CrowdStrike extends Falcon AI Detection across key gateways

CrowdStrike has extended Falcon AI Detection and Response across several AI gateway partners, including Google Cloud, Microsoft Azure and TrueFoundry.

The rollout also covers JetStream Security, Kong, LiteLLM and Maxim AI, broadening oversight of AI traffic moving through gateways, application programming interfaces and models.

The move comes as companies put more AI systems into live use and face a more fragmented security picture across tools and providers. That can leave gaps in visibility and increase the risk of prompt injection, data leakage and manipulation by attackers.

CrowdStrike's approach is to connect Falcon AIDR with partner gateways through native integrations, allowing security teams to view telemetry across AI interactions and apply policy controls where prompts, agents and models operate.

Daniel Bernard, chief business officer at CrowdStrike, set out the company's view of the market.

"Every enterprise has AI in production across multiple gateways, models, and APIs. That's not a future state, it's today's attack surface," said Bernard.

He added: "With Falcon AIDR and our AI gateway ecosystem, CrowdStrike makes the Falcon platform AI's security control plane, so organisations benefit from unified visibility and protection wherever their AI runs."

Partner network

The integrations are designed to place controls at the AI gateway layer, which sits between users or applications and the underlying models. This would allow organisations to inspect AI interactions, apply access controls and enforce security policies without changing existing system architectures.

Falcon AIDR is also intended to secure prompts, responses, agents and models in real time. CrowdStrike cited prompt injection, jailbreak attempts, and accidental or malicious data exposure among the risks it is trying to address.

Another element of the rollout is data sharing with Falcon Next-Gen SIEM, CrowdStrike's security information and event management product. This would allow AI-related activity to be viewed alongside endpoint, identity, cloud, software-as-a-service and third-party data sources.

Several partners said the integrations were aimed at improving oversight as AI services spread across enterprise systems.

"AI is rapidly becoming embedded across enterprise applications and APIs, which makes consistent governance and visibility critical. Integrating Falcon AIDR with Google Cloud Apigee helps organisations secure AI interactions across distributed environments while maintaining the flexibility developers need to scale AI services," said Vineet Bhan, director of security and identity partnerships at Google Cloud.

JetStream Security framed the issue as one that extends beyond network controls.

"AI isn't a network problem; it's a runtime reality. You can't secure AI with gateways alone that only provide surface-level visibility. CrowdStrike and JetStream operate where AI activity actually happens - across endpoints, browsers, SaaS applications, and the cloud. Together, we're helping organisations bring consistent visibility and control across AI environments as they scale," said Raj Rajamani, co-founder and chief executive officer at JetStream Security.

Kong pointed to governance concerns for companies operating models and agents at scale.

"Enterprises running AI at scale need security that's both model- and agent-aware, as well as threat-aware. With AI connectivity from Kong and complete AI attack path protection from CrowdStrike, organisations can scale AI without losing control of how those interactions are governed and secured. Together, this gives customers a more complete approach to securing AI in production," said Ken Kim, senior vice president of partners and alliances at Kong.

LiteLLM focused on avoiding disruption to software teams.

"Enterprises need to secure AI interactions without disrupting developer workflows. By integrating LiteLLM with Falcon AIDR, organisations can apply real-time protection and policy enforcement across AI traffic as it moves through their environments," said Krrish Dholakia, co-founder and chief executive officer at LiteLLM.

Maxim AI highlighted the role of inspecting prompts before they reach language models.

"Organisations need consistent enforcement for AI security policies across growing LLM environments. With Bifrost and Falcon AIDR, customers can inspect prompts before they reach models and automatically block or rewrite risky AI responses in real time. As AI agents become core to enterprise workflows, the gateway and the security layer need to operate as one. Our partnership with CrowdStrike is a strong signal of where the industry is heading, and we're proud to be building it together," said Vaibhavi Gangwar, co-founder and chief executive officer at Maxim AI.

Microsoft Azure described the integration in terms of centralised oversight across applications and services.

"Enterprises are looking for ways to operationalise AI securely across increasingly complex API ecosystems. By extending Falcon AIDR through Microsoft Azure API Management, organisations can apply centralised policy enforcement and gain deeper visibility into AI activity across applications, services, and models," said Balan Subramanian, general manager of Azure App Platform Services at Microsoft.

TrueFoundry said the architecture of modern AI deployments made consistent controls necessary across different frameworks and gateways.

"Customers are building AI applications across multiple frameworks and gateways, and security must follow that architecture. Our integration with Falcon AIDR enables organisations to enforce consistent guardrails across AI workflows as they scale," said Anuraag Gutgutia, co-founder and chief operating officer at TrueFoundry.