By As generative AI (GenAI) becomes a core part of modern applications, securing AI-powered systems is no longer optional—or siloed. According to Stas Neyman, Director of Product Marketing at Akamai, AI security must become a foundational component of your architecture.
In a recent TFiR conversation, Neyman pushed back on the idea that AI security is just a bolt-on at the edge. “I truly believe that AI security is part of a broader strategy,” he said. That strategy includes segmenting infrastructure, monitoring applications, securing APIs, and—crucially—sharing threat intelligence across all layers.
📹 Going on record for 2026? We're recording the TFiR Prediction Series through mid-February. If you have a bold take on where AI Infrastructure, Cloud Native, or Enterprise IT is heading—we want to hear it. [Reserve your slot
Akamai’s approach emphasizes defense in depth. If a threat actor probes a firewall on one side of the enterprise, that intel can inform API security posture elsewhere. “We can say, ‘Hey, this IP is suspicious. If it’s hitting your API, raise the risk,’” Neyman explained.
But Akamai is also rolling out tools purpose-built for the new wave of AI-driven software. Enter Firewall for AI, a solution that protects generative AI apps by identifying and securing the APIs that interact with large language models (LLMs). “We actually tell you which APIs are interacting with your GenAI,” Neyman said. “Then we feed that information into the firewall.”
That handoff is critical. Firewalls don’t automatically know what to monitor—they need context. Akamai provides that foundation by mapping API-to-AI interactions and flagging which endpoints require guardrails, input validation, or closer scrutiny.
The result is a tightly integrated security workflow that spans discovery, monitoring, and real-time enforcement. It also acknowledges the architectural reality: APIs are the connective tissue of GenAI systems. They’re where sensitive data flows, model outputs are served, and decisions are made.
“Without us,” Neyman added, “the firewall wouldn’t know what to protect.”
For teams building or securing AI-native applications, this is a glimpse of what modern appsec looks like. Visibility-first, intelligence-driven, and built to defend a fast-evolving attack surface.
