Not long ago, artificial intelligence in the workplace was a helpful assistant, drafting emails, summarising meetings and answering queries on cue. At re:Invent 2025, Amazon Web Services (AWS) unveiled a new class of โfrontier agentsโ, AI technology designed not just to assist, but to autonomously manage complex tasks for hours or even days at a time, freeing up workers to focus on more strategic, impactful, or just more fun work.
Modern professionals focus on high-value creative strategy while autonomous agents manage complex software development.Credit: iStock
These agents donโt wait for prompts. They diagnose problems and make decisions within defined guardrails independently, with its goal to complete the task assigned, while checking in with a human to make the final decision. In effect, they operate more like โwork assistantsโ than tools.
Solving complex challenges in minutes, not hours
Among the new releases showcased at re:Invent were three AWS frontier agents. Kiro, an autonomous software development agent, can understand and maintain context across an entire codebase, writing and refining code as a human developer would. DevOps Agent investigates and resolves tech system failures, often stitching together signals across company systems that would otherwise take skilled engineers hours to untangle. And Security Agent performs penetration testing and vulnerability assessments at a scale and speed that was previously impossible.
During testing at Commonwealth Bank of Australia, the DevOps Agent identified the root cause of a complex network and identity-management issue in under 15 minutes, a task that may have taken senior engineers hours. โAWS DevOps Agent thinks and acts like a seasoned DevOps engineer, helping our engineers build a banking infrastructure thatโs faster, more resilient, and designed to deliver better experiences for our customers,โ says Jason Sandery, head of Cloud Services at Commonwealth Bank. โThis isnโt just about faster resolution times, itโs about maintaining the trust our customers put in us.โ
To ensure these agents operate securely and responsibly, there are guardrails to provide real-time governance and centralised security controls that protect sensitive data and prevent misuse as the agent is working. These enterprise-grade safeguards ensure compliance and responsible AI deployment.
The agents sit alongside a major expansion of AWSโs Nova AI models, which underpin how these systems perceive, reason and communicate. The new Nova 2 Pro and Nova 2 Lite models balance accuracy and speed for enterprise workloads, while Nova 2 Sonic introduces natural, real-time voice interaction across more than 30 languages.
In practical terms, that opens the door to AI-powered customer service that can hold fluid, conversational dialogue rather than scripted exchanges.
Nova Forge allows organisations to build their own AI models by securely blending their proprietary data with AWSโs Nova models for frontier intelligence. This approach also gives businesses tighter control over accuracy, privacy and intellectual property.
