Enterprises face a significant challenge when deploying AI agents: the desktop and legacy applications that power most business workflows are simply inaccessible to modern AI systems. According to a 2024 Gartner report, 75% of organizations run legacy applications that lack modern APIs, and 71% of Fortune 500 companies operate critical processes on mainframe systems without adequate programmatic access. For many organizations, this has meant choosing between delaying AI adoption or undertaking expensive and risky modernization projects.

Today, we are announcing that Amazon WorkSpaces now enables AI agents to securely operate desktop applications without requiring application modernization. The same managed virtual desktops that millions of employees use and trust can now also serve AI agents, turning WorkSpaces into infrastructure for scaling enterprise productivity, not just delivering it. Because agents operate within your existing WorkSpaces environment, there are no APIs to build, no application migrations to plan, and no new infrastructure to manage.

Some of our customers had an early opportunity to give their agents a WorkSpace. Chris Noon, Director, Nuvens Consulting shared with us, “WorkSpaces lets our clients give AI agents the same secure, governed desktop environment their employees already use — no custom API integrations, full audit trails, and enterprise-grade isolation out of the box. For regulated industries, that’s not a nice-to-have — it’s the baseline.”

Secure cloud desktop access for AI agents
With WorkSpaces, AI agents can securely access and operate desktop applications running inside managed WorkSpaces environments to complete complex business workflows. Agents authenticate through AWS Identity and Access Management (IAM) and connect via Workspaces with complete audit trails available through AWS CloudTrail and Amazon CloudWatch. Because agents operate within secure WorkSpaces environments rather than on local machines, your existing security controls and compliance policies remain fully intact.

Amazon Workspaces supports the industry-standard Model Context Protocol (MCP), which means WorkSpaces works with any agent framework, such as LangChain, CrewAI and Strands Agents.

Let’s try it out
To set up a WorkSpaces environment for AI agents, I started in the AWS Management Console by creating a new WorkSpaces Applications stack—the environment definition that controls how agents connect and what they’re allowed to do.

From the Amazon WorkSpaces console, I chose Create stack and configured the basics: name, fleet association, and VPC endpoints. In Step 3 of the stack creation workflow, I noticed the new AI agents section with two options. The first, No AI agent access, is the default configuration for standard WorkSpaces designed for people. The second, Add AI Agents, allows AI agents to securely access and operate applications using their own identity and permissions. I selected Add AI Agents to enable agent connections on this stack.

Next, I will enable storage before configuring the agent access settings to define how agents interact with the desktop.

Under Agent features, I enabled three capabilities. Computer input allows the agent to click, type, and scroll within the desktop. Computer vision allows the agent to capture screenshots of the desktop, which is how it “sees” the application. Finally, screenshot storage configures where session screenshots are stored for audit and debugging.

Under Desktop screen layout, I set the screen resolution to 1280×720 and image format to PNG. The resolution determines the fidelity of what the agent sees during a session—a complex application with dense UI elements might benefit from higher resolution, while a terminal-style interface works well at 720p.

With my stack configured, WorkSpaces exposes a managed MCP endpoint. I pointed my agent framework to this endpoint, provided IAM credentials for authentication, and my agent began interacting with the desktop applications installed on the fleet’s image.

To see this in action, here’s an agent built with the Strands Agent SDK and Amazon Bedrock handling a prescription refill, looking up the patient record, searching for the medication, placing the order, and confirming a successful refill, all inside a sample pharmacy system with no API.

The application doesn’t know an agent is driving it. Nothing about the software was modified, rebuilt, or integrated. The agent worked with it exactly as it exists today.

Now available
This feature is available today in public preview at no additional cost in US East (N. Virginia, Ohio), US West (Oregon), Canada (Central), Europe (Frankfurt, Ireland, Paris, London), and Asia (Tokyo, Mumbai, Sydney, Seoul, Singapore) Regions.

Get started building today using our GitHub repo, or visit the WorkSpaces page for more details.