Treating AI Like a Workforce
At Risk-!n 2026, Philippe Séjalon sat down with Shira Kaplan, Founder of Cyverse AG, fresh off the stage after a talk that sparked real conversations. One of her most striking analogies: AI agents should be treated like junior employees. Fast, capable, but also inexperienced and not completely reliable. According to Kaplan, this is not just her view. It is what her customers are saying too. These agents are automating tasks, reviewing documents, and even semi-processing payments. The question, she insists, is not whether they are useful, but how we monitor and supervise them properly.
Contracts With AI Agents?
Kaplan shared a striking example from her customer base, which includes CISOs of large enterprises in Switzerland. One client raised a compelling point: if you sign a contract with a junior employee to establish trust and accountability, why not do the same with an AI agent? Kaplan admitted the idea is ahead of its time, but acknowledged it is the direction things are heading. The lack of any contractual framework around AI agents is a genuine concern for security leaders today.
Agentic AI and the Expanding Attack Surface
On the topic of agentic AI, Kaplan was direct. These agents make organizations more efficient, but they also make them more vulnerable. Every new AI agent introduced into an organization increases its attack surface. Governing and securing these agents is now an emerging field, and Kaplan is confident that proper supervision is achievable, even if the tools and processes are still maturing.
AI Attacking AI
Perhaps the most sobering part of the conversation was about the evolving threat landscape. Kaplan described a reality that is already unfolding: AI on the attack is targeting AI on the defense. This is not science fiction. And the fundamental asymmetry remains unchanged. The attacker only needs to find one open path. The defender has to close every single door. That structural disadvantage for defenders, she noted with a smile, is precisely why her business exists.
Discover Cyverse AG
Cyverse AG is a cybersecurity company focused on helping large enterprises navigate the growing risks brought on by artificial intelligence and autonomous systems. Working closely with CISOs across Switzerland, the company addresses the security challenges of AI agents, agentic workflows, and expanding digital attack surfaces. Cyverse AG brings a pragmatic, business-driven approach to AI governance and defense, bridging the gap between fast-moving technology adoption and the security frameworks organizations need to stay protected. As the line between AI-powered attack and defense continues to blur, Cyverse AG positions itself at the forefront of this critical field.
