Amazon

A technical walkthrough shows how to secure AI agents on Amazon ECS using Amazon Bedrock AgentCore Identity by implementing the OAuth 2.

A new Exa integration for the Strands Agents SDK exposes two tools — exa_search and exa_get_contents — via the strands — agents-tools package, returning structured page text and metadata ready for LLM consumption

A new Amazon Quick release introduces capabilities that let organizations query large, governed enterprise datasets in natural language and get fast, auditable SQL-backed answers.

Anthropic’s Claude Platform is now generally available on AWS, letting customers access the native Claude experience directly through their AWS accounts without separate credentials, contracts, or billing relationships.

Miro rolled out BugManager, an AI bug-triage system built with AWS PACE and running on Amazon Bedrock that uses RAG and multimodal parsing with Anthropic Claude Sonnet 4.

Halliburton and the AWS Generative AI Innovation Center built a proof — of-concept assistant that converts natural — language queries into executable seismic processing workflows and adds a question — answering layer

Published guidance for builders who need predictable, time‑bound GPU capacity for ML workloads, comparing on‑demand, Spot, and EC2 Capacity Blocks for ML reservation modes.

A technical post describes using reinforcement learning with verifiable rewards (RLVR) on SageMaker AI, combined with Group Relative Policy Optimization (GRPO) and few-shot examples, to make reward signals more

Tomofun reconfigured its inference fleet so BLIP vision — language models can run on Amazon EC2 Inf2 (Inferentia2) instances, keeping its GPU path and client APIs while lowering operational costs for continuous

OS Level Actions let AgentCore Browser control full‑desktop UI through the InvokeBrowser API: Each call carries one action and returns SUCCESS or FAILED; screenshots are returned as base64‑encoded PNGs.

Hapag‑Lloyd’s Digital Customer Experience and Engineering team replaced a manual, spreadsheet — driven feedback review process with an automated generative AI pipeline that ingests comments, extracts sentiment

AgentCore Optimization (preview) introduces a Recommendations API plus batch evaluation and live A/B testing to close the observe — evaluate–improve loop for production agents, automating detection and remediation

At an April 28 event, AWS executives outlined agentic AI updates: Amazon Quick gains a desktop app, broader integrations and an app‑building preview; Amazon Connect is split into four domain solutions;

Amazon Quick added a generative AI tool that converts natural‑language prompts and selected datasets into a native Analysis that can be refined and published as a dashboard.

AgentCore Optimization, now in preview, closes the observe — evaluate-improve loop for AI agents by generating recommended fixes from production traces and validating them with batch evaluation or controlled A/B tests

AWS has built TARA, an internal conversational analytics assistant that uses the Dataset Q&A feature of Amazon Quick to let technical field leaders ask multi‑dimensional questions in natural language and receive fast,

AWS has expanded AWS Transform to support automated migration of Tableau and Power BI dashboards into Amazon QuickSight by offering partner EZConvertBI agents from Wavicle via AWS Marketplace.

Reinforcement Fine — Tuning (RFT) significantly enhances the alignment of large language models (LLMs) by addressing issues like inaccuracies and misalignments.

AWS has published guidance on configuring Amazon Bedrock AgentCore Gateway to securely access private resources within an Amazon Virtual Private Cloud (VPC), mitigating operational overhead and enhancing security for AI agent deployments.

AWS has demonstrated a new architecture utilizing agentic AI assistants from Amazon Quick to democratize data analytics, transforming it into a self-service capability for users of all technical levels.

Amazon Web Services (AWS) has launched the Generative AI Model Agility Solution, a framework to simplify the migration and upgrading of Large Language Models (LLMs) in production.

Amazon has recently launched 'Join the chat,' an AI-powered audio Q&A feature integrated into its product pages, allowing customers to ask questions and receive real-time, conversational spoken responses.

Amazon Web Services (AWS) achieved its fastest growth in 15 quarters, with net sales surging 28% to $37.6 billion, primarily driven by robust demand from the AI sector.

PwC has launched its AI-driven annotation (AIDA) solution, developed on AWS, to transform unstructured contracts into searchable insights.