Amazon Kendra is an intelligent enterprise search service designed to help users find information across connected content repositories. Instead of relying only on exact keyword matches, users can ask natural-language questions and retrieve relevant information from indexed enterprise content.
What problem does Kendra solve?
Organizations often store useful information across document repositories, intranets, file shares, and business systems. Employees may know that an answer exists but not where the document is stored or which keywords the author used. Kendra provides a managed search index over supported data sources so users can search this content from one experience.
Core concepts: index, data source, and query
- Index: the searchable collection that holds indexed content and associated metadata.
- Data source: a supported repository or connector configuration from which Kendra retrieves content for indexing.
- Document: the content item that is indexed and made available to search.
- Query: a user’s question or search request. Kendra can handle factoid questions, descriptive questions, and keyword searches.
- Metadata and access controls: document attributes and permissions help filter results and control which content a user can discover.
How a Kendra search workflow works
- Create an index in a supported AWS Region.
- Configure one or more supported data sources and the required IAM permissions.
- Synchronize the data source so content and metadata are added to the index.
- Submit user queries through the Kendra query interface or API.
- Present relevant results, excerpts, and document references to the user, respecting the application’s access-control design.
What kinds of questions can Kendra answer?
A Kendra index can support factoid questions such as who, what, when, or where; descriptive questions such as how; and keyword searches when the user’s intent is less explicit. Search quality depends on the content being indexed, metadata quality, connector configuration, and the wording of the query.
Amazon Kendra vs. Amazon Bedrock Knowledge Bases
| Capability | Primary role | Exam clue |
|---|---|---|
| Amazon Kendra | Enterprise search over indexed content | Employees need intelligent search across company repositories |
| Amazon Bedrock Knowledge Bases | Retrieve context for foundation-model applications and RAG | Ground a generative model’s response in a knowledge base |
| Amazon OpenSearch Service | Search and analytics engine with configurable indexing and querying | Build and operate a search/analytics solution with more control |
Exam traps and design considerations
- Kendra is an enterprise search service; it is not itself a foundation model that generates free-form answers.
- Indexing and querying are separate stages. A document must be brought into the index before it can be found through that index.
- Connector support, synchronization behavior, Region availability, and edition or feature differences should be checked in current AWS documentation.
- Search results must respect document permissions. Do not assume that indexing content alone automatically solves every application authorization requirement.
- Do not confuse enterprise search with RAG. Kendra returns relevant search results; a RAG application retrieves context and passes it to a foundation model for generation.
Test yourself
Question 1
A company wants employees to ask natural-language questions across a large collection of internal documents stored in supported repositories. Which AWS service is designed for intelligent enterprise search?
- Amazon Kendra
- Amazon Polly
- AWS Glue
- Amazon Rekognition
Reveal answer
Correct answer: A. Kendra provides intelligent search over indexed enterprise content.
Question 2
An application must retrieve passages from a knowledge base and provide them to a foundation model so the model can generate a grounded response. Which capability is the closest fit?
- Amazon Kendra only
- Amazon Bedrock Knowledge Bases
- Amazon Polly
- AWS CloudHSM
Reveal answer
Correct answer: B. Bedrock Knowledge Bases are designed to retrieve context for generative AI and RAG workflows. Kendra’s primary role is enterprise search.
Key takeaway
For exam scenarios, map Kendra to intelligent enterprise search across indexed repositories. Map Bedrock Knowledge Bases to retrieval that supplies context to a foundation model for RAG-based generation.
Related AWS topics
Amazon Bedrock Knowledge Bases · Amazon OpenSearch Service · Amazon S3
✦ CloudExamPro Premium
Ready for the AWS Certified AI Practitioner (AIF-C01)?
Move beyond memorizing terms. Practice with realistic, scenario-based questions and detailed explanations that help you understand the AWS AI services and concepts behind each answer.
Original practice questions. No dumps. No shortcuts.