AWS Makes AgentCore Harness Generally Available: From Idea to Production AI Agent in Minutes

Amazon Web Services has announced the general availability of Amazon Bedrock AgentCore harness, a managed capability designed to let developers create production-grade AI agents without writing orchestration code or building a container.

The announcement was made at AWS Summit New York 2026, with the official blog post published on June 18, 2026. While this is the general availability milestone, the underlying technology isn't brand new — a preview version, originally called the Managed Agent Harness, first appeared in April 2026.

The headline message from AWS is simple: "Go from idea to production-grade agent in minutes."

The standout claim is the speed and simplicity of getting started. With just two API calls — CreateHarness to define an agent, and InvokeHarness to run it — you can have an agent running in seconds.

The idea is that developers declare what an agent does: the model it uses, the tools it calls, and the instructions it should follow. From there, AgentCore handles the rest. According to Amazon's newsroom, the service "does the hard work for you," assembling the orchestration loop, tool execution, memory management, context handling, and error recovery.

Notably, the GA version has been streamlined since the April preview, which was described as requiring three API calls. The current release brings that down to two.

Each agent runs in its own isolated environment complete with a filesystem and shell, allowing it to read files, run commands, and write code safely.

The agents can remember users and conversations across sessions, pick up skills you point them at — including an AWS-curated catalog — browse the web, and call your tools through a gateway or MCP. They can even switch model providers mid-session without losing context.

Every step streams back in real time and is automatically traced to Amazon CloudWatch, giving teams visibility into what the agent is doing. And while no orchestration code or container is required, developers retain the option to build those themselves if they choose.

The launch is part of a wider wave of AI agent announcements at AWS Summit New York 2026, signalling how quickly the major cloud providers are moving to make agent-building more accessible.

Independent analysts and technical media have confirmed the launch venue and timing. Coverage from Constellation Research and developer-focused outlets noted that by simply configuring models, tools, skills, and instructions, production-quality AI agents can now be up and running in minutes.

For businesses watching this space, the trend is clear: the technical barrier to deploying useful AI agents is falling fast. What once required specialist engineering effort is increasingly being packaged into managed services.

The promise of going from idea to working AI agent in minutes is genuinely exciting — but speed of deployment is only half the story. An AI agent is only as useful as the data it can reach.

An agent that can browse the web, call tools, and remember conversations still needs accurate, connected information about your business to make smart decisions. If your customer records sit in one system, your financials in another, and your operational data somewhere else entirely, even the most capable agent will be working with a fragmented picture.

This is where a unified, AI-ready data foundation matters. By bringing your business data from across your CRM, accounting, and other systems into one place, you give any AI tool — whether it's an agent platform like this one or your own analytics — a single, trustworthy source to work from.

At Intellova, that's exactly what we focus on: unifying your data so that when the next breakthrough in AI arrives, your business is ready to use it from day one.