Connect Salesforce Commerce Cloud with Klevu to power natural language search, AI-led relevance, merchandising rules, synonyms, and product recommendations that lift product discovery and conversion.
• Product, category, inventory, price, and attribute data is exported from Salesforce to Klevu as a normalized feed, with field mappings for titles, descriptions, images, and facetable attributes.
• Locale-specific values are mapped to Klevu store views so language, currency, and region-specific availability can be indexed separately where applicable.
• Delta updates sync changed records on a schedule or event basis, while full reindex jobs rebuild the Klevu index when structural changes occur.
• Search queries are routed from the storefront to Klevu APIs, which return ranked results, autosuggest, and zero-result handling based on the current index and relevance model.
• Merchandising rules, synonyms, and boosts are applied at query time in Klevu, with selected attributes from Salesforce used as filters, facets, and ranking signals.
• Recommendation placements call Klevu endpoints using context such as product ID, category, or cart contents, and returned SKUs are matched back to Salesforce identifiers for rendering.
• Sync and query activity is logged for monitoring, with validation on required fields (SKU, availability, and primary category) to prevent unindexable products from entering the feed.
.png)
We map Salesforce product, price, inventory, and category data to Klevu feeds so search, category discovery, and recommendations update on a set schedule or near real time.
Yes – Klevu supports synonym dictionaries, redirects, and query rules, so shoppers using natural language get relevant results instead of dead ends.
Merchandisers can set ranking rules, boosts and buries, banners, and query-specific results without changing Salesforce core logic.
It can – we structure feeds by site, locale, or currency and align attribute sets so relevance tuning stays correct per market.
Missing attributes, inconsistent category paths, and poor variant handling usually cause weak ranking – we fix the mapping and enrichment so Klevu has clean signals.






