Manufacturing Software Cuts Order-Ready Design Time from Hours to Seconds with AWS Bedrock
At a glance
Our customer’s clients were spending hours manually designing custom furniture concepts, sifting through 500-1,000 products to create personalized recommendations to buyers.
NMD deployed an AI-powered context engine using AWS Bedrock that instantly generates images of new products that match the buyer’s needs and style preferences, creating a 90%+ reduction in time to create a buyer proposal. Most importantly, the buyer could then request production of that item.
The POC moved to production and a second POC is underway, quoting production of new AI-generated furniture designs.
Industry
Use Case
- Personalization & Recommendation
- Analytics & Business Intelligence
- Natural Language Processing
- Computer Vision Applications
Solution implemented
- Deployed AWS Bedrock for AI embeddings and generative design of custom furniture products
- Built vector database similarity matching using AWS Lambda to analyze buyer-product alignment
- Implemented RAG architecture processing multimodal catalog data and buyer purchase history on S3
- Configured AWS Bedrock models to generate product images matching buyer style preferences
- Orchestrated automated workflow from design generation to quote preparation using AWS Lambda
The value equation
- 6 weeks to deploy a Personalization & Recommendation Solution
- 90%+ reduction in buyer proposal creation time achieved
- Successfully moved from POC to production
- Automated quoting for AI custom furniture added in POC #2
Company Snapshot
A 3D visualization platform for furniture and home decor manufacturers that enables interactive product customization, virtual showrooms, and AI-powered design matching to accelerate sales cycles and create immersive online shopping experiences.
Location
US
Customer Situation
Our customer, a 3D visualization platform for furniture and home decor manufacturers, was spending hours manually designing custom furniture concepts and sifting through 500-1,000 products to create personalized recommendations resulting in long times to proposal.
In the furniture manufacturing industry, sales teams typically manage 500-1,000 products per manufacturer and must manually match these against buyer catalogs of 1,000-3,000 products, creating buyer decisional overload and mismatching.
NMD Solution
NMD reviewed and found an opportunity to create a system to automatically generate new furniture designs that match buyer preferences and style that would reduce buyer dissonance and increase time-to-design relevance.
Enabling automated custom furniture design and recommendations required implementing an AI-powered generative design engine using AWS Bedrock for AI image generation and embeddings, Amazon S3 for data storage, and AWS Lambda for orchestrating the design generation and matching logic.
Solution Deployed
- AWS Bedrock-powered context engine with vector database to automatically generate product recommendations and new furniture designs
- AI image generation that creates visuals of new furniture products matching buyer style preferences and needs
- AI embeddings system analyzing profiles and product catalog data across 500-1,000 products, processing descriptions, specifications, and images
- Knowledge graphs and LLMs to understand preferences, purchase history, and style
- Similarity matching algorithms orchestrated via AWS Lambda to intelligently align products with interests
- Automated marketing content generation creating detailed product descriptions, personalized sales messages, and custom collection presentations
- Multimodal AI capabilities processing both structured catalog data and visual product information to generate entirely new product concepts
Within 6 weeks of deploying NMD’s Personalization & Recommendation Solution for Manufacturing solution, customer successfully completed POC #1 validating the technical approach for AI-generated furniture designs and product matching, creating a 90%+ reduction in time to create a buyer proposal, and is now in production.
The proven capability has now been expanded into POC #2: an automated quoting system that generates production quotes for AI-designed custom furniture, creating an end-to-end pipeline from AI-generated design concept through manufacturing quote.
Ready to Transform Your Manufacturing Business?
See how New Math Data can transform your manufacturing business with AWS-powered innovation.