AS: Data Cloud Search & Knowledge Check

What types of data can Data Cloud search be grounded on? How does this help?

Data Cloud search can be grounded on unstructured and structured data to enhance the use of generative AI, analytics, and automation tools across Salesforce.

What 2 types of search indexes can be created in Data Cloud?

Vector or hybrid search indexes.

What can be used to define a search index in Data Cloud

A search index configuration

How can a search index be created?

What are the 3 available options?

A search index configuration can be created by navigating to the Search Index tab in the Data Cloud app.

The available options for creating a configuration are Easy Setup, Advanced Setup, and From a Data Kit.

What 3 ways do search indexes help?

Help with more accurate and relevant AI-generated content, deeper insights from analytics, and more efficient automation workflows.

What is a vector search embedding? How does one help?

A vector search embedding is a numerical representation of a unit of text, such as an article or passage from a larger document, that is retrieved or used during a response generation.

Vector search helps understand semantic similarities and context between embeddings

What can a vector search index be created for?

How does it handle the referenced data?

A vector search index configuration can be created for a data model object.

The referenced data is broken up into semantically related chunks to generate searchable vectors, which can then be used to find semantically similar items.

What is hybrid search?

Hybrid search combines vector search for semantic similarity with keyword search for lexical similarity.

It understands semantic similarities and context while focusing on specific domain vocabulary.

What two things can a hybrid search index be created for? What does it provide? What does Data Cloud generate for one?

A hybrid search index configuration can be created for a data model object (DMO) or an unstructured data model object (UDMO) to provide relevant information to generative AI applications.

Data Cloud generates a vector index and a keyword index.

What is an example of vector search finding 2 similar strings?

Vector search recognizes that “How do I reset my password?” and “How can I change my login credentials?” are semantically similar

What is an example of keyword search finding 2 similar strings?

Keyword search recognizes that Model X200 Printer and Model X210 Printer are lexically similar.

When creating a search index configuration, how are the fields to chunk and chunking strategy set?

Fields to chunk can be added and the chunking strategy can be added per field.

What does setting a vector strategy on a search index configuration do?

It measures the unstructured data for semantic relevance.

What can be assigned to an Agentforce agent to improve accuracy, add personalization, and build trust in generative AI responses.

Agentforce Data Library

Which standard action is used by an agent to answer queries based on data in the corresponding Agentforce Data Library?

Answer Questions with Knowledge

What does chunking do and what does it allow?

Chunking breaks long text into smaller, semantically meaningful chunks stored in chunk DMOs, so retrievers can pull just the information that answers a user’s question

What needs to be created to chunk and vectorize unstructured data in Data Cloud?

Search Index Configuration

What types of Einstein Search retrievers can be added to a custom prompt template?

Default and Individual Retrievers

Which chunking strategy can be utilized to ensure that meaning inherent in HTML tags is used to chunk a document into passages.

Semantic-based passage extraction

What are 3 examples of formats supported for storing unstructured data in Data 360

HTML, txt, PDF

What type of retriever supports defining filters to narrow the search focus to more relevant data

Individual Retriever

Which field in Prompt Builder shows active retrievers that can be added to a prompt template

Resource field

What can be embedded in a prompt template to search for, and return, relevant information from a data library

Retriever

What can be used to assign a Data Library to an Agentforce agent.

Agentforce Builder

Which two search types are available for creating a search index configuration in Data Cloud?

Vector Search and hybrid search

Which 3 settings are available for configuring a retriever in Prompt Builder.

Search Text, Output Fields, and Number of Results

What type of data has no specific, consistent format and cannot be easily stored in a relational database

Unstructured data

What 2 objects are used to reference unstructured data in Data 360?

UDLO - Unstructured Data Lake Object

UDMO - Unstructured Data Model Object

Which type of retriever represents a collection of individual retrievers?

Ensemble retriever

What can be enabled at the individual retriever level to allow users to compare the LLM response with the source data

Citations