NEW! SearchBlox is now an Adobe Gold Technology Partner - Add Agentic Search to Adobe Experience Manager I Get Started.

NEW! SearchBlox is now an Adobe Gold Technology Partner - Add Agentic Search to Adobe Experience Manager I Get Started.

NEW! SearchBlox is now an Adobe Gold Technology Partner - Add Agentic Search to Adobe Experience Manager I Get Started.

AMAZON KENDRA - SEARCHAI - ENTERPRISE SEARCH

AMAZON KENDRA - SEARCHAI - ENTERPRISE SEARCH

Amazon Kendra Is in Maintenance Mode.
Plan Your Search Migration.

Amazon Kendra Is in Maintenance Mode.
Plan Your Search Migration.

Amazon Kendra Is in Maintenance Mode.
Plan Your Search Migration.

Preserve what works, compare the alternatives, and prove SearchAI with your own content before moving production.

Preserve what works, compare the alternatives, and prove SearchAI with your own content before moving production.

Preserve what works, compare the alternatives, and prove SearchAI with your own content before moving production.

August 3, 2026 - 5 min read

August 3, 2026 - 5 min read

August 3, 2026 - 5 min read

Your Kendra investment should move forward - not start over.

Your Kendra investment should move forward - not start over.

Amazon Kendra entered Maintenance Mode on June 30, 2026. AWS says existing customers can keep using the service and will continue to receive support, bug fixes and security updates. Since July 30, 2026, Kendra has not accepted new customers or feature requests.

Amazon Kendra entered Maintenance Mode on June 30, 2026. AWS says existing customers can keep using the service and will continue to receive support, bug fixes and security updates. Since July 30, 2026, Kendra has not accepted new customers or feature requests.

For current Kendra teams, the real question is what comes next for connectors, metadata, permissions, relevance rules and every search experience built around them.

For current Kendra teams, the real question is what comes next for connectors, metadata, permissions, relevance rules and every search experience built around them.

A CONTROLLED KENDRA MIGRATION

Prove SearchAI before you change production.

Prove SearchAI before you change production.

Keep Kendra in production while you test SearchAI against the content, permissions and queries your users depend on.

Keep Kendra in production while you test SearchAI against the content, permissions and queries your users depend on.

Before the move

Know what must survive. Record the sources, metadata, permissions, relevance rules and queries Kendra supports today.

Know what must survive. Record the sources, metadata, permissions, relevance rules and queries Kendra supports today.

SearchAI on AWS Marketplace

During comparison

Test the real experience. Run Kendra and SearchAI in parallel with the same content, identities and judged queries.

Test the real experience. Run Kendra and SearchAI in parallel with the same content, identities and judged queries.

Compare with your own data

After validation

Move only when the evidence holds. Shift traffic in stages after relevance, access, freshness and performance meet the agreed thresholds.

Move only when the evidence holds. Shift traffic in stages after relevance, access, freshness and performance meet the agreed thresholds.

Explore SearchAI

WHY SEARCHAI

Compare the foundations behind the next search experience.

Compare the foundations behind the next search experience.

A migration is also a platform decision. Compare what your team will need after Kendra.

A migration is also a platform decision. Compare what your team will need after Kendra.

Enterprise search experience

Enterprise search experience

Managed retrieval and RAG are core. Search interfaces and features may require additional application work.

Managed retrieval and RAG are core. Search interfaces and features may require additional application work.

Keyword, vector and Hybrid Search support enterprise search within the platform.

Keyword, vector and Hybrid Search support enterprise search within the platform.

Deep intent understanding

Deep intent understanding

Semantic retrieval supports natural-language queries; decomposition and experience logic are built around the knowledge base.

Semantic retrieval supports natural-language queries; decomposition and experience logic are built around the knowledge base.

Deep intent decomposes multi-constraint queries and applies context, synonyms and business rules in the search layer.

Deep intent decomposes multi-constraint queries and applies context, synonyms and business rules in the search layer.

Connected enterprise sources

Connected enterprise sources

Seven native connectors are documented. Other sources need an S3 or custom ingestion path.

Seven native connectors are documented. Other sources need an S3 or custom ingestion path.

300+ connectors and crawlers, plus APIs, support enterprise content ingestion.

300+ connectors and crawlers, plus APIs, support enterprise content ingestion.

RAG and cited answers

RAG and cited answers

Built for managed retrieval and generation within AWS.

Built for managed retrieval and generation within AWS.

RAG works with the same governed content used for enterprise search.

RAG works with the same governed content used for enterprise search.

Knowledge graphs and agentic search

Knowledge graphs and agentic search

Graph and agent capabilities can be assembled with AWS services and application code.

Graph and agent capabilities can be assembled with AWS services and application code.

Knowledge graphs, retrieval, answers and agents operate on one governed platform.

Knowledge graphs, retrieval, answers and agents operate on one governed platform.

Permission-aware retrieval

Permission-aware retrieval

Access controls must be mapped and validated in the target architecture.

Access controls must be mapped and validated in the target architecture.

Permission-aware retrieval can support both search results and generated answers.

Permission-aware retrieval can support both search results and generated answers.

Personalization

Personalization

Personalization is designed in the surrounding application and service architecture.

Personalization is designed in the surrounding application and service architecture.

Knowledge-graph and journey signals support personalized results and recommendations.

Knowledge-graph and journey signals support personalized results and recommendations.

Enterprise and eCommerce stacks

Enterprise and eCommerce stacks

Integrations are implemented through the ingestion and application architecture.

Integrations are implemented through the ingestion and application architecture.

Enterprise connectors plus Shopify, BigCommerce, WooCommerce, Adobe Commerce and Stripe integrations.

Enterprise connectors plus Shopify, BigCommerce, WooCommerce, Adobe Commerce and Stripe integrations.

What your team operates

What your team operates

Your team assembles the complete experience across the required AWS services and integrations.

Your team assembles the complete experience across the required AWS services and integrations.

Search and AI experiences are managed through a shared application layer.

Search and AI experiences are managed through a shared application layer.

Cost model

Cost model

Knowledge base, query, model and supporting service usage shape cost.

Knowledge base, query, model and supporting service usage shape cost.

Fixed-cost licensing options are available; final cost depends on deployment and scope.

Fixed-cost licensing options are available; final cost depends on deployment and scope.

Amazon Bedrock Managed Knowledge Base

Managed retrieval and RAG are core. Search interfaces and features may require additional application work.

SearchBlox SearchAI

Keyword, vector and Hybrid Search support enterprise search within the platform.

SearchAI is not only a Kendra destination.

It is the foundation for what search becomes next.

SearchAI is not only a Kendra destination.

It is the foundation for what search becomes next.

Hybrid retrieval, RAG, knowledge graphs, personalization and agents share one governed platform across enterprise and eCommerce experiences.

Hybrid retrieval, RAG, knowledge graphs, personalization and agents share one governed platform across enterprise and eCommerce experiences.

01

Reconnect the sources your teams rely on.

Reconnect the sources your teams rely on.

Kendra may connect SharePoint, Confluence, S3, websites, databases, file systems and business applications. Rebuild those connections in SearchAI, then verify freshness, metadata and ingestion behavior for each source.

Kendra may connect SharePoint, Confluence, S3, websites, databases, file systems and business applications. Rebuild those connections in SearchAI, then verify freshness, metadata and ingestion behavior for each source.

Source reconnection checklist for a Kendra to SearchAI migration
Side by side Kendra and SearchAI relevance test example

02

Prove relevance with the queries users actually ask.

Prove relevance with the queries users actually ask.

Build a judged test set from Kendra analytics, zero-result searches and known problem queries. Run the same queries in both systems, compare the top results and record every improvement or regression before moving production traffic.

Build a judged test set from Kendra analytics, zero-result searches and known problem queries. Run the same queries in both systems, compare the top results and record every improvement or regression before moving production traffic.

03

Rebuild permissions before content goes live.

Rebuild permissions before content goes live.

Map identities, groups and document-level access in SearchAI. Test authorized and unauthorized users across results, previews, answers and citations before cutover.

Map identities, groups and document-level access in SearchAI. Test authorized and unauthorized users across results, previews, answers and citations before cutover.

Permission validation matrix for enterprise search users

SEARCHAI 12.1 · UNIFIED AI PLATFORM

One governed SearchAI foundation, from source to answer.

One governed SearchAI foundation, from source to answer.

Enterprise content flows into one governed SearchAI core, where connectors, permissions, hybrid retrieval, RAG and knowledge graphs work together. The same foundation powers enterprise search, grounded answers, ChatBot, Assist and agents - deployed on premises, in a private cloud or through AWS Marketplace.

Enterprise content flows into one governed SearchAI core, where connectors, permissions, hybrid retrieval, RAG and knowledge graphs work together. The same foundation powers enterprise search, grounded answers, ChatBot, Assist and agents - deployed on premises, in a private cloud or through AWS Marketplace.

SearchAI unified architecture from enterprise sources to search and AI experiences

Cloud / On-Premises

Private & Secure

Fixed-Cost Options

WHAT SHOULD YOU PROVE?

Six checks to run before you migrate.

Six checks to run before you migrate.

Run the same test set against Kendra and SearchAI. Record failures and regressions, then move traffic only after the agreed thresholds are met. SearchAI is also available on AWS Marketplace.

Run the same test set against Kendra and SearchAI. Record failures and regressions, then move traffic only after the agreed thresholds are met. SearchAI is also available on AWS Marketplace.

01 Search quality

Compare ranking, filters and content coverage with agreed queries.

02 Permissions

Test real roles across results, previews, generated answers and citations.

03 Grounded RAG

Confirm every answer uses the right evidence and cites it clearly.

04 Deep intent

Try constrained queries, business language and ambiguous terms.

05 Knowledge graphs

Confirm linked products, policies and cases improve context and retrieval.

06 Agentic readiness

Test approved actions, tool access, logs and audit trails before enabling agents.

TRUSTED BY 600+ ENTERPRISES

Turn the Kendra decision into a migration plan.

Turn the Kendra decision into a migration plan.

We will map the sources, metadata, permissions, integrations and search behavior you need to preserve, then define what SearchAI must prove.

We will map the sources, metadata, permissions, integrations and search behavior you need to preserve, then define what SearchAI must prove.

Enterprise team planning a search migration