Salesforce Data Cloud offers powerful search capabilities to help businesses unlock the full potential of their data. By building search indexes on unstructured, semi-structured, or structured data, organizations can retrieve meaningful insights, power AI-driven applications, and deliver exceptional user experiences. In this guide, we’ll explore the supported Search Index Types and how to choose the best one for your needs.
Empowering AI with Search in Data Cloud
Generative AI becomes exponentially more valuable when grounded on customer-specific data. Salesforce applications like Agentforce, Tableau, and Flow Builder leverage search in Data Cloud to retrieve relevant CRM data, ensuring:
- Accurate and relevant AI-generated outputs.
- Deeper insights for analytics.
- Streamlined and efficient automation workflows.
This is achieved using search indexes, which allow applications to extract semantic and lexical similarities from user queries. By aligning AI-generated outputs with user intent, search in Data Cloud enhances personalization, relevance, and precision.
Types of Search Indexes in Data Cloud
Data is ingested into Data Cloud from various sources. This ingestion is typically performed using Data Streams, which create pipelines that bring data from CRM systems, external databases, and other platforms into the Data Cloud environment. Once ingested, the data can be indexed and searched using vector or hybrid search techniques.
Data Cloud supports two main types of search indexes:
1. Vector Search
Vector search, also known as semantic search, retrieves data based on semantic similarities to the search query. It is especially useful for:
- Long-form queries seeking general information.
- Unstructured data like call transcripts, videos, and audio files.
How it Works:
- Data is ingested into Data Cloud from various sources.
- The data is chunked and transformed into vector embeddings.
- These embeddings are stored in a vector index, enabling semantic matches for queries.
2. Hybrid Search (Beta)
Hybrid search combines semantic understanding (vector search) with keyword-based matching to deliver the most precise and relevant results. This method is ideal when:
- Queries include specific keywords alongside general context.
- Precision and comprehensive matching are required.
How it Works:
- Data is ingested, chunked, and vectorized.
- Two indexes are created: a vector index and a keyword index.
- A fusion ranker function merges results from both indexes to rank the most relevant matches.
Choosing the Right Search Type:
Vector Search:
- Best for long queries (more than five words).
- Ideal for general information or context-heavy searches.
Hybrid Search:
- Suited for queries with a mix of specific keywords and contextual content.
- Provides the most accurate results by balancing semantic and keyword matches.
Search indexes in Salesforce Data Cloud can also support advanced customer analytics by enabling faster lookup of unified customer data. When organizations implement Identity Resolution, multiple records belonging to the same customer are linked together to form a unified profile. Indexed datasets make it easier for AI applications and analytics tools to retrieve relevant identity data quickly.
Salesforce Data Cloud provides a comprehensive platform for ingesting, transforming, unifying, and activating customer data. To explore the complete capabilities and architecture of Salesforce Data Cloud, visit our Salesforce Data Cloud page.
Summary
Salesforce Data Cloud supports Vector Search for contextual, long-form queries and Hybrid Search for combining semantic and keyword matches. These search indexes enable businesses to ground AI on their data, ensuring accurate insights, relevant outputs, and efficient workflows.
For additional questions on Experience please reach out to support@astreait.com