Search behavior rarely matches catalog structure.
Customers search by requirement, occasion, and product attribute—often with typos and their own words. Zhara reads that context with natural-language understanding, typo tolerance, and synonyms, then returns the most relevant products.
Interpret product requirements.
Combine material, function, fit, and category signals to rank products that satisfy the complete request.
Multi-attribute searchForgive typos, resolve synonyms.
Typo-tolerant search and synonym groups connect customer language such as “office backpack” with catalog terms including laptop bags and commuter backpacks.
Typo tolerance & synonymsRank results by contextual relevance.
An occasion-based query can surface coordinated apparel and accessories even when the exact phrase is absent from product titles.
Semantic relevance