal.

The first output is always visual: the user sees their own room redesigned. Only after this moment do product details, swaps, and refinements appear.

The flow follows a clear sequence:

Add photos -> Generate proposal -> Explore changes -> Shop the room

This structure keeps the experience focused on the core promise: helping users move from inspiration to a confident, shoppable version of their own space.

A key part of the concept is that the generated room is not just an image - it is connected to real IKEA products.

After the room is generated, users can open the product list to see what was used in the design. From there, they can review items, swap alternatives, and move toward purchase.

This keeps the experience grounded in action. The AI output is not only inspirational; it becomes a bridge into the IKEA catalog.