SEATTLE: Amazon Web Services launches a preview feature letting AI agents access and operate WorkSpaces virtual desktops with assigned identities.
The capability marks a significant step toward autonomous enterprise computing. AI agents now perform tasks on virtual machines just like human employees do. Each agent operates under its own assigned identity within the WorkSpaces environment.
The preview targets enterprise customers building agentic workflows at scale. Companies can deploy agents that handle repetitive desktop tasks across business systems. Agents access legacy applications, web tools, and productivity software without human intervention.
AWS positions the launch as foundational infrastructure for the agent era. The cloud giant competes directly with Microsoft and Google on agentic platforms. Each hyperscaler races to embed AI agents inside their core enterprise stacks.
WorkSpaces serves thousands of enterprise customers globally with virtual desktop infrastructure. The platform hosts Windows and Linux desktops accessible from any device. Adding agent identities lets companies extend automation across these environments seamlessly.
Identity-based access creates clear audit trails for agent actions on desktops. Security teams track which agent performed which task at what time. The model addresses governance concerns blocking many enterprise agent deployments today.
“Agents need first-class identities to operate safely in regulated environments,” an AWS spokesperson says about the preview launch.
The release arrives amid a broader AWS push into agentic AI infrastructure. The company recently expanded Bedrock with new agent orchestration capabilities. Q Developer also continues gaining traction among enterprise coding teams.
Microsoft pushes similar agent capabilities through its Copilot Studio platform. Google promotes agent features through Workspace and Vertex AI offerings. The three-way race intensifies competition for enterprise automation budgets globally.
AWS has not disclosed pricing or general availability timing for the feature.
Previous
