As enterprises accelerate their adoption of generative AI (GenAI), securing AI-powered applications has become mission-critical. In this clip, Rupesh Chokshi, SVP and GM of Application Security at Akamai, introduces Firewall for AI, a solution designed to safeguard AI interactions in real time.

Tackling AI-Specific Threats Head-On

Akamai’s Firewall for AI addresses threats unique to large language model (LLM) usage—such as prompt injection, data leakage, and toxic outputs. “We built it for customers we serve… to secure the AI traffic end to end—whether it’s inbound or outbound prompts,” says Chokshi. With visibility at the edge, the firewall can intercept and either alert, modify, or block risky AI interactions dynamically.

Real-World Detection at Scale

During a technical preview with a digital mortgage services provider, the firewall detected 5–6% of interactions involving sensitive data leakage, PII exposure, or prompt injection attempts. “The product capabilities are real, and the threats are real,” Chokshi notes, reinforcing the tool’s relevance for businesses rolling out GenAI assistants and agents.

Flexible, Low-Friction Deployment

Deployment is a strong suit: existing Akamai customers can activate the firewall inline, while others can integrate via API. This frictionless setup ensures rapid time-to-value across regulated and non-regulated sectors alike.

Compliance-Driven by Design

With frameworks like OWASP’s LLM Top 10, NIST, and the EU AI Act shaping governance expectations, Akamai’s Firewall also serves as a compliance enabler. “More and more customers are being asked: what’s your governance model for GenAI?” Chokshi observes.

As AI becomes central to customer experience and business automation, Akamai’s Firewall for AI positions itself not just as a gatekeeper—but as a critical enabler of secure, compliant innovation.