Tech Stack
Interior Vision AI – The Future of Personalized Home Design
#Usecase
Overview
Interior Vision AI is an Agentic Generative AI interior design assistant that reimagines how users discover, design, and shop for home furniture.
Instead of scrolling endlessly through furniture catalogs, users simply describe how they want their space to feel — for example, “I want to set up a warm and fresh workspace.”
The AI agent then initiates a guided conversation to understand the user’s preferences, analyzes uploaded room photos, curates matching IKEA furniture, and generates a photorealistic visualization of the redesigned room.
Users can continue refining their space — asking the AI to add, remove, or swap items, or even re-optimize the layout under a specific budget.
Problem Statement
Most e-commerce furniture platforms fail to bridge the gap between imagination and visualization. Users struggle to picture how furniture would actually look in their own spaces.
Static product photos and manual filtering cause:
- Poor engagement and decision fatigue.
- Inconsistent personalization.
- Missed opportunities for upselling or cross-selling.
Interior Vision AI addresses these issues by combining conversational AI, computer vision, and generative design, transforming furniture shopping into a personalized, interactive, and visually guided experience.
Users struggle to picture how furniture would actually look in their own spaces. Static product photos and manual filtering cause:
- Poor engagement and decision fatigue.
- Inconsistent personalization.
- Missed opportunities for upselling or cross-selling.
Goal
To create an intelligent conversational design assistant that redefines how users visualize and purchase home furniture by merging:
- Generative AI for creative visualization.
- Computer Vision for spatial understanding.
- Product Recommendation Systems for personalized suggestions.
The goal is to turn ideas into visual realities and directly connect users to buyable products that match their aesthetic, space, and budget.
- Generative AI for creative visualization.
- Computer Vision for spatial understanding.
- Product Recommendation Systems for personalized suggestions.
The goal is to turn ideas into visual realities and directly connect users to buyable products that match their aesthetic, space, and budget.
Solution
A multi-agent system that collaborates to deliver an end-to-end, personalized design journey:
- Understands user intent through natural conversation.
- Analyzes uploaded room photos to detect layout, lighting, and tone.
- Retrieves and recommends IKEA products that match the user’s style.
- Generates photorealistic visualizations of the redesigned space.
- Iterates based on feedback (“make it cozier”, “add plants”, “fit under ₹30,000”).
- Connects to checkout, enabling users to buy the selected furniture instantly.
A multi-agent system that collaborates to deliver an end-to-end, personalized design journey:
- Understands user intent through natural conversation.
- Analyzes uploaded room photos to detect layout, lighting, and tone.
- Retrieves and recommends IKEA products that match the user’s style.
- Generates photorealistic visualizations of the redesigned space.
- Iterates based on feedback (“make it cozier”, “add plants”, “fit under ₹30,000”).
- Connects to checkout, enabling users to buy the selected furniture instantly.
Agents
Agents
- Intent Understanding Agent – Extracts design goals, preferences, and constraints.
- Room Analysis Agent – Uses vision AI to analyze user-uploaded room photos.
- Product Retrieval Agent – Searches IKEA’s catalog via embeddings and metadata.
- Design Generation Agent – Uses GenAI to create room visualizations.
- Budget Optimization Agent – Recommends cost-effective alternatives.
- Feedback Iteration Agent – Updates designs in real time based on user feedback.
Agents
- Intent Understanding Agent – Extracts design goals, preferences, and constraints.
- Room Analysis Agent – Uses vision AI to analyze user-uploaded room photos.
- Product Retrieval Agent – Searches IKEA’s catalog via embeddings and metadata.
- Design Generation Agent – Uses GenAI to create room visualizations.
- Budget Optimization Agent – Recommends cost-effective alternatives.
- Feedback Iteration Agent – Updates designs in real time based on user feedback.
System Architecture
[Image]
- LLM / Conversational Layer: Gemini 2.0 Pro via Vertex AI Agent Builder
- Vision Analysis: Cloud Vision API for room layout and lighting detection
- Image Generation: Imagen / ControlNet for photorealistic room renderings
- Product Discovery: Vertex AI Search + Recommendations AI
- Database: Firestore / BigQuery for user preferences and product metadata
- Frontend: Next.js + Tailwind (Chat interface with real-time image previews)
- Backend: Cloud Run / Cloud Functions for multi-agent orchestration
- Embeddings Store: Vertex AI Embeddings API / Pinecone
- LLM / Conversational Layer: Gemini 2.0 Pro via Vertex AI Agent Builder
- Vision Analysis: Cloud Vision API for room layout and lighting detection
- Image Generation: Imagen / ControlNet for photorealistic room renderings
- Product Discovery: Vertex AI Search + Recommendations AI
- Database: Firestore / BigQuery for user preferences and product metadata
- Frontend: Next.js + Tailwind (Chat interface with real-time image previews)
- Backend: Cloud Run / Cloud Functions for multi-agent orchestration
- Embeddings Store: Vertex AI Embeddings API / Pinecone
Impact
- Enhanced personalization: Delivered tailored room designs that reflect each user’s unique taste and mood.
- Improved engagement: Interactive GenAI-driven conversations increased session duration and user satisfaction.
- Higher conversion rates: Visualizing products in users’ own spaces led to more confident purchase decisions.
- Reduced decision fatigue: Simplified the process of furniture selection and design iteration.
- Real-world retail transformation: Demonstrated how Agentic AI can merge creativity, visualization, and commerce seamlessly.
- Enhanced personalization: Delivered tailored room designs that reflect each user’s unique taste and mood.
- Improved engagement: Interactive GenAI-driven conversations increased session duration and user satisfaction.
- Higher conversion rates: Visualizing products in users’ own spaces led to more confident purchase decisions.
- Reduced decision fatigue: Simplified the process of furniture selection and design iteration.
- Real-world retail transformation: Demonstrated how Agentic AI can merge creativity, visualization, and commerce seamlessly.