60% to nearly 90% accuracy in AI internal outputs
Kravet Inc. is a 5th-generation, global leader in luxury textiles, furniture, wall coverings, and home décor. With 1,000+ employees, thousands of product lines, and one of the world's largest design archives, the company manages complex, unstructured knowledge across sales, supply chain, operations, and HR.
Challenges
Thousands of files contained outdated, conflicting, or unreadable information, leading to misleading AI responses.
Product data on the website wasn't fully accessible and didn't allow exact-match search by SKU.
Unstructured formats (PDF scans, mixed file types) failed to support reliable knowledge retrieval.
AI-generated answers were unpredictable, pulling information from inconsistent sources.
Our solution
Meet Kravet's internal AI assistant—built for a global company with over 1,000 employees and designed to sift through massive amounts of unstructured data to streamline knowledge retrieval.
We proposed a zero-risk pilot to train AI on Kravet's data “as is,” allowing us to identify the root causes of inaccuracies and craft a scalable approach.
Delivered a 3-week POC with built-in RAG.
Ingested 1,000+ static files, the full Kravet blog, and product website pages.
Conducted iterative testing with client-curated test questions.
Identified problematic data sources and collaboratively redesigned the information pipeline.
The results
Accuracy gains achieved
Accuracy improved from <60% to nearly 90% through iterative, data-driven refinement.
Rapid prototype delivery
First functional prototype
delivered in 3 weeks.
Enterprise-scale knowledge
AI assistant now answers questions across 125,000+ products—with details on materials, colors, collections, and specs.
Real-time inventory
Real-time inventory insights delivered through
seamless system integrations.
Automated client communications
AI can now draft client-ready emails, combining product info + inventory checks + tone of voice selection.
