Amazon Bedrock is a fully managed service that provides secure, enterprise-grade access to high-performing foundation models from leading AI companies, enabling you to build and scale generative AI applications. Today, Amazon Bedrock introduces a console experience designed for how customers actually build with foundation models: experiment, iterate, and scale. This is the same Amazon Bedrock service customers already use, with a refreshed workflow optimized for the bedrock-mantle endpoint, which supports the OpenAI Responses API, OpenAI Chat Completions API, and the Anthropic Messages API.
The new experience makes it simple to find the right model and move quickly from evaluation to production. Customers can browse the full Amazon Bedrock model catalog, including the latest Claude, GPT, and open-weight models, and compare them side by side on capabilities, modality support, context window, and applicable service quotas in a single view, removing the need to stitch together documentation, and limit calculators. Work is organized into projects, where customers can run evaluations and review usage insights in one streamlined workflow that mirrors the lifecycle of building a generative AI application. Each project also includes project-aware documentation: code samples, SDK snippets, and API references are automatically prefilled with the project’s selected model ID, region, bedrock-mantle endpoint URL, and API key reference, and they update in place as customers change models or settings. Developers can copy a snippet straight from the console into their application and run it without modification.
To get started, sign in to the AWS Management Console, open Amazon Bedrock, and choose the new experience from the navigation. Create a project, pick a model, and begin sending requests through the bedrock-mantle endpoint using your existing OpenAI or Anthropic client libraries with an Amazon Bedrock API key. The new console experience is available in all AWS Regions where the bedrock-mantle endpoint is offered: US East (N. Virginia, Ohio), US West (Oregon), Asia Pacific (Jakarta, Mumbai, Sydney, Tokyo), Europe (Frankfurt, Ireland, London, Milan, Stockholm), and South America (São Paulo). To try the new experience, visit the Amazon Bedrock console.
Amazon Bedrock is a fully managed service that provides secure, enterprise-grade access to high-performing foundation models from leading AI companies, enabling you to build and scale generative AI applications. Today, Amazon Bedrock introduces a console experience designed for how customers actually build with foundation models: experiment, iterate, and scale. This is the same Amazon Bedrock service customers already use, with a refreshed workflow optimized for the bedrock-mantle endpoint, which supports the OpenAI Responses API, OpenAI Chat Completions API, and the Anthropic Messages API. The new experience makes it simple to find the right model and move quickly from evaluation to production. Customers can browse the full Amazon Bedrock model catalog, including the latest Claude, GPT, and open-weight models, and compare them side by side on capabilities, modality support, context window, and applicable service quotas in a single view, removing the need to stitch together documentation, and limit calculators. Work is organized into projects, where customers can run evaluations and review usage insights in one streamlined workflow that mirrors the lifecycle of building a generative AI application. Each project also includes project-aware documentation: code samples, SDK snippets, and API references are automatically prefilled with the project’s selected model ID, region, bedrock-mantle endpoint URL, and API key reference, and they update in place as customers change models or settings. Developers can copy a snippet straight from the console into their application and run it without modification. To get started, sign in to the AWS Management Console, open Amazon Bedrock, and choose the new experience from the navigation. Create a project, pick a model, and begin sending requests through the bedrock-mantle endpoint using your existing OpenAI or Anthropic client libraries with an Amazon Bedrock API key. The new console experience is available in all AWS Regions where the bedrock-mantle endpoint is offered: US East (N. Virginia, Ohio), US West (Oregon), Asia Pacific (Jakarta, Mumbai, Sydney, Tokyo), Europe (Frankfurt, Ireland, London, Milan, Stockholm), and South America (São Paulo). To try the new experience, visit the Amazon Bedrock console.
You can now deploy Amazon MQ for RabbitMQ in the AWS European Sovereign Cloud (Germany) Region. This new independent cloud for Europe is located entirely within the EU, designed to help customers in regulated industries and public sector organizations meet their sovereignty requirements.
Amazon MQ is a managed message broker service that makes it easy to set up and operate message brokers in the cloud. Amazon MQ for RabbitMQ manages the provisioning, patching, and maintenance of RabbitMQ brokers, letting you focus on building applications without managing messaging infrastructure. You can migrate existing RabbitMQ workloads without rewriting application code and benefit from the same familiar APIs and protocols. Amazon MQ for RabbitMQ in the AWS European Sovereign Cloud supports RabbitMQ engine version 4.2 and Graviton3-based m7g instance types for high-performance messaging ranging from m7g.medium to m7g.16xlarge.
You can now deploy Amazon MQ for RabbitMQ in the AWS European Sovereign Cloud (Germany) Region. This new independent cloud for Europe is located entirely within the EU, designed to help customers in regulated industries and public sector organizations meet their sovereignty requirements. Amazon MQ is a managed message broker service that makes it easy to set up and operate message brokers in the cloud. Amazon MQ for RabbitMQ manages the provisioning, patching, and maintenance of RabbitMQ brokers, letting you focus on building applications without managing messaging infrastructure. You can migrate existing RabbitMQ workloads without rewriting application code and benefit from the same familiar APIs and protocols. Amazon MQ for RabbitMQ in the AWS European Sovereign Cloud supports RabbitMQ engine version 4.2 and Graviton3-based m7g instance types for high-performance messaging ranging from m7g.medium to m7g.16xlarge. To get started, see the Amazon MQ product page or the Amazon MQ Developer Guide.
Amazon Cognito now supports multi-Region replication, enabling you to synchronize user and machine identity data — including credentials, user pool configurations, and federation setups — to a secondary user pool in a standby Region you designate in near real-time. This capability helps you improve the resilience of your authentication system by providing a standby replica that can accept traffic in case there is a regional service disruption. In the event of a disruption in the primary Region, you can redirect traffic to the secondary user pool. Signed-in users continue accessing their applications without re-authenticating, and registered users can sign in with their existing credentials. Authentication methods continue to work in the secondary Region, including username/password, federation with social identity and SAML/OIDC providers, and machine-to-machine authorization flows. Multi-Region replication is available as an add-on for user pools in Essentials or Plus feature tiers. You can start using this feature in the following AWS Regions: US East (Ohio, N. Virginia), US West (N. California, Oregon), Asia Pacific (Mumbai, Seoul, Singapore, Sydney, Tokyo), Canada (Central), Europe (Frankfurt, Ireland, London, Paris, Stockholm), and South America (São Paulo). To get started, configure multi-Region replication using the AWS Management Console, AWS Command Line Interface (CLI), or AWS Software Development Kits (SDKs) by adding a replica user pool. Visit the pricing page for pricing details and the developer guide for instructions.
Amazon Cognito now supports multi-Region replication, enabling you to synchronize user and machine identity data — including credentials, user pool configurations, and federation setups — to a secondary user pool in a standby Region you designate in near real-time. This capability helps you improve the resilience of your authentication system by providing a standby replica that can accept traffic in case there is a regional service disruption. In the event of a disruption in the primary Region, you can redirect traffic to the secondary user pool. Signed-in users continue accessing their applications without re-authenticating, and registered users can sign in with their existing credentials. Authentication methods continue to work in the secondary Region, including username/password, federation with social identity and SAML/OIDC providers, and machine-to-machine authorization flows. Multi-Region replication is available as an add-on for user pools in Essentials or Plus feature tiers. You can start using this feature in the following AWS Regions: US East (Ohio, N. Virginia), US West (N. California, Oregon), Asia Pacific (Mumbai, Seoul, Singapore, Sydney, Tokyo), Canada (Central), Europe (Frankfurt, Ireland, London, Paris, Stockholm), and South America (São Paulo). To get started, configure multi-Region replication using the AWS Management Console, AWS Command Line Interface (CLI), or AWS Software Development Kits (SDKs) by adding a replica user pool. Visit the pricing page for pricing details and the developer guide for instructions.
A principios de este año, las demos de OpenClaw parecían estar por todas partes, aunque muchas se resumían en el mejor de los casos a un truco llamativo de fiesta («Mira, mi agente pidió una pizza»). Pero hizo que Omar Shahine, un empleado de Microsoft desde hace mucho tiempo, pensara: ¿Qué utilidad podrían tener en verdad las claws (módulos de extensión funcional)?
Resultó que mucha. En su tiempo libre, Shahine creó Lobster, un asistente personal de IA basado en OpenClaw. Tiene su propio Apple ID y dirección de correo electrónico, así que puede enviar mensajes de texto desde cualquier dispositivo con iMessage. En un inicio dividió a Lobster en un trío de agentes, cada uno con su propio perfil de seguridad y acceso a herramientas (ocho semanas después, ese número había aumentado a nueve agentes siempre activos). Lobster se encarga de la logística de viaje, envía recordatorios familiares de manera proactiva con antelación y, en general, ayuda a Shahine y a su familia a mantenerse organizados y a hacer las cosas. Y tras presentar Lobster al grupo de AI Accelerator de Microsoft, Shahine consiguió un nuevo trabajo: llevar OpenClaw a M365 y a la nube como CVP de lo que se denominó «Project Lobster».
Al mismo tiempo, Jakob Werner, miembro del personal técnico de Microsoft, seguía una idea similar, pero con un giro: un agente basado en una aplicación de escritorio inspirado en OpenClaw. El objetivo era ofrecer un asistente personal de IA potente y seguro para la empresa, que cualquiera dentro de Microsoft pudiera utilizar. En solo un par de semanas, lo que a nivel interno se denominaba «Clawpilot» ya había sido descargado por miles de empleados de Microsoft, y esa comunidad sigue en crecimiento.
Cuando Shahine empezó a reunir un pequeño equipo de constructores entusiastas—Ocean’s 11, claro está—Werner se unió de inmediato a sus filas. De manera reciente, ambos se reunieron en Redmond, Washington, para comparar ideas sobre cómo construir estos agentes siempre activos y autónomos y cómo navegar por los mundos de la seguridad empresarial, la memoria agéntica y más.
Abrazar el espíritu del código abierto
El equipo de Project Lobster representa una nueva manera de trabajar dentro de Microsoft, impulsada por los avances de la IA. Es un grupo muy unido que prefiere colaborar de forma asíncrona. Hay un consenso general en contra de las reuniones. Todos contribuyen a la base de código, incluido Shahine. Y no hay un asistente ejecutivo tradicional entre sus filas: cada miembro del equipo utiliza prototipos a lo largo del día para sumergirse por completo en la tecnología mientras la construyen. Incluso hay una comunidad de código abierto en crecimiento alrededor del equipo que refleja lo que se encuentra en proyectos de código abierto fuera de los muros de Microsoft.
«Nunca he visto un proyecto dentro de la empresa donde tanta gente apareciera con sus ideas y su código y trabajara para producir una PR «, dice Shahine.
Nunca he visto un proyecto dentro de la empresa donde tanta gente haya llegado con sus ideas.
De hecho, la expectación interna en torno al Project Lobster ha sido tal que el equipo ha recibido solicitudes de propuesta de cambios (PR – Pull Request, forma de proponer cambios en el código dentro de un proyecto) por todas partes durante la fase inicial de construcción, que revisaron para determinar si cumplían el estándar necesario para incluirlo en el producto. Incluso algunos de los cambios de Shahine no pasaron a la lista. El foco debía mantenerse en el objetivo central del producto: crear un agente personal siempre activo para el trabajo. Un ayudante de IA que aprende tus objetivos, se adapta a tus patrones de trabajo diarios y actúa con contexto, para identificar problemas antes de que surjan, mantener los proyectos en marcha e impulsar resultados sin una intervención constante. Un agente que pueda detectar cuando un calendario se ha saturado y proponer cambios específicos antes de que empiece la semana o identificar cuándo una decisión está estancada y redactar un seguimiento específico para desbloquearlo.
«Tenemos que determinar si una relación pública dada cambia o no la idea central del producto —y la velocidad de esa reseña es la velocidad humana, no la de la IA», señala Werner. «Cualquiera puede hacer una PR súper rápido ahora. Lo que intentamos es ayudar a la comunidad y enseñar a los colaboradores cómo revisar las PR.»
Aunque el trabajo comenzó como un experimento interno, pronto se convirtió en un esfuerzo centrado en el cliente que culminó con la introducción de Microsoft Scout, un agente personal siempre activo impulsado por la tecnología open-source OpenClaw.
De experimento a producto listo para empresas
Microsoft Scout opera de manera autónoma—con su propia identidad—y actúa en su nombre. Funciona en la nube, el escritorio y el navegador web, por lo que puede conectarse entre las superficies que usan—Teams, Outlook, OneDrive y SharePoint—y los sistemas donde circula el trabajo, incluido el correo electrónico, calendario y contactos.
A diferencia de cualquier claw común en la naturaleza, Microsoft Scout combina el código OpenClaw con la identidad empresarial, la gobernanza y la seguridad. Cada paquete se ingiere a través de una cadena de suministro de Microsoft seleccionada y firmada, y cada llamada a herramienta, solicitud de modelo y salto de red está mediado por un tiempo de ejecución de confianza cero: el contenedor del agente se trata como no confiable, con la identidad, los tokens y la política controlados por Microsoft fuera de él. Con Agent 365, los administradores disponen de un único plano de control, y Microsoft Purview ofrece a los equipos de seguridad la misma señal de cumplimiento y DLP que ya reciben de otras superficies M365.
«Es una herramienta súper poderosa», reconoce Werner. «Y para ser seguros en la empresa, necesitábamos asegurarnos de que la gobernanza de los datos fuera correcta, que la privacidad fuera correcta y que no cancelara una reunión y enviara toda tu información personal a esa cadena de correos electrónicos. Si envío a mi agente contigo, no debería contarte todo sobre mí. Estas áreas son posibles de contener, pero también tuvimos que hacerlo de manera equilibrada, que no limite las posibilidades a nada.»
Es un intercambio que merece la pena hacer. Y con las pruebas y confiables ofertas de seguridad empresarial de Microsoft y la investigación e innovación continuas en el sector, el equipo contó con una base sólida desde la cual afrontar el desafío.
El papel de la memoria agéntica
Para que un agente personal de IA siempre activo sea en verdad útil, debe ser proactivo—y eso requiere contexto impulsado por Work IQ. Con el tiempo, Microsoft Scout entiende cómo trabajan, utiliza las mismas herramientas de productividad que ustedes y les quita cosas de encima sin necesidad de constantes indicaciones. Aprende sus objetivos, se adapta a sus patrones de trabajo diarios y actúa con intención. A diferencia de las olas tecnológicas anteriores, este es un software en verdad personalizado. Eso es transformador, pero no está exento de compensaciones.
«OpenClaw, Claude Code, GitHub Copilot CLI, son arneses de codificación agéntica que, de manera básica, recuerdan—escriben cosas igual que hacen las personas», señala Shahine. «Escriben las cosas como un diario. Pero así como necesita recordar cosas, también necesita olvidar algunas cosas.»
Así como necesita recordar cosas, también necesita olvidar algunas cosas.
Como ejemplo, Shahine señala la introducción de la memoria en ChatGPT. Pasó un tiempo para contarle a ChatGPT que su hija tenía 17 años mientras que su hijo 13. Pero un año después, esa información seguía estática. El sistema no tenía el concepto de que algunos hechos debían cambiar con el tiempo, mientras que otras piezas de información —como tu nombre— permanecerían igual.
«En la fase de diseño, pensaba en el humano y en cómo los humanos memorizan las cosas», dice Werner. «Olvido cosas que son irrelevantes porque no las usé. Así que he creado un sistema en el que, si voy a usarlo de manera repetida, se va a quedar. Pero si no voy a usarlo con regularidad, quiero que el sistema se olvide. No quiero tener un diario infinito de cosas, ¿verdad? Así que hay una especie de capas de memoria, y con el tiempo desaparece si no se usa. Mientras tanto, la relevancia de otros recuerdos crece a medida que los usas más.»
Formación de un nuevo centro de gravedad
Cuando unieron fuerzas por primera vez, Werner le presentó a Shahine el concepto de gravedad, el marco en torno al cual operaba.
«Para construir un producto en verdad bueno, no creo que pueda hacerlo yo mismo», explica Werner. «Tenemos que colaborar con otras personas. Pero, ¿cómo influimos en otras personas para que colaboren con nosotros? Y la mentalidad que uso e intento inculcar en mi equipo es la gravedad. Construimos algo y lo hacemos tan grande en influencia—no en el número de características, sino en su influencia—que cuando surgen ideas nuevas y emocionantes, quieren intentar unirse a la gravedad de nuestro trabajo en lugar de disolver el enfoque.»
«Y no sabía en realidad de qué hablabas hasta que anunciaron mi nuevo papel», admite Shahine. «Pero desde entonces, he recibido cientos, si no miles, de mensajes de personas que quieren ayudar, personas que quieren aprender, personas que quieren mostrarme lo que han hecho y clientes que quieren saber cuanto antes cuándo van a poner las manos en lo que construimos. Hay muchas otras palabras para eso—atracción del usuario, señal—pero tu mantra de gravedad en verdad resuena conmigo ahora.»
Los empleados de Microsoft ya han comenzado a utilizar una experiencia temprana de escritorio de Microsoft Scout. Lo construimos para aprender cómo los agentes siempre activos aparecen en trabajos reales, y vemos cómo asume coordinación, afronta riesgos antes y mantiene el trabajo en movimiento sin necesidad de que se lo pidan de manera constante.
Ahora extendemos esa experiencia inicial a las organizaciones Frontier. Microsoft Scout está disponible como una versión experimental a través de Frontier, para dar a los clientes la oportunidad de explorar cómo puede encajar en sus propios flujos de trabajo.
El acceso requiere inscripción en Frontier, configuración de políticas de Intune y una certificación opt-in. Los usuarios con una licencia de GitHub Copilot pueden descargar e instalar la experiencia. Más información.
Den clic aquí para conocer más sobre el nuevo blog de Microsoft de constructuroes, para constructores.
AWS IoT Device Management adds MQTT session data to connectivity status API, enabling you to troubleshoot connectivity issues and audit connection patterns across your Internet of things (IoT) device fleet.
This launch brings AWS IoT Device Management’s existing connectivity status API to full parity with AWS IoT Core’s recently launched GetConnection API, enabling you to retrieve detailed connection and MQTT session information for the IoT device by its thing name. In addition to the connection status, timestamp, and disconnect reason already available, you now get visibility into MQTT session timeout and session expiry values, along with optional socket level details such as source and destination IP addresses, ports, and client VPC endpoint ID. Access to socket information is controlled through granular IAM policies, so you can restrict it to the teams that need it.
A key advantage of the connectivity status API over AWS IoT Core’s GetConnection API is data retention. While GetConnection retains connection and session details for 30 minutes after a device disconnects, the connectivity status API stores this information indefinitely. This means you can investigate disconnect reasons, review session metadata, and troubleshoot issues long after a device goes offline.
This enhancement is available in all AWS regions where AWS IoT Device Management is supported. AWS IoT Device Management only supports devices registered in AWS IoT Core Thing Registry. To learn more, visit the AWS IoT Device Management documentation and reference guide.
AWS IoT Device Management adds MQTT session data to connectivity status API, enabling you to troubleshoot connectivity issues and audit connection patterns across your Internet of things (IoT) device fleet. This launch brings AWS IoT Device Management’s existing connectivity status API to full parity with AWS IoT Core’s recently launched GetConnection API, enabling you to retrieve detailed connection and MQTT session information for the IoT device by its thing name. In addition to the connection status, timestamp, and disconnect reason already available, you now get visibility into MQTT session timeout and session expiry values, along with optional socket level details such as source and destination IP addresses, ports, and client VPC endpoint ID. Access to socket information is controlled through granular IAM policies, so you can restrict it to the teams that need it. A key advantage of the connectivity status API over AWS IoT Core’s GetConnection API is data retention. While GetConnection retains connection and session details for 30 minutes after a device disconnects, the connectivity status API stores this information indefinitely. This means you can investigate disconnect reasons, review session metadata, and troubleshoot issues long after a device goes offline. This enhancement is available in all AWS regions where AWS IoT Device Management is supported. AWS IoT Device Management only supports devices registered in AWS IoT Core Thing Registry. To learn more, visit the AWS IoT Device Management documentation and reference guide.
Amazon SageMaker Data Agent, available in SageMaker Unified Studio now supports conversation history, enabling data practitioners to maintain continuity across analytical sessions. Data analysts and data scientists can now seamlessly reference previous agent-generated code, resume multi-step analyses, and review past troubleshooting interactions within their notebooks and Query Editor workflows.
With conversation history, you can pick up exactly where you left off by accessing a scrollable list of past conversations through the clock icon in the chat panel header. Each conversation includes auto-generated titles and timestamps for easy identification. Whether you’re resuming complex multi-step analyses, reusing agent-generated code, or continuing troubleshooting from earlier notebook runs, conversation history keeps the context preserved. Data teams save time, eliminate rework, and move faster across concurrent projects, staying focused on insights rather than rebuilding context.
Amazon SageMaker Data Agent, available in SageMaker Unified Studio now supports conversation history, enabling data practitioners to maintain continuity across analytical sessions. Data analysts and data scientists can now seamlessly reference previous agent-generated code, resume multi-step analyses, and review past troubleshooting interactions within their notebooks and Query Editor workflows.
With conversation history, you can pick up exactly where you left off by accessing a scrollable list of past conversations through the clock icon in the chat panel header. Each conversation includes auto-generated titles and timestamps for easy identification. Whether you’re resuming complex multi-step analyses, reusing agent-generated code, or continuing troubleshooting from earlier notebook runs, conversation history keeps the context preserved. Data teams save time, eliminate rework, and move faster across concurrent projects, staying focused on insights rather than rebuilding context.
Conversation history is available in all AWS Regions where Amazon SageMaker Data Agent is currently available. To learn more about Amazon SageMaker Data Agent and how to leverage conversation history in your analytical workflows, visit the Amazon SageMaker product page or explore the Amazon SageMaker Unified Studio documentation.
Amazon SageMaker Unified Studio now enables you to schedule, parameterize, and orchestrate notebook runs directly from the notebook interface without managing external orchestration infrastructure. This makes it easier for customers to take notebooks from experimentation to production, automating recurring workloads such as daily reports, data quality checks, and model retraining.
You can trigger on-demand background runs on dedicated compute without interrupting interactive sessions and create scheduled or recurring runs. With notebook parameterization, you can reuse a single notebook across different inputs, for example, generating shipping performance reports for multiple carriers, by defining parameters and overriding their values per schedule or on-demand run. You can also orchestrate multi-notebook workflows using the Notebook Operator in the Workflows tool, chaining notebooks so that outputs from one run feed as inputs to the next. When a scheduled or background run fails, AI-assisted troubleshooting using SageMaker Data Agent helps you identify the root cause and suggests fixes directly in the notebook, reducing time to resolution. You can also use the Data Agent to create schedules and start notebook runs using natural language, without having to navigate. To get started, open a notebook in your SageMaker Unified Studio project, choose the menu on the Run all button, and select Run in background. To create a schedule, choose the schedule icon in the notebook header or ask the Data Agent to set one up for you.
Amazon SageMaker Unified Studio now enables you to schedule, parameterize, and orchestrate notebook runs directly from the notebook interface without managing external orchestration infrastructure. This makes it easier for customers to take notebooks from experimentation to production, automating recurring workloads such as daily reports, data quality checks, and model retraining.
You can trigger on-demand background runs on dedicated compute without interrupting interactive sessions and create scheduled or recurring runs. With notebook parameterization, you can reuse a single notebook across different inputs, for example, generating shipping performance reports for multiple carriers, by defining parameters and overriding their values per schedule or on-demand run. You can also orchestrate multi-notebook workflows using the Notebook Operator in the Workflows tool, chaining notebooks so that outputs from one run feed as inputs to the next. When a scheduled or background run fails, AI-assisted troubleshooting using SageMaker Data Agent helps you identify the root cause and suggests fixes directly in the notebook, reducing time to resolution. You can also use the Data Agent to create schedules and start notebook runs using natural language, without having to navigate. To get started, open a notebook in your SageMaker Unified Studio project, choose the menu on the Run all button, and select Run in background. To create a schedule, choose the schedule icon in the notebook header or ask the Data Agent to set one up for you.
You can use notebook scheduling in all AWS Regions where Amazon SageMaker Unified Studio is supported. To learn more, see the AWS blog and user guide.
AWS Step Functions now enables you to add AI agent reasoning steps to your workflow through an optimized integration with the managed harness (currently in preview) in Amazon Bedrock AgentCore. AWS Step Functions is a visual workflow service that orchestrates AWS services with built-in error handling, parallel execution, and human approval steps. The AgentCore harness lets you declare an agent through configuration where you specify the model, tools, and behavior. AgentCore provides the managed environment that runs the agent loop end-to-end.
With this integration, you can automate reasoning tasks in your workflow such as classifying a document or extracting elements from an unstructured form. You can run multiple agents in parallel or in sequence at different decision points in a single workflow and add human approval before critical actions. The workflow execution history shows agent input, output, token usage, and duration with links to agent turn details in Amazon CloudWatch, so you can trace and audit every agent decision. You can reuse an existing harness or create a new one directly from the Workflow Studio, the Step Functions visual builder. With per-invocation overrides such as the model, system prompt, and tools, you can adapt the agent to each workflow context without duplicating configurations. Agent context can be persisted across invocations using a session ID that works within or across workflow executions.
The harness integration is available in the following AWS Regions where the AgentCore harness preview is available: US East (N. Virginia), US West (Oregon), Europe (Frankfurt), and Asia Pacific (Sydney). Standard Step Functions pricing applies for workflow execution with no additional integration charges, and standard Amazon Bedrock and AgentCore pricing applies for model inference and associated AgentCore resources.
To learn more about adding agentic reasoning to your workflows, visit AWS Step Functions documentation.
AWS Step Functions now enables you to add AI agent reasoning steps to your workflow through an optimized integration with the managed harness (currently in preview) in Amazon Bedrock AgentCore. AWS Step Functions is a visual workflow service that orchestrates AWS services with built-in error handling, parallel execution, and human approval steps. The AgentCore harness lets you declare an agent through configuration where you specify the model, tools, and behavior. AgentCore provides the managed environment that runs the agent loop end-to-end.
With this integration, you can automate reasoning tasks in your workflow such as classifying a document or extracting elements from an unstructured form. You can run multiple agents in parallel or in sequence at different decision points in a single workflow and add human approval before critical actions. The workflow execution history shows agent input, output, token usage, and duration with links to agent turn details in Amazon CloudWatch, so you can trace and audit every agent decision. You can reuse an existing harness or create a new one directly from the Workflow Studio, the Step Functions visual builder. With per-invocation overrides such as the model, system prompt, and tools, you can adapt the agent to each workflow context without duplicating configurations. Agent context can be persisted across invocations using a session ID that works within or across workflow executions.
The harness integration is available in the following AWS Regions where the AgentCore harness preview is available: US East (N. Virginia), US West (Oregon), Europe (Frankfurt), and Asia Pacific (Sydney). Standard Step Functions pricing applies for workflow execution with no additional integration charges, and standard Amazon Bedrock and AgentCore pricing applies for model inference and associated AgentCore resources.
To learn more about adding agentic reasoning to your workflows, visit AWS Step Functions documentation.
Amazon Bedrock now supports GPT‑5.4 from OpenAI in AWS GovCloud (US-West) — giving government and regulated industry customers access to OpenAI’s most capable frontier model for professional work, backed by the enterprise-grade security and goverment compliance scope of AWS GovCloud (US).
GPT‑5.4 supports native computer-use capabilities, and deep reasoning across coding, documents, and multi-step agentic tasks — all running on Bedrock’s high-performance inference engine with isolated queues and durable state for fault-tolerant workloads. Your data stays in-partition and is never used to train models.
Amazon Bedrock now supports GPT‑5.4 from OpenAI in AWS GovCloud (US-West) — giving government and regulated industry customers access to OpenAI’s most capable frontier model for professional work, backed by the enterprise-grade security and goverment compliance scope of AWS GovCloud (US).
GPT‑5.4 supports native computer-use capabilities, and deep reasoning across coding, documents, and multi-step agentic tasks — all running on Bedrock’s high-performance inference engine with isolated queues and durable state for fault-tolerant workloads. Your data stays in-partition and is never used to train models.
For Regional availability of GPT-5.4 see the AWS Regions page. Read the launch blog to learn more, for documentation and a step-by-step walkthrough, see the Amazon Bedrock docs and the getting started blog.
Amazon Application Recovery Controller (ARC) Region switch helps customers orchestrate the failover of their multi-Region applications to achieve a bounded recovery time in the event of a Regional impairment. Today, we are announcing three new execution blocks — the Amazon Aurora serverless scaling execution block, the Amazon Aurora provisioned scaling execution block, and the Amazon Neptune global database failover execution block — which automate database scaling and failover for multi-Region workloads.
Customers running Amazon Aurora global database in active-passive configurations typically maintain a scaled-down secondary cluster to minimize cost. During failover, they must manually right-size and scale the secondary cluster to handle production traffic before routing requests — adding critical minutes to recovery time. The new Amazon Aurora serverless and Amazon Aurora provisioned scaling execution blocks automate right-sizing and scaling the secondary cluster as part of the Region switch plan, so it’s ready for production traffic when failover completes.
Customers running Amazon Neptune global database face a similar challenge: failover requires scripting or manually deciding whether to switchover or detach-and-promote depending on the outage type — all under the pressure of an active incident. The new Amazon Neptune global database failover execution block automates both planned switchover and unplanned failover scenarios within a single plan, eliminating custom scripting during recovery.
Amazon Application Recovery Controller (ARC) Region switch helps customers orchestrate the failover of their multi-Region applications to achieve a bounded recovery time in the event of a Regional impairment. Today, we are announcing three new execution blocks — the Amazon Aurora serverless scaling execution block, the Amazon Aurora provisioned scaling execution block, and the Amazon Neptune global database failover execution block — which automate database scaling and failover for multi-Region workloads. Customers running Amazon Aurora global database in active-passive configurations typically maintain a scaled-down secondary cluster to minimize cost. During failover, they must manually right-size and scale the secondary cluster to handle production traffic before routing requests — adding critical minutes to recovery time. The new Amazon Aurora serverless and Amazon Aurora provisioned scaling execution blocks automate right-sizing and scaling the secondary cluster as part of the Region switch plan, so it’s ready for production traffic when failover completes. Customers running Amazon Neptune global database face a similar challenge: failover requires scripting or manually deciding whether to switchover or detach-and-promote depending on the outage type — all under the pressure of an active incident. The new Amazon Neptune global database failover execution block automates both planned switchover and unplanned failover scenarios within a single plan, eliminating custom scripting during recovery. All three blocks support cross-account orchestration, enabling a single plan to coordinate database operations across multiple accounts and Regions. To learn more, read documentation of Amazon Aurora provisioned scaling, Amazon Aurora serverless scaling and Amazon Neptune global database failover
