Marcello Filacchioni : 24 September 2025 18:05
At Fal.Con 2025, the annual conference that brings together thousands of cybersecurity experts from around the world, CrowdStrike made one thing clear: cyber defense is entering a new era, that of “AI agents.”
For years, CrowdStrike has led the industry with endpoint protection and its Endpoint Detection and Response model. Today, the same approach is being applied to artificial intelligence. With the acquisition of the startup Pangea , the company aims to secure every aspect of enterprise AI: from models to virtual agents, to simple conversations with a chatbot.
This is how the concept of AI Detection and Response (AIDR) was born, a sort of “antivirus of the future” capable of intercepting sophisticated attacks such as prompt injections and preventing abuse or risky uses of generative systems.
CEO George Kurtz spoke of a true paradigm shift for Security Operations Centers. Today, attacks are no longer measured in days or hours, but in seconds. This is why CrowdStrike proposes the Agentic SOC : no longer a team of analysts inundated with alerts, but an environment where intelligent digital agents work alongside people , analyzing anomalies, making decisions, and acting autonomously.
At the heart of this revolution for Crowdstrike is Charlotte AI , the system that orchestrates agents and coordinates their actions. What’s more, thanks to Agent Works , companies can create their own customized agents with a no-code interface, as if they were “hiring” new digital colleagues specialized in security.
CTO Elia Zaitsev then presented the Enterprise Graph , a model that reconstructs the entire corporate infrastructure in real time —users, assets, identities, and data—offering a unified, natural-language queryable view. This approach allows for the transition from vulnerability identification to automatic remediation plan generation in a matter of moments, drastically reducing response times.
Also presented at Fal.Con was the Adversary Strategy Program , which allows CrowdStrike to replicate and anticipate hackers’ moves to make the platform increasingly resilient. This has led to solutions like Apex , an AI model capable of unmasking malicious use of legitimate processes, and new techniques to combat ransomware and the abuse of remote management tools.
In a world where “prompts are the new malware,” as was said on stage, CrowdStrike aims to turn the fear of AI into an opportunity: if cybercriminals leverage generative models to accelerate their attacks, companies can respond with AI agents that work tirelessly, around the clock, alongside security teams.
The promise is clear : with Falcon, Pangea, and the intelligent agent ecosystem, CrowdStrike wants to do with AI what it did with endpoints years ago: raise the bar and redefine, once again, what it means to “stop breaches.”
Marcello Filacchioni