AWS Unveils AI Agents That Find and Fix Code Vulnerabilities on Their Own

Amazon Web Services just raised the bar for enterprise AI at its annual Summit in New York. The cloud giant unveiled a wave of new products designed to let businesses deploy AI agents that work autonomously across security, software development, and data management. The announcements signal a major push to move AI agents from experimental projects into everyday production use.

Continuum Takes Aim at Code Security

The headline launch is AWS Continuum, an AI-native security service that continuously scans for code vulnerabilities, checks whether they are actually exploitable, and fixes them within guardrails set by human teams. Unlike traditional scanning tools that just flag problems, Continuum validates and prioritizes issues by business context before applying patches automatically. Organizations can start in a supervised learn mode and later graduate to enforce mode for fully automated remediation. AWS also introduced Continuum threat modeling, which generates comprehensive threat models directly from design documents or source code.

Knowledge Graphs and the Agent Ecosystem

Alongside Continuum, AWS launched Context, a service that automatically maps relationships across a company’s existing data into a knowledge graph. This gives AI agents governed access to business rules, domain knowledge, and data connections at runtime without requiring custom retrieval pipelines. Think of a knowledge graph as a smart map that shows how different pieces of company information relate to each other, helping AI agents find answers faster. AWS also expanded Amazon Bedrock AgentCore with managed knowledge bases, built-in web search, and new security integrations from partners like Check Point and Zscaler. Task volume on AgentCore has grown fifteenfold in the past six months, showing rapid enterprise adoption.

The results are already showing up in real deployments. Indian fintech startup Dhan used AWS AI development tools to build a new charting platform in just eight weeks with a single engineer, a project originally estimated to take twelve developers up to two years. Southwest Airlines also announced AWS as its preferred cloud provider, with plans to migrate its entire infrastructure by 2028 and put twenty seven hundred developers on AI-powered development tools.

These launches make it clear that Amazon is no longer just selling cloud storage and compute power. The company is building the AI workforce that runs on top of it, and betting that businesses willing to trust autonomous agents will pull ahead of those that do not.