Amazon Kendra
Definition
AWS's intelligent enterprise search service powered by machine learning, designed to help organizations find information quickly and efficiently.
Use Cases
- Vanderbilt University: Campus-wide knowledge search for IT and administrative support content — Indexed internal knowledge bases and documents and exposed a search experience that supports natural-language queries; used Kendra’s relevance tuning and connectors to keep content synchronized (Faster answers for staff and students, reduced time spent searching across multiple portals, and improved self-service for common support questions)
- Liberty Mutual: Internal search across policies, procedures, and operational documentation for employees — Deployed Amazon Kendra as an enterprise search layer over multiple internal repositories and integrated it into internal tools so employees could ask questions in natural language (Improved findability of internal information and reduced time to locate the right policy or procedure, supporting quicker customer service and operations workflows)
- Siemens: Enterprise knowledge discovery across technical documentation and internal content — Used Amazon Kendra to index large volumes of documents and provide a unified search experience with relevance ranking and filtering; integrated search into internal applications (More efficient access to technical knowledge and reduced time spent searching across disparate document stores)
Provider Equivalents
- AWS: Amazon Kendra
- Azure: Microsoft Copilot Studio (formerly Power Virtual Agents) + Azure AI Search
- GCP: Vertex AI Search (part of Vertex AI Agent Builder)
- OCI: OCI Search with OpenSearch
Frequently Asked Questions
- What's the difference between Amazon Kendra and Amazon OpenSearch Service?
- Amazon Kendra is designed for enterprise search with built-in machine learning relevance, natural-language question answering, and many prebuilt connectors (like SharePoint, Confluence, and S3). Amazon OpenSearch Service is a managed search and analytics engine where you design the index, ranking, and query experience yourself. Use Kendra when you want high-quality enterprise search with less ML and relevance engineering; use OpenSearch when you need full control over indexing, custom scoring, log analytics, or complex search architectures.
- When should I use Amazon Kendra?
- Use Amazon Kendra when you need employees or customers to quickly find answers across many document repositories (wikis, file shares, ticketing systems, intranets) using natural-language queries. It’s a good fit for help desks, HR/IT knowledge bases, compliance and policy search, legal research, and customer support portals—especially when content is spread across multiple systems and traditional keyword search returns too many irrelevant results.
- How much does Amazon Kendra cost?
- Amazon Kendra pricing depends on the edition/capacity you choose and how you use it. Key cost drivers typically include index capacity (how much content you index), query volume, and any additional features or connectors you enable. Costs can vary significantly between small departmental deployments and large enterprise indexes, so estimate based on expected document volume, update frequency, and monthly queries, and validate using the AWS Pricing page and the AWS Pricing Calculator for Amazon Kendra.
Category: ai
Difficulty: intermediate
Related Terms
See Also