Clarity Search

This article describes how Clarity's search engine works and its key features.

What is Clarity Search

Clarity Search is a vector-based search engine that processes your product catalog to enable semantic product discovery within Clarity conversations.

📘

Note

This is the default search available for all Clarity customers.

How Clarity Search works

Clarity Search transforms your product catalog into an intelligent, searchable vector database.

Product feed vectorization

For Engagement or Discovery customers, Clarity ingests your product feed and converts each product into vector embeddings. This vectorization process creates a product vector database, helping Clarity understand product relationships beyond simple text matching.

Semantic search

When a shopper searches a query, Clarity Search performs vector-based semantic matching to find products that match the query's intent. For queries like "I want something for a beach vacation," Clarity Search detects the underlying need and recommends swimwear, sunscreen, and beach accessories.

Key features

Product attribute selection

Choose which product attributes Clarity Search processes and which it ignores. Clarity can work with any subset of the provided product attributes, including trending or product popularity metrics.

👍

Note

Some attributes, like title, description, and so on, are required to generate relevant responses.

Adding computed attributes

Clarity enriches your product feed with computed attributes derived from behavioral data stored in Bloomreach Engagement. By tracking events such as product views or purchases for each product, Clarity can provide more precise answers to shopper queries.

For instance, if popularity information is available in the product feed, Clarity can answer queries such as "show me the most popular products."

Use cases

• Boost or rerank search results by trending or bestseller data.
• Create product collections based on co-viewed or co-purchased items.

Common computed attributes

• Product views (PDP views).
• Purchase frequency.
• Popularity scores.
• Best-seller indicators.
• Co-viewed products.
• Co-purchased products.

Add computed attributes to product feed

Follow the steps below to add computed attributes to your product feed:

1. Create a report with the metrics you want to add to the product feed.
2. Set up a repeated import to automatically update the product feed with these metrics.
3. The import runs regularly as per the frequency you set, augmenting your product feed with fresh data.

Boosting mechanism

Clarity Search can perform basic merchandising by boosting certain products to increase their visibility in search results. To configure boosting by a specific attribute, contact your Bloomreach representative.

Catalog enrichment

If enabled, Clarity enriches product feeds by generating new attributes or augmenting existing ones. This capability allows the agentic framework to answer user queries specific to products more effectively.

Filtering by specific product attributes

Clarity Search offers robust filtering capabilities, allowing shoppers to specify product attributes in their queries. For instance, a query like " I want red T-shirts under $50" will trigger a semantic vector search for products that are red in color and priced below $50.