On June 17, 2026, the Javits Center in New York became the epicenter of enterprise agentic AI. At the AWS Summit New York 2026, Amazon unveiled a suite of tools that redefine how businesses — including SMBs — can build, deploy, and operate autonomous AI agents without specialized engineering teams. The three headliners were Kiro (the successor to Amazon Q Developer), Amazon Bedrock AgentCore with a dramatically expanded feature set, and Amazon Quick, a new agentic automation platform. The keynote, led by Dr. Swami Sivasubramanian, VP of Agentic AI at AWS, also featured a live production case study from Southwest Airlines, demonstrating that AI agents are already running at industrial scale in the real world.
What Did AWS Announce at the New York Summit?
The most disruptive announcement was the evolution of Kiro, AWS's agentic IDE built on the VS Code open-source base. Unlike its predecessor Amazon Q Developer — a chat-first assistant that responded to prompts with code — Kiro takes a fundamentally different approach: it generates a full project specification before writing a single line of code. Using EARS (Easy Approach to Requirements Syntax) formal notation, Kiro automatically produces three documents — requirements.md, design.md, and tasks.md — that guide the agent through property-verified implementation. The new Pro Max tier adds higher usage limits, access to the latest frontier models (including Claude Opus 4.8 and GPT-5.5 on Bedrock), and expanded agentic capabilities for dev teams. Amazon Bedrock AgentCore received four critical upgrades: Managed Harness (define an agent with just a model, a system prompt, and tools — no orchestration code required), the AgentCore CLI available in 14 AWS regions at no additional cost, Filesystem Persistence (agents can pause mid-task and resume exactly where they left off), and AgentCore Skills with native support for Claude Code, Codex, and Cursor.
"With Kiro and AgentCore, AWS has just lowered the entry barrier so that any SMB in Houston or Latin America can have its first AI agent in production in hours, not months."
Davarion Group & LabsReal Impact for SMBs
- 01Kiro Pro Max lets developers with no prior AI experience build complete agents in minutes: the IDE auto-generates the project architecture before coding begins, cutting errors and development time by up to 60%.
- 02AgentCore's Managed Harness eliminates orchestration code entirely: define your model + prompt + tools and AgentCore handles the rest — ideal for SMBs with small technical teams.
- 03Filesystem Persistence is a critical breakthrough for businesses that need long-running agents: an agent can process accounting documents, pause overnight, and resume the next morning at the exact same point, with full context intact.
- 04Southwest Airlines demonstrated live that AI agents already manage complex operations at production scale — a clear signal the technology is production-ready for companies of any size, not just tech giants.
- 05Key risk: the proliferation of AWS agentic tools demands a solid governance strategy from day one; without properly defining permissions in AgentCore Identity and AgentCore Policy, agents can take unintended actions on business-critical systems.
What AWS demonstrated today in New York is that the AI agent era for business is not the future — it is the operational present. The combination of Kiro (automating the agentic software development cycle), AgentCore (providing the execution runtime, memory, identity, and observability infrastructure), and Amazon Quick (connecting agents to enterprise workflows) forms a complete stack that enables an SMB to build its first automation agent without hiring an ML team. The AgentCore CLI is already available in 14 AWS regions — including us-east-1 and us-west-2, which serve Houston and all of Latin America — meaning any company with an AWS account can start TODAY. The 13 built-in evaluators in AgentCore Evaluations also ensure that deployed agents work correctly before touching production, dramatically reducing operational risk.
At Davarion Group & Labs we work with these exact technologies — AWS Bedrock, AgentCore, and models like Anthropic's Claude — to build custom automation agents for businesses in Houston, TX and across Latin America. If today's announcement raises questions about how your company can deploy its first AI agent on AWS, visit davarion.com and let's talk: we can have a working agent in your environment in under two weeks.
