AWS launches Continuum and Context to secure AI agents

Lead

On June 21, 2026, Amazon Web Services announced two new services—Continuum and Context—to give AI agents the business awareness and security safeguards they have been missing.

Context

During its New York summit, AWS highlighted a growing problem: AI agents can generate code at speed but often lack the necessary understanding of an organization’s specific data landscape, leading to mistakes and hidden security flaws. The company said the gap stems from agents operating without a clear picture of corporate processes, policies, and data relationships.

Continuum is designed to automatically detect, prioritize, and remediate code vulnerabilities, acting as a continuous guardrail for AI‑generated software. At the same time, Context builds a knowledge graph from a company’s internal data sources, feeding that structured information to agents so they can make decisions that align with business rules and objectives. Both services are positioned as complementary fixes for the same underlying issue—agents that write code quickly but get things wrong too often.

Impact

Enterprises that have begun experimenting with AI‑driven development tools can now address two critical risk vectors at once. By surfacing code flaws in real time, Continuum reduces the likelihood of security incidents that could arise from unchecked AI output. Meanwhile, Context’s knowledge graph equips agents with the nuanced understanding required to respect compliance regimes, data governance policies, and operational constraints.

For developers, the dual offering promises smoother workflows: fewer manual code reviews and fewer costly rework cycles caused by context‑blind mistakes. For security teams, the automated vulnerability handling means a more predictable threat surface, allowing resources to focus on higher‑level risk management rather than chasing down AI‑induced bugs.

What’s Next

AWS indicated that both services are now available to customers, though detailed rollout timelines for different regions were not disclosed. Companies interested in tightening AI agent security are expected to pilot the tools through the AWS console, integrating them with existing Bedrock or SageMaker environments.

Looking ahead, AWS is likely to expand the knowledge‑graph capabilities of Context to ingest more diverse data formats and to enhance Continuum’s remediation engine with deeper integration into CI/CD pipelines. As AI agents become more embedded in enterprise workflows, the company’s focus on business context and security suggests a broader push to make AI‑assisted development both fast and trustworthy.