Add highlights

Signed-off-by: Sayali Gaikawad <gaiksaya@amazon.com>

release-notes/opensearch-release-notes-2.13.0.md

@@ -1,9 +1,30 @@
 <h1>OpenSearch and OpenSearch Dashboards 2.13.0 Release Notes</h1>
 <h2>RELEASE HIGHLIGHTS</h2>
 
+OpenSearch 2.13 includes several new features designed to help you build AI-powered applications, along with upgrades to help you access and analyze your operational data, and new ways to support resiliency for your OpenSearch clusters.
+
 <h2>NEW FEATURES</h2>
 
-<h2>EXPERIMENTAL FEATURES</h2>
+The OpenSearch flow framework is generally available, allowing you to automate the configuration of search and ingest pipeline resources required by advanced search features like semantic, multimodal, and conversational search.
+
+AI connectors receive several enhancements, including predefined templates that let you automate setup for machine learning models that are integrated through OpenSearch connectors to APIs such as those from OpenAI, Amazon Bedrock, and Cohere. New settings for AI connectors allow users to configure timeouts and automatically deploy models.
+
+The OpenSearch Assistant Toolkit is now generally available, allowing developers to build interactive, AI-powered assistant experiences in OpenSearch that let users query their operational and security log data using natural language. 
+
+The agent framework is now generally available, allowing you to automate machine learning  tasks using agents and tools. 
+
+OpenSearch now supports guardrails for large language models (LLMs). The agent framework adds support for user-defined regex rules that can be used to filter inappropriate text generation that could be produced by integrated LLMs.
+
+You can now index quantized vectors with FAISS engine based k-NN indexes. Instead of storing vectors that require 4 bytes per dimension, you can compress the dimensions down to 2 bytes, which can reduce memory requirements and improve query latency.
+
+You can now post-filter hybrid search results, allowing you to apply aggregations to the results to support use cases such as faceting.
+
+This release includes updates to tools that enable you to query external data sources, including a new Bloom filter, a type of skipping index that can increase performance for data types like IP addresses and hostnames where there are many different values that can be stored. In addition, you can now manage tables and accelerations visually in Query Workbench, instead of using SQL statements. 
+
+This release adds I/O-based admission control, a proactive mechanism designed to support cluster resiliency by protecting against spikes or increases in capacity if a cluster’s I/O usage breaches a defined threshold.
+
+New cross-cluster monitors are added to the alerting plugin, allowing you to manage alerts across clusters. This release also introduces the option to set up a cluster that is dedicated to alerting, separating alerting resources from indexing and querying workloads.
+
 
 <h2>RELEASE DETAILS</h2>
 <ul>