Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Some of the security industry is still cautiously evaluating its relationship with AI. They are weighing questions, sitting with uncertainty, and waiting for something to ease their concerns about trusting AI in production. This post isn't for that group. This is for AI tool developers already in motion. The ones who vibe-coded a log parser over a weekend, spun up local inference on dedicated hardware, or ran cross-model research pipelines across multiple data sources.

AI tools are moving faster than the security controls meant to govern them.In this episode of Defender Fridays, Cisco's Cybersecurity Technical Solutions Architect Katherine McNamara walks through changes in the threat landscape as organizations rush to integrate AI without applying basic security discipline. When Katherine meets with customers to discuss AI security, the conversation almost always starts and ends in the same place: data leakage. Someone might upload sensitive files to a public LLM.

Most multi-agent security deployments fail in production not because the agents can't act, but because there's no shared context layer between them. When something goes wrong, the audit trail doesn't exist. In LimaCharlie, solving that problem is architectural, and the solution starts with how individual agents are defined.

Anyone who's stood up a SIEM from scratch knows the feeling: weeks of infrastructure work, integration headaches, and a services team alongside for the whole process. That experience shaped how people think about adopting anything new in security ops. The instinct is to treat AI the same way: budget for it, plan for it, bring in specialists. This instinct is costing teams real time. Traditional infrastructure takes great effort to stand up. Infrastructure-as-code happens in seconds.

Security operators often struggle with the escalating friction that naturally occurs in their detection and response (D&R) workflow. Detections fire in one tool. Investigations happen in another. Case tracking lives in a third. For MSSPs managing dozens of client environments, fragmentation compounds quickly. Analyst time bleeds into context-switching. SLAs are hard to track. When something goes wrong, reconstructing what happened across multiple platforms is painful.

Most current demonstrations of AI in security operations are lackluster. You ask a chat interface a question, get a summary, and maybe a suggested next step. The operator still does all the work, at human speed. Meanwhile, adversaries are already deploying AI offensively against their targets. AI in SecOps must ultimately be an operator. Otherwise, the gap between adversary and defender will become too wide to bridge. LimaCharlie Co-founder, Christopher Luft, demonstrates a simple way to get started.

Most AI detection engineering puts a human in the loop at every step. David Burkett envisions an efficient and effective pipeline architecture that does not. David is a security researcher at Corelight Labs and a longtime LimaCharlie community member. He appeared on a recent episode of Defender Fridays to walk through his vision of a fully agentic detection engineering pipeline. His system uses LimaCharlie as its operational backbone.

Detection engineering is fundamentally a translation problem: rules need to be converted between formats, IOCs need to be converted into detection logic, and noisy alerts need to be converted into precise suppressions. That translation work is what consumes analyst time, and it's what Claude Code handles well.

A credential access event fired. An AI agent investigated it, correlated it against running processes, assessed the risk, and closed the ticket. No analyst touched it. The entire loop ran in minutes. This is what security operations look like when AI can actually operate in the environment rather than advise from outside it. Security operations have always required a special kind of person.

Every new customer shouldn't cost you headcount. But with stack diversity, it usually does.