You are here:
Use Search for AI, Automation, and Analytics
Grounding Data 360 search on unstructured and structured data enhances your use of generative AI, analytics, and automation tools across Salesforce. Grounding brings customer-specific data, including diverse forms of unstructured data (such as text documents, multimedia files, social media content, web content, chat logs, sensor data, and customer feedback), into applications like Agentforce, Tableau, and Flow Builder. This alignment results in more accurate and relevant AI-generated content, deeper insights from analytics, and more efficient automation workflows for your teams and customers.
As of October 14, 2025, Data Cloud has been rebranded to Data 360. During this transition, you may see references to Data Cloud in our application and documentation. While the name is new, the functionality and content remains unchanged.
In Data 360, you can create vector or hybrid search indexes depending on data and query needs.
Vector Search
In Data 360, you can create searchable vector embeddings from your unstructured and structured data. A vector embedding is a numerical representation of a unit of text (an article or a passage from a larger document) that captures its semantic meaning. Large language models (LLMs) group embeddings with similar semantic meanings so that generative AI applications can effectively interpret their relationships and contextual relevance.
Use vector search when your data source and search queries benefit more from semantically aware matches within a large dataset. For example, if you send a query such as “What is Acme famous for?”, the LLM retrieves chunks that have the highest vector search score that relates to the closest semantic match with the search query.
Before you create a vector search index, unstructured data stored in data model objects (DMOs) or unstructured data model objects (UDMOs) are chunked into manageable, semantically meaningful units. These units are stored in two objects: a chunk data model object (CDMO) and an index data model object (IDMO). The CDMO is a Data 360 object that stores the actual content chunks, and the IDMO stores the vector embedding.
You can perform vector searches on that data from apps such as Prompt Builder, AI agents, or Tableau.
Hybrid Search
To deliver relevant and accurate responses in generative AI applications, Large Language Models (LLMs) often require both vector and keyword search. While vector search excels at grasping semantic similarities, it can lack precision with specific domain vocabulary. Conversely, keyword search is excellent for lexical matches but doesn't understand semantic similarity. In Data 360, hybrid search combines these methods, using semantically aware vector search for contextual understanding and keyword search for precise lexical matching.
When your data source and search queries benefit from both semantically aware vector search and the precision of a keyword search, use hybrid search. For example, in the query “Does Acme motor XYZ123 use hydraulic pumps?” the addition of keyword search promotes higher-ranking positions for more relevant content, thus providing the results with better grounding.
After you create a hybrid search index on your DMOs or UDMOs, Data 360 generates content chunks and vector embeddings for a vector search index and a keyword search index. Hybrid search queries the vector search index to retrieve information based on semantic similarity, and it queries the keyword search index to retrieve information based on lexical similarity. Hybrid search then merges and ranks the results provided by both types of searches to generate the best response. You can also configure additional ranking factors, such as recency or popularity of records, to influence the search results returned.
Key Use Cases for Data 360 Search
Key use cases for Data 360 search include:
- In Retrieval Augmented Generation (RAG): Data 360 search is fundamental to RAG, grounding AI responses in relevant unstructured data. This directly leads to more accurate and contextually appropriate AI outputs that are based on your specific knowledge base.
- In Service Cloud: AI agents use search to query for precise information within knowledge articles, emails, support tickets, and chat logs. This enables them to provide more accurate answers, craft better responses, and even proactively suggest solutions to customer issues.
- In Sales Cloud: AI agents can query unstructured sales data—like emails, call transcripts, and notes—to generate highly relevant meeting briefings and summaries. This insight also fuels personalized content and customer-specific recommendations for new emails, significantly boosting customer engagement and sales team productivity.
- In automation flows: AI agents can automatically identify semantically similar records, such as Cases, based on their descriptions. This capability is crucial for flagging duplicates, accelerating resolution times, and accurately classifying records by finding similar historical examples.
- In analytics: Within Tableau, Data 360 search facilitates gaining deeper insights from unstructured data, notably through advanced topic classification. This enhances data analysis by revealing patterns and themes that traditional methods sometimes miss.
- Chunk Data
To add your data to Data 360’s search index, you must first chunk it. Chunking break your data into meaningful chunks, and Data 360 turns those chunks into machine-readable vector embeddings. - Vector Search
In Data 360, you can create searchable semantic vector embeddings from your unstructured and structured data. Vector search helps your generative AI, automation, and analytics applications understand semantic similarities and context between embeddings. - Hybrid Search
To provide relevant information to an LLM with gen AI applications and generate the correct response, requires both vector search and keyword search. Vector search understands semantic similarities and context, but lacks specific domain vocabulary. Keyword search excels at lexical similarity but not at semantic similarity. In Data 360, hybrid search combines the strengths of semantically aware vector search with the precision and speed of keyword search. - Manage Search Indexes
Create, view, edit, and delete search index configurations. - Optimizing Search Indexes: Field Selection and Chunking
When you create a search index in the advanced builder, you can optimize your search index to deliver more accurate results by paying attention to the field selection and chunking strategies you use. - Search Index Reference
Data 360 search supports many different file formats with unstructured data, embedding models, and pre-filter operations on search results. - Using Retrieval Augmented Generation
Retrieval Augmented Generation (RAG) in Data 360 is a framework for grounding large language model (LLM) prompts. By augmenting the prompt with accurate, current, and pertinent information, RAG improves the relevance and value of LLM responses for users. - Billing Considerations for Unstructured Data and Search Index
When you use unstructured data and search index configurations, your data is stored and processed in Data 360. Use of Data 360 features for unstructured data has billing implications. Use of Data 360 services impacts the consumption of credits used for billing. There are four billing components for unstructured data in Data 360: data ingestion, data storage, data processing, and data queries. Each component has a distinct applicable usage type. Note that there are two types of unstructured data connectors: connectors that reference data that resides on an external data source, and connectors that ingest data from an external data source into Data 360. Billing works differently for the two connector types.
