Right now, everyone's selling AI-powered security operations. The pitch sounds great: faster detection, smarter triage, and less noise. However, if your logging is disorganized, your playbooks don't exist, and no one is responsible for the process, AI won't improve your SOC. It'll just make it faster at doing the wrong thing. In this episode of Follow the White Rabbit, Link11 CISO Kofi Osae-Attah sits down with Erik Van Buggenhout, NVISO co-founder and SANS instructor, to cut through the hype and discuss what AI in the SOC looks like in practice.

Erik is an AI optimist but, more importantly, he's a realist. He has spent years building security operations at scale and knows exactly where automation succeeds and where it falls short. His take? Up to 70% of incoming alerts can be automated without AI. Static playbooks, when built and maintained properly, do most of the heavy lifting, cheaply and reliably. AI earns its place where context matters, such as in dynamic environments, nuanced triage, and situations where a rigid playbook runs out of answers. The sweet spot isn't AI everywhere. It's AI where judgment is needed and automation everywhere else.

However, the conversation goes deeper than tools. Who's accountable when an AI agent makes a wrong decision? What will happen to the career path of junior analysts when L1 work disappears? Why does the security industry keep rebranding the same problems with new buzzwords every three years? Erik doesn't sugarcoat any of it, which is exactly what makes this episode worth your time.

1. AI won't fix a broken SOC. Garbage in, garbage out—faster. Before buying any AI tooling, first sort out your log sources, processes, and ownership.
2. Seventy percent automation is already possible without AI. Static playbooks that are properly maintained can handle most of the alert volume. AI is the next layer, not the foundation.
3. AI genuinely adds value through context. Managed service providers can't know every customer environment in detail. AI coupled with retrieval-augmented generation can provide that context on a large scale without requiring humans to memorize everything.
4. Humans remain accountable. AI agents operate with identities and permissions, but responsibility ultimately rests with the person operating them. Having a human in the loop isn't optional; it's a structural necessity.
5. The industry's buzzword cycle is exhausting and confusing. SIEM became XDR, and now XDR is becoming AI SOC. Same problem, new name. Erik argues for a more pragmatic and less dramatic approach to what's actually changing.

Listen in and subscribe to Follow the White Rabbit.

If this episode made you think twice about that AI SOC pitch in your inbox, good! Subscribe on your preferred platform and leave a review. It only takes 30 seconds, and it helps us reach those who need to hear this the most. Share it with your security team, your CISO, or anyone who's been handed an AI tool without a plan.

You'll find Erik on Linkedin and here more about NVISO.

If you want to dive deeper:

SOC-CMM – SOC Capability Maturity Model

SANS Institute – Purple Teaming & SOC Courses

MITRE ATT&CK – Detection & Response Framework

Gartner on AI in Security Operations (2024)

North Korea is infiltrating Fortune 500 companies with fake employees. They create authentic LinkedIn profiles, excel in remote interviews, collect salaries, and secretly steal intellectual property, cryptocurrency, and system access. This isn't a future threat. It's happening right now across more than 40 countries. In this episode of Follow the White Rabbit, Link11 ISO Kofi Osae-Attah sits down with Kritika Roy, a senior threat intelligence researcher at DCSO in Berlin. Together, they map the threat landscape that most security teams only partially see.

Kritika's work sits at the intersection of geopolitics and cybersecurity — and that intersection is where the full picture emerges. China is running long-term intelligence operations aligned with its five-year economic plan. Russia is focused on disruption and sabotage, especially since invading Ukraine. Iran is tracking dissidents and targeting organizations with Israeli ties. And North Korea? It's doing it all — stealing money to fund weapons programs, embedding operatives inside companies, and learning by doing. The line between nation-state espionage and cybercrime has blurred to the point of being nearly indistinguishable. Threat actors are buying ransomware on the dark web as if it were Amazon. Attribution is becoming more difficult. Defenders are falling behind.

The most important insight from this conversation isn't technical; it's contextual. Geopolitics determines who targets you, when, and why. A NATO summit, a trade dispute, or an election can trigger a wave of tailored phishing campaigns and targeted intrusions. Kritika's advice to security teams isn't to become intelligence agencies. Rather, it's to read the news, understand the motivations behind attacks, and stop treating every threat with the same level of urgency. Prioritize based on context. If you're hiring remotely, ask your candidates what the local food is like. You'll be surprised at how much that one question can reveal.

1. North Korean IT workers are already inside companies. They are hired through legitimate job platforms, work as regular employees, and use their access to steal money, intellectual property, and system knowledge. The fix? At a minimum, conduct one in-person interview.
2. Geopolitics is a threat intelligence tool. Phishing lures are timed to coincide with summits, elections, and conflicts. Knowing what's happening in the world allows you to anticipate what's coming at your organization.
3. The four main threat actors have different goals. China wants intelligence. Russia wants to cause disruption. North Korea wants money and knowledge. Iran targets dissidents and organizations related to Israel. Knowing who you're up against changes everything about how you defend yourself.
4. The line between cybercrime and nation-state activity is disappearing. Nation-state actors are purchasing off-the-shelf malware on the dark web. Attribution is becoming more difficult. Security teams need to adapt their thinking.
5. Fundamentals still win. Patch management, identity security, endpoint visibility, and regular red team exercises are not boring basics; they're essential. They're the difference between being resilient and being exposed.

Subscribe to Follow the White Rabbit.

If this conversation changed the way you think about hiring, threat intelligence, or geopolitics, tell someone. Subscribe on your preferred platform, leave a review, and share this episode with your security and HR teams. Both need to hear it.

Take a look at Kritika Roy's Linkedin profile or the DCSO Website

MITRE ATT&CK – North Korea Threat Groups

FBI Advisory: North Korean IT Worker Threat (2024)

Mandiant / Google: APT Overview by Nation State