Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Amazon Bedrock

AI overview

Sign in and the AI will write an overview from our coverage.

Headlines · 115

  1. Standard Bank deploys automated generative AI risk assessment with AWS

    Standard Bank Group, working with AWS Professional Services, deployed an automated assessment that scores generative AI use cases against international standards and internal policies. AWS Machine Learning Blog reports that the system uses an unmodified foundation model on Amazon Bedrock as an LLM-as-a-judge to collect evidence, identify mitigations, produce reports and a dashboard, and support existing risk-committee approval gates.

    AWS Machine Learning Blog · 🔥 4
  2. Kimi Code CLI v2.7.0 Adds Native Amazon Bedrock Provider

    AWS Machine Learning Blog reports that Kimi Code CLI v2.7.0 adds a native Amazon Bedrock provider that uses the standard AWS credential chain, including named profiles and SSO. The integration can connect Kimi K3.1 Pro Highspeed on Amazon Bedrock without separately managing a Bedrock API key, although suitable AWS credentials, IAM permissions, region, and model access are still required.

    AWS Machine Learning Blog · 🔥 0
  3. Agent Anomaly Detection expands beyond Google Cloud

    Google says Agent Anomaly Detection has entered Private Preview with support for monitoring agents outside Google Cloud when their traces and logs are sent to a Google Cloud project. Eligible agents can include ADK for Python applications running on AWS Lambda, Microsoft Foundry, or on-premises infrastructure. In vendor-provided tests, the service flagged catalog scraping, resource exhaustion, and unauthorized refunds.

    Google Developers Blog · 🔥 0
  4. Amazon Bedrock AgentCore Adds Strands Defaults and Lifecycle Hooks

    AWS is upgrading the Amazon Bedrock AgentCore harness with Strands harness defaults, built-in tools for planning, web retrieval, filesystem access, context offloading, and task delegation. It also adds lifecycle hooks that let AWS Lambda functions allow or deny actions and send notifications through Amazon SNS or EventBridge. The Strands-backed experience began rolling out to new and existing harnesses on November 11, with no new resource type or customer action required.

    AWS Machine Learning Blog · 🔥 0
  5. Accenture and AWS Detail Six Ways to Reduce AI Token Costs

    The AWS Machine Learning Blog describes a six-lever approach Accenture used to reduce token spending in four client engagements spanning telecom, insurance, mining and a large enterprise deployment. The methods cover cost attribution, workload-aware model routing, context reduction, semantic caching, model distillation and infrastructure optimization, and are available through Amazon Bedrock. Where savings were measured, Accenture reported token-spending reductions of about 40–85%, though client names were withheld and the results may vary by workload.

    AWS Machine Learning Blog · 🔥 0
  6. Claude Code v2.1.328 adds auto mode and hook safeguards

    Claude Code v2.1.328 adds auto mode for Claude Opus 4.6 and Sonnet 4.6 across Amazon Bedrock, Google Cloud’s Agent Platform, Microsoft Foundry, and the Claude apps gateway. The release also lets failed MCP tool hooks block their associated action, warns about incomplete OpenTelemetry configuration, supports additional policy controls, and improves self-hosted runners, alongside numerous bug fixes.

    Claude Code Releases · 🔥 0
  7. Airbnb runs coding agents on Amazon Bedrock with cost-per-task evals

    AWS Machine Learning Blog reports that thousands of engineers and engineering-adjacent staff at Airbnb use the internal AirChat tool to run coding agents each week. The agents send requests through Airbnb's LLM gateway on Amazon Bedrock, where the company evaluates models using its own tasks and cost per task before rollout, then monitors production latency, throttling, token consumption, and spend. Airbnb also uses evaluation results to identify equivalent models and rebalance heavy users when costs rise.

    AWS Machine Learning Blog · 🔥 0
  8. AWS Details an Explainable RAG Design for Aurora PostgreSQL

    AWS published a design pattern for implementing the Explainability dimension of Responsible AI using a RAG-based customer support chatbot built on Amazon Aurora PostgreSQL-Compatible Edition. The walkthrough uses Aurora ML, pgvector, Amazon Bedrock, Bedrock Guardrails, citations, and database audit logging to record retrieval provenance, including retrieved and cited sources. AWS says Aurora ML can invoke SageMaker endpoints and Bedrock foundation models directly from SQL. The post is Part 1 and implements one of AWS's eight Responsible AI dimensions in depth, rather than all eight.

    AWS Machine Learning Blog · 🔥 0
  9. AWS details Social Department's QuickSearch video discovery system

    AWS Machine Learning Blog says Social Department built QuickSearch, a production system that retrieves entertainment footage for shareable scenes, quotable quotes, memeable moments, and trend alignment. Its AWS architecture combines TwelveLabs Marengo 3.0 and Pegasus 1.2 models on Amazon Bedrock, Claude models for interpretation and editorial ranking, and hybrid retrieval through Amazon OpenSearch Serverless. The system uses source-grounding checks, Amazon Bedrock Guardrails, and human approval before publishing.

    AWS Machine Learning Blog · 🔥 0
  10. Stretto deploys Tara AI voice agent for bankruptcy intake

    AWS reported that Stretto, the AWS Generative AI Innovation Center and CI&T built Tara, a speech-to-speech AI voice agent powered by Amazon Nova 2 Sonic, and deployed it for phone-based bankruptcy intake. Tara guides callers through structured interviews, collects personal and financial details, and applies attorney-defined rules to qualify leads for attorney review. After each call, a Claude 3 Haiku pipeline extracts fields from the transcript and stores them in Amazon Aurora. Stretto reports approximately 96% structured-field extraction accuracy in testing, an average completed-call duration of about eight minutes, a 20% lead-conversion increase compared with the web form, and a 95% reduction in per-minute voice cost versus its previous solution.

    AWS Machine Learning Blog · 🔥 0
  11. Disney and AWS unveil AI-powered Star Wars Anzellans

    Walt Disney Imagineering and Amazon Web Services have unveiled three AI-powered Anzellan Audio-Animatronic prototypes: Vilo Vele, Kado Bre'a and Jibu Bre'a. Using cameras, microphones, sensors and AWS cloud models, the characters perceive nearby guests and choose from dialogue and movements prewritten and recorded by Disney Imagineers, rather than generating unrestricted responses. The prototypes are scheduled to debut at AWS re:Invent in Las Vegas from November 30 to December 4, but Disney has not announced a park deployment.

    CNET · 🔥 0
  12. AWS urges teams to question every generative AI model call

    AWS Machine Learning Blog offers a decision framework for choosing between deterministic code, constrained decision calls, and appropriately sized Amazon Bedrock models at each workflow step. It says teams often send validation, formatting, and routing tasks to LLMs, adding costs, latency, and output variability at production scale. The guidance also recommends deterministic fast paths, labeled-data evaluation, configuration flags, and continuous integration for model changes.

    AWS Machine Learning Blog · 🔥 0
  13. AWS FDE expands AI deployment service to 15 countries

    AWS says Forward Deployed Engineering is now available to commercial and public-sector customers in 15 countries. The service uses 45-day sprints to build AI systems in customers’ own AWS accounts, with each sprint priced under the FDE Terms and billed through the account. AWS describes the offering as combining production deployment, reusable foundations, workflows and governance.

    AWS Machine Learning Blog · 🔥 0
  14. LoanPro deploys multi-agent loan-servicing assistant with AWS

    LoanPro built a production multi-agent loan-servicing assistant with AWS Partner AllCloud and the AWS Generative AI Innovation Center. The system uses Amazon Bedrock and Amazon Bedrock AgentCore to retrieve live borrower data, generate account summaries and recommend servicing actions. Its Data Retrieval Tool spans eight LoanPro API endpoints and synthesizes about 1,400 fields per account. AgentCore Gateway exposes the APIs as tools over the Model Context Protocol, handles authentication and manages connections, while AWS Lambda provides the integration layer. LoanPro said the deployment applies role-based access controls and compliance guardrails and records AI actions in its audit trail. AllCloud’s published case study reported customer call times reduced by up to 15%, but the supplied report provides no independent measurement methodology.

    AWS Machine Learning Blog · 🔥 0
  15. AWS Outlines a Phased Responsible-AI Roadmap for AI Agents

    AWS published a roadmap for platform and application teams building agentic AI systems, organized around governance foundations, pilot deployment, and evaluation and hardening. It includes a Responsible AI Assessment Skill, a maturity scorecard, governance controls, human approval mechanisms, observability, and evaluation practices spanning eight responsible-AI pillars. AWS says the framework is a starting point rather than a compliance checklist and advises organizations to tailor it with their legal, compliance, and ethics teams.

    AWS Machine Learning Blog · 🔥 0
  16. ArisGlobal Cuts NavaX Inference Costs with Bedrock Prompt Caching

    AWS reports that life-sciences software company ArisGlobal rebuilt NavaX’s Amazon Bedrock prompt structure to cache its shared prefix rather than each case document, enabling reuse across requests and tenants. ArisGlobal says the change reduced effective input costs by approximately 50%, contributed to about 40% overall monthly inference savings, and supported roughly 15% more enterprise tenants. AWS cautions that these vendor-reported production results vary with prompt structure, traffic patterns, and cache-hit rates.

    AWS Machine Learning Blog · 🔥 0
  17. WAIC CONNECT HK plans three-stop Hong Kong AI City Walk

    WAIC CONNECT HK announced a limited-capacity AI City Walk in Hong Kong on December 9, with planned stops at Amazon Web Services, Hong Kong University of Science and Technology and Hong Kong Science Park. The 50–70-person tour is scheduled to assemble at 9:30 and end at approximately 17:45, focusing on enterprise AI applications, local generative AI research and commercialization support. Exact visit points and the route remain subject to organizer confirmation, while final sharing and demonstration arrangements will depend on the hosts. The organizers also announced that 2027 WAIC HONG KONG will take place at Hong Kong Science Park on December 10–11, 2026.

    36氪 人工智能 · 🔥 0
  18. Duco rebuilds financial-data agent stack on Amazon Bedrock AgentCore

    AWS reports that Duco rebuilt its agentic AI stack for financial-data automation using Amazon Bedrock and Amazon Bedrock AgentCore. The architecture combines strict per-tenant isolation, a multi-model strategy and on-demand MCP skills loading to reduce context overhead. AWS says reconciliation deployments are 10 times faster; for one global bank, onboarding costs fell by 70% and straight-through processing increased by more than 20%. The report attributes these results to Duco's client deployments.

    AWS Machine Learning Blog · 🔥 0
  19. Multi-agent coordination patterns for manufacturing

    A tutorial organizes multi-agent coordination into two tiers: Tier 1 uses predefined flows, including sequential pipelines, conditional workflows and routers, while Tier 2 lets a model choose the path at runtime. It recommends beginning with deterministic Tier 1 workflows and reserving Tier 2 for tasks whose subtasks emerge during execution. The tutorial also identifies anti-patterns such as unbounded iteration, unstructured natural-language handoffs, premature adoption of multiple agents and unapproved real-world actions. Suggested safeguards include hard iteration and token limits, typed handoffs, observability, business rules and human approval before execution.

    AWS Machine Learning Blog · 🔥 0
  20. MoEngage Deploys Merlin AI on Amazon Bedrock AgentCore

    AWS Machine Learning Blog described how MoEngage built Merlin AI, a set of AI agents that operate on live platform data, using Amazon Bedrock, Amazon Bedrock AgentCore, and the Model Context Protocol. Its MCP Connector exposes existing OpenAPI-compatible campaign, analytics, segmentation, content, and flow capabilities as tools, with JWT validation, isolated sessions, and tenant-scoped access. MoEngage reports customer improvements including faster campaign launches and higher push click-through rates, but the figures are vendor-reported.

    AWS Machine Learning Blog · 🔥 0
  21. AWS Details AI-Assisted Assembler-to-Java Modernization Workflow

    AWS Machine Learning Blog presents a five-phase workflow for modernizing legacy mainframe Assembler applications into Java Spring Boot services on AWS. The process combines Amazon Bedrock and Kiro IDE for business-rule extraction with an intermediate COBOL stage and AWS Transform for COBOL-to-Java conversion, while requiring subject matter expert review and incremental validation. AWS says the approach can reduce modernization time and preserve business logic, but those benefits are vendor claims rather than independently reported results.

    AWS Machine Learning Blog · 🔥 0
  22. Amazon Bedrock adds agentic tabular search

    AWS says Amazon Bedrock Managed Knowledge Base now supports tabular search across CSV, XLSX, and XLS files through agentic retrieval. The feature can calculate across complete tables, combine results with facts from related documents, and run code in an isolated sandbox over original files. It uses the existing AgenticRetrieveStream request, while AWS says trace data identifies which files the code read.

    AWS Machine Learning Blog · 🔥 0
  23. Amazon Bedrock Adds Explicit Prompt Caching for Open-Weight Models

    AWS announced that open-weight models on Amazon Bedrock now support explicit prompt caching, alongside model-dependent implicit caching. Prompt caching reuses computed state for matching prompt prefixes, while explicit mode lets developers mark reusable content with cache breakpoints through the Chat Completions API. AWS says this can reduce input costs and response latency, but the reported benefits remain vendor claims.

    AWS Machine Learning Blog · 🔥 0
  24. Route coding agents across Bedrock models with NeMo Switchyard

    A guide shows how to deploy LiteLLM with the NVIDIA NeMo Switchyard routing plugin on AWS, allowing coding-agent requests to be routed among models on Amazon Bedrock. Stage routing uses recent tool activity and request-history results to select one of two candidates, while Random routing supports weighted traffic splits and A/B comparisons. Sample benchmark runners compare the cost per solved task for a routed model group with single-model baselines. The Switchyard LiteLLM integration is experimental, and the sample code is intended for non-production use in a non-production account.

    AWS Machine Learning Blog · 🔥 0
  25. AWS Shows Per-User Cost Attribution for Bedrock AI Tools

    AWS Machine Learning Blog explains how Amazon Bedrock traces inference calls to IAM principals and carries organizational tags into AWS Cost and Usage Reports and Cost Explorer, without requiring API changes or tool-specific instrumentation. The guide covers identity setup through AWS IAM Identity Center, reporting with Amazon Athena, and usage controls ranging from budgets to CloudWatch alarms, DynamoDB-based quotas, and gateway enforcement. AWS says the same endpoint-level approach applies to Claude Code, Claude Desktop, OpenAI Codex, ChatGPT Work, and OpenCode.

    AWS Machine Learning Blog · 🔥 0
  26. Four Amazon Bedrock request settings that can raise inference costs

    AWS Machine Learning Blog identifies four Amazon Bedrock request-level choices that can change inference costs without altering the model or prompt: prompt-cache duration, service tier, reasoning effort, and maxTokens. The article explains how each option affects processed tokens or billing rates, shares test results including higher output-token use at greater reasoning effort, and recommends checking response metadata, logs, and billing data. AWS also provides Bedrock sample notebooks and workshop code, while noting that effects vary by model, workload, AWS Region, and current pricing.

    AWS Machine Learning Blog · 🔥 0
  27. AWS Guide Adds Per-Identity Bedrock Spend Tracking and Limits

    AWS Machine Learning Blog describes an AWS-native method for attributing Amazon Bedrock costs to IAM principals, reporting usage, and enforcing spend thresholds. The approach combines Bedrock’s automatic caller identity records with Cost Explorer or CUR 2.0 for reporting and AWS Budgets actions that can apply a Deny IAM policy when a threshold is crossed. AWS says the common case requires no application code, while shared roles and gateway roles can limit attribution granularity.

    AWS Machine Learning Blog · 🔥 0
  28. Anthropic launches self-hosted Claude Managed Agents beta

    Anthropic announced self-hosted Claude Managed Agents in private beta, letting customers run the agent harness, sandboxes, credentials, session history, database, and workers in their own cloud or on-premise environment. Model requests can be routed to Anthropic, Amazon Bedrock, Google Agent Platform, Microsoft Foundry, or a customer gateway, while security features include host allowlists, credential injection outside sandboxes, and OpenTelemetry activity records. The Next Web notes that model-call content still reaches the selected provider, customers assume server operation, and features such as webhooks and scheduled runs are unavailable.

    Claude Blog · 🔥 0
  29. A practical framework for agent cost and ROI on Amazon Bedrock

    AWS Machine Learning Blog outlines a framework for measuring the true cost of an AI agent workflow per completed task rather than per token. It identifies tokenizer efficiency, output alignment, agentic loop efficiency and reasoning overhead as hidden cost drivers, then recommends caching, model routing, output constraints and batch inference to reduce expenses. To assess ROI, the article proposes documenting a human baseline, running controlled and holdout experiments, and reviewing a five-dimension scorecard tied to operational decisions.

    AWS Machine Learning Blog · 🔥 0
  30. AWS details a self-service LLM gateway for Amazon Bedrock

    AWS Machine Learning Blog presents a reference pattern for running LiteLLM on AWS Fargate as a self-service gateway in front of Amazon Bedrock. The OpenAI-compatible API gives developers dedicated keys and access to supported foundation models while adding per-user spend controls, CloudWatch monitoring and optional Langfuse tracing. The tutorial uses Claude Sonnet and Claude Haiku for its examples and includes AWS CDK constructs for application inference profiles, IAM and alarms.

    AWS Machine Learning Blog · 🔥 0

Next →

Experience and discussion from the community

Share my Amazon Bedrock experienceAsk about Amazon Bedrock

Nobody has shared their experience with Amazon Bedrock yet.