Wiz expands AI security coverage across cloud & edge

Wiz has expanded its AI Application Protection Platform with new coverage across cloud platforms, AI development tools and edge services. It also introduced Red Agent in public preview.

Announced alongside Google Cloud Next, the additions extend Wiz's security platform into AI coding tools, agent studios, multicloud platform services and internet edge infrastructure.

Among the updates is support for Databricks, giving customers more visibility into where sensitive data sits and how identities, infrastructure and access patterns combine to create risk. Coverage has also expanded across AI studio environments including AWS Agentcore, Gemini Enterprise Agent Platform, Microsoft Azure Copilot Studio and Salesforce Agentforce.

Wiz is also broadening support across multicloud platform-as-a-service environments, including AgentCore, as customers run applications across a mix of cloud, software-as-a-service, platform services and on-premise infrastructure.

Red agent

A central part of the update is the public preview launch of Red Agent, an AI system designed to identify and validate complex vulnerabilities by modelling attacker behaviour.

The release builds on the earlier AI-APP launch, which was aimed at securing AI applications from software code through to runtime. That broader platform also includes Blue Agent, which is generally available, and Green Agent, which is in public preview.

The latest additions are intended to address a changing threat environment as AI models and autonomous agents gain greater access to live data and sensitive systems. Wiz argues this raises the impact of basic security failures such as misconfigurations, excessive identity permissions and unpatched software flaws.

Intel centre

Another new product is the Technology Intel Centre, which brings together information on feature releases, migration changes and end-of-life notices across cloud and AI providers inside the Wiz platform.

The service is designed to show customers which of their resources are affected by those external changes. For organisations using Wiz Cloud Cost, it will also indicate whether updates could alter cloud spending.

The expansion also reaches earlier into software development. Wiz said AI-assisted coding is speeding up application development, but also introducing new security weaknesses and visibility gaps for security teams.

According to Wiz Research, analysis of real applications built with AI-assisted coding tools found that 20% contained significant security issues, including broken access controls and exposed data endpoints.

Code tools

To address that, Wiz is adding three new functions to Wiz Code. The first, AI-BOM, is designed to create an inventory of the AI frameworks, models and integrated development environment extensions used across an organisation's coding estate.

The tool can identify frameworks such as LangChain and developer tools including Gemini Code Assist, GitHub Copilot and Cursor. The aim is to reduce so-called shadow AI, where staff use AI coding tools outside standard review processes.

The second addition is a set of security guardrails that scan AI-generated code in tools such as Lovable before it is committed to source control. These controls can apply organisational security rules at the point code is generated and stop insecure code from progressing.

The third feature focuses on remediation. Wiz is releasing pre-built skills for AI development environments including Claude Code and Cursor, allowing AI agents to pull in validated findings from the Wiz Security Graph, analyse code and apply fixes within the developer console.

Edge coverage

Wiz is also extending visibility beyond core cloud infrastructure through integrations with Cloudflare, Akamai, Vercel and Google Cloud Apigee.

Those integrations are designed to bring edge services, front-end deployments, DNS configurations and API gateways into the Wiz Security Graph. In practice, that means security teams can connect risks found in those external layers to vulnerabilities deeper in their infrastructure.

For Cloudflare, the integration is intended to show how AI applications are exposed through DNS and infrastructure and to identify unprotected endpoints. With Akamai, the platform analyses edge DNS and property manager settings alongside cloud workloads. Vercel support brings projects, domains and firewall settings into the same view, while Apigee support maps API gateways and environments.

These additions reflect how businesses now operate across dispersed environments where security teams need a single view of risk spanning code, cloud, AI services and the internet edge.

Wiz said its objective is to give customers coverage "from the first line of AI-generated code, through AI and agent studios, all the way to the edge of the cloud."