AWS unveils Continuum, Context to secure enterprise AI agents
Serge Bulaev
AWS introduced two new services, Continuum and Context, to help manage and and secure AI agents in large companies. Continuum may help find and fix software weaknesses by checking code and suggesting solutions, while Context appears to build a knowledge graph so AI agents can access data safely and with rules. Both services are in early testing as of June 2026, and there is no public pricing yet. Experts suggest these tools might address common problems with AI agent coordination and security, but extra safety steps may still be needed. AWS has not said when these tools will be widely available.
To help secure enterprise AI agents, Amazon Web Services unveiled Continuum and Context, a pair of new services designed to coordinate and govern AI systems across large organizations. The offerings focus on the data layer and DevSecOps pipeline rather than the language models themselves.
AWS Continuum is a gated-preview AWS service for code vulnerabilities that discovers, prioritizes, validates, and remediates security risks across the software lifecycle. The service is currently in a gated preview without public pricing. Meanwhile, Context builds a knowledge graph from enterprise data, enabling AI agents to access governed business rules and data relationships at runtime.
These preview releases directly address two major pain points in scaling multi-agent AI: persistent memory and secure release management. According to industry analysts, poor coordination between AI coding agents can introduce defects that halt entire releases. Continuum is designed to mitigate this risk by detecting such issues before they spread throughout the system.
What Are the Core Functions of Continuum and Context?
AWS Continuum is a gated-preview AWS service for code vulnerabilities that discovers, prioritizes, validates, and remediates security risks across the software lifecycle. AWS Context builds a governed knowledge graph from company data, allowing AI agents to securely access approved information and business rules, ensuring compliance and consistency across automated workflows.
- Context: Constructs a continuously updated knowledge graph, enabling AI agents to query approved data under strict policy controls.
- Continuum: Scans code repositories, ranks vulnerabilities by business impact, and automates pull requests for fixes that align with existing DevOps workflows.
AWS documentation highlights key governance features, including identity propagation, comprehensive audit trails, and native integration with Amazon Bedrock for model access. This aligns with a growing enterprise trend, where orchestration platforms are evaluated on their monitoring, policy enforcement, and interoperability capabilities, not just model performance.
Availability and Pricing
Both Continuum and Context are in preview phases. While official list pricing is not yet available, prospective customers should budget for standard AWS infrastructure costs, including token usage, storage, and data transfer. Enterprise support plans may also be required for organizations implementing these services.
How Continuum and Context Fit in the Market
Continuum and Context position AWS to compete directly with governance layers from Microsoft and Google. According to industry reports, enterprises are demanding centralized control planes to manage AI workflows, identity, and compliance. By integrating a knowledge graph and a security remediation loop, AWS aims to provide a native solution that prevents customers from turning to third-party orchestration frameworks like CrewAI or LangGraph.
Addressing New AI Cybersecurity Threats
The rise of agentic AI accelerates both development and exploitation, creating new security challenges. For example, one audit found that a significant portion of AI agent skills in a public marketplace were malicious. Continuum is designed to counter these threats by validating a vulnerability's exploitability before recommending a patch - a critical step that traditional SAST tools often miss as AI blurs the lines between proprietary and generated code.
Security experts recommend that organizations using these preview services implement additional safeguards, such as SBOM scanning and cryptographic verification, alongside Continuum's automated fixes. This suggests Continuum is intended to be a key component within a comprehensive security and governance framework, not a standalone solution.
AWS has not yet provided a timeline for regional rollouts or general availability. To gain early access, organizations must apply for the gated preview, which requires demonstrating a clear enterprise DevSecOps use case.