Elastic 9.3: Chat with your data, build custom AI agents, automate everything

Today, we are pleased to announce the general availability of Elastic 9.3 as the latest version of the Elasticsearch Platform — the world’s most popular open source platform for transforming both structured and unstructured data into trusted answers and outcomes.

In addition to including new features that help developers with context engineering and agent building, Elastic 9.3 introduces a broad set of new capabilities to Elastic Search & AI, Elastic Observability, and Elastic Security.

Highlights for Search & AI in 9.3:

• Elastic Agent Builder, now generally available,delivers a set of AI-powered capabilities that enable developers to natively chat with their Elasticsearch data and simplify the development of custom AI agents. And, with Elastic 9.3, Agent Builder works with Elastic Workflows — giving agents the ability to take reliable actions.

• Three Jina AI models (jina-embeddings-v3, jina-reranker-v2-base-multilingual, and jina-reranker-v3) are now generally available via the Elastic Inference Service (EIS) — delivering fast, GPU-accelerated multilingual embeddings and high-precision reranking. Even better, moving forward, all new state-of-the-art Jina AI models will be provided through the Elastic Inference Service as they become available.

• GPU-accelerated vector indexing has arrived as a technical preview. By integratingNVIDIA cuVS, an open source library for GPU-accelerated vector search and data clustering, into Elasticsearch, self-managed Elastic customers can now accelerate data indexing using NVIDIA GPUs. These customers can expect to see a 12x improvement in indexing throughput and 7x faster force merging. Offloading vector indexing to GPUs frees up CPU cycles that can be redirected to improve search performance.

• EIS via Cloud Connectis now generally available for self-managed customers. Cloud Connect now enables self-managed clusters to offload inference to managed GPU infrastructure in Elastic Cloud, while keeping data, storage, and indexing local. Users can deploy models without provisioning GPUs or managing model operations, making semantic search, multilingual retrieval, reranking, and agentic retrieval augmented generation (RAG) workflows possible in private or regulated environments.

Find more details in the blogs linked above and in the Search & AI 9.3 release notes.

Highlights for Elastic Observability in 9.3:

• Pattern-based compression for log messages via the introduction of the new pattern_text field type (now generally available) reduces the storage footprint of message fields by up to 50%.

• Available as a technical preview, customers can now benefit from up to a 5x reduction in ES|QL query latency on metrics data as well as richer and more comprehensive analytics with more time series aggregation commands, exponential histograms, and lightweight metrics downsampling.

• Also now available as a technical preview, the new Amazon Bedrock AgentCore integration provides users with an end-to-end observability solution for agentic AI applications running on Amazon Bedrock AgentCore.

• Improvements to Elastic Streams, available now as a technical preview, use an agentic workflow to enable users to parse logs directly from the message field in a log document with just the click of a button.

• Elastic Workflows, available now as a technical preview, gives operators the ability to create workflows that can be used for orchestration and autoremediation. It also integrates with Elastic Agent Builder, enabling operators to create tools that can integrate with external systems and create truly agentic workflows for both investigations and root cause analysis.

Find more details in the blogs linked above and in the Elastic Observability 9.3 release notes.

Highlights for Elastic Security in 9.3:

• Automatic Migration for Rules has expanded its capabilities to include support for QRadar SIEM (tech preview), making it easier than ever for customers to migrate to and adopt Elastic’s AI-driven detection and investigation security solution.

• Entity Analytics now offers Entity AI Summary (generally available) — an AI generated summary of entity risk scores with clear recommended actions based on associated anomalies, vulnerabilities, misconfigurations, and asset criticality. 

• For security use cases, Elastic Workflows provides a native engine for scripted automation to eliminate manual triage and execute reliable response actions. It also integrates with Elastic Agent Builder, enabling analysts to invoke AI agents grounded in their operational data and context to reason through complex investigations. In fact, with Elastic 9.3, Agent Builder features the next leap in conversational AI, a new prebuilt threat hunting agent that unifies Alerts, Attack Discovery, and Entity Risk Scores to accelerate investigations.

• Automatic Gap Filling (generally available) provides security teams with new functionality to automatically backfill detected rule execution gaps; think: fast recovery for any potentially missed alerts and fewer false negatives.

• And, in terms of endpoint security, Elastic Security now includes several new features (all generally available): Memory Dump (as a new Windows response action), two new osquery extensions (Browser History and Amcache), and a new library of prebuilt osquery queries and packs.

Find more details about these features in the Elastic Security 9.3 release notes.

The release and timing of any features or functionality described in this post remain at Elastic's sole discretion. Any features or functionality not currently available may not be delivered on time or at all.

In this blog post, we may have used or referred to third party generative AI tools, which are owned and operated by their respective owners. Elastic does not have any control over the third party tools and we have no responsibility or liability for their content, operation or use, nor for any loss or damage that may arise from your use of such tools. Please exercise caution when using AI tools with personal, sensitive or confidential information. Any data you submit may be used for AI training or other purposes. There is no guarantee that information you provide will be kept secure or confidential. You should familiarize yourself with the privacy practices and terms of use of any generative AI tools prior to use. 

Elastic, Elasticsearch, and associated marks are trademarks, logos or registered trademarks of Elasticsearch B.V. in the United States and other countries. All other company and product names are trademarks, logos or registered trademarks of their respective owners.