RSAC 2026: AI Agents, Secure Browsers, and Zero Trust Dominate Day One

RSA Conference 2026 kicked off in San Francisco on Monday with a flood of product announcements. The common thread: AI agents are coming to enterprise networks, and the security industry is racing to lock them down before attackers get there first.

Cisco introduces zero trust for AI agents

Cisco made the biggest WAF-adjacent move of the day. Their new Zero Trust Access for AI agents shifts the security model from access control to action control. Instead of handing agents long-lived credentials (the way humans typically authenticate), Cisco ties permissions to specific workflows and tasks. A monitoring layer evaluates whether agent actions are appropriate in real time.

For organizations running network security stacks alongside cloud WAFs like Cloudflare or AWS WAF, this matters. AI agents that interact with APIs, modify configurations, or trigger deployments need the same kind of least-privilege enforcement that WAF rules apply to HTTP traffic. Cisco is betting that action-level controls will become the standard for agentic security.

Palo Alto targets SMBs with Prisma Browser

Palo Alto Networks launched Prisma Browser for Business, a secure browser aimed at small businesses. It bundles phishing protection, ransomware defense, and data leak prevention into a browser that "doesn't require customers to make security decisions," according to CEO Nikesh Arora.

This is a direct play against browser-based attacks that often bypass traditional WAFs. For teams already using Prisma Cloud WAAS for application-layer defense, the browser adds client-side coverage that fills a gap most WAFs cannot reach.

CrowdStrike expands SIEM to Microsoft Defender

CrowdStrike announced Falcon Next-Gen SIEM now ingests Microsoft Defender for Endpoint telemetry. This is a market expansion play. WAF logs from providers like Cloudflare, Akamai, or Imperva are only useful if your SIEM can correlate them with endpoint data. CrowdStrike is making that correlation easier for the massive Microsoft Defender install base.

Wiz launches AI application protection

Google-owned Wiz debuted its AI Application Protection Platform (AI-APP), calling it the "natural evolution" of the CNAPP. It combines visibility, risk analysis, and runtime protection for AI applications in a single graph-powered platform. For teams deploying AI workloads behind WAFs like Google Cloud Armor or F5 Advanced WAF, Wiz adds the application-layer context that WAF rules alone cannot provide.

Other notable launches

Arctic Wolf announced an "agentic SOC" with hundreds of AI agents handling security operations workflows. Dell introduced quantum-ready security features for its commercial PCs. Rubrik launched a semantic AI governance engine for controlling autonomous agents. The trend is clear: every major vendor is building controls for AI agents, not just defending against them.

WAFplanet take

Day one of RSAC 2026 made one thing obvious: the security industry has shifted from "how do we detect AI-powered attacks" to "how do we govern AI agents in our own stack." That is a meaningful change.

For WAF teams, the implications are practical. AI agents will increasingly interact with APIs that your WAF protects. They will modify firewall rules, trigger deployments, and access sensitive data. The question is no longer whether your ModSecurity rules can block a malicious payload. It is whether your entire security architecture can tell the difference between a legitimate agent action and a compromised one.

Cisco's action-control approach is the right direction. WAF-level request filtering and agent-level action filtering are two sides of the same coin. Expect this pattern to show up across the WAF vendor landscape within the year.