AWS used its 2026 New York summit to introduce new capabilities for Amazon Bedrock AgentCore, a mobile version of Kiro, release-management features for AWS DevOps Agent, and autonomous modernization workflows in AWS Transform.
AWS announced a slate of AI and developer tooling updates at the AWS Summit in New York, 2026, with a focus on production AI agents, enterprise knowledge systems, coding assistance and application modernization.
In an AWS News Blog roundup of the summit announcements, AWS said it introduced new Amazon Bedrock AgentCore capabilities intended to help organizations build and operate production AI agents. The company described the updates as part of a broader push around agentic systems that can use company data, external information and operational controls.
AWS provided more detail in an AWS Machine Learning Blog post, saying the new Amazon Bedrock AgentCore capabilities include access to organizational, web and paid knowledge sources. According to AWS, the update is designed to let agents draw on broader information while remaining governed by enterprise policy controls.
The AWS Machine Learning Blog also said the release includes production optimization features and a generally available managed agent harness. AWS described the managed harness as infrastructure for running agents in production, rather than requiring customers to assemble every operational component themselves.
AWS also announced Amazon Bedrock Managed Knowledge Base, according to a separate AWS News Blog post. The company positioned the capability as a way to build enterprise generative AI applications using proprietary data and managed retrieval-augmented generation infrastructure.
Retrieval-augmented generation, often shortened to RAG, is commonly used to connect AI applications to company documents, databases or other private information. In its announcement, AWS said Bedrock Managed Knowledge Base is intended to help teams build generative AI applications that can retrieve relevant enterprise data without requiring customers to manage all of the underlying retrieval infrastructure.
The announcement fits with AWS’s broader Bedrock strategy: giving customers more managed components for building AI applications, while keeping those applications connected to enterprise data sources and governance requirements. The provided AWS materials do not state pricing or regional availability details in the excerpts, so those details should be checked against AWS’s service documentation before implementation.
The summit roundup from the AWS News Blog also said AWS announced Kiro for iOS. AWS described Kiro as part of its coding automation and developer tooling portfolio. The iOS release suggests AWS is extending access to the tool beyond desktop development environments, although the provided source excerpt does not specify feature parity, supported workflows or availability by region.
For developers, the significance of an iOS version will depend on what tasks AWS enables on mobile. The AWS News Blog excerpt identifies the product announcement but does not provide enough detail to confirm whether the mobile app supports code review, task planning, project monitoring or direct code generation.
AWS also said it announced release-management features for AWS DevOps Agent. According to the AWS News Blog summary, these features are part of AWS’s coding automation announcements at the summit.
Release management is a critical area for software teams because it involves coordinating deployment readiness, approvals, rollback planning and production changes. However, the source excerpt does not describe the exact functions AWS DevOps Agent now performs. Based on the announcement summary, the update should be understood as a release-management expansion rather than a fully described autonomous deployment system.
Another major item in the AWS News Blog roundup was AWS Transform, which AWS said now includes autonomous modernization workflows. The company framed the update around application modernization, a recurring enterprise concern as organizations migrate older workloads, refactor legacy applications and update infrastructure.
The wording in the AWS announcement points to more automation in modernization projects, but the provided excerpt does not specify which application types, programming languages or migration paths are supported. It also does not provide evidence about success rates, customer adoption or cost savings, so those claims should not be inferred.
Taken together, the AWS announcements show the company continuing to package generative AI capabilities as managed enterprise services. Amazon Bedrock AgentCore and Bedrock Managed Knowledge Base address the data and governance side of AI applications, while Kiro, AWS DevOps Agent and AWS Transform focus on software development and modernization workflows.
The most concrete information from AWS’s own posts is that the company is adding broader knowledge access, policy controls, production-oriented agent infrastructure, managed RAG capabilities, mobile access for Kiro, release-management features for AWS DevOps Agent and autonomous modernization workflows for AWS Transform.
Enterprises evaluating the announcements should review AWS’s service pages for availability, pricing, security documentation and integration requirements. The AWS blog excerpts describe the direction of the releases, but implementation decisions will depend on the details of each service and the constraints of existing development, data and compliance environments.
AWS announced a slate of AI and developer tooling updates at the AWS Summit in New York, 2026, with a focus on production AI agents, enterprise knowledge systems, coding assistance and application modernization.
The company described the updates as part of a broader push around agentic systems that can use company data, external information and operational controls.
AWS provided more detail in an AWS Machine Learning Blog post, saying the new Amazon Bedrock AgentCore capabilities include access to organizational, web and paid knowledge sources.
Continue reading
