Schedule

James Bond has a licence to kill, but not a free pass. Every mission is bound by agreed rules, and anyone who deviates from the plan must be judged against the overarching objective. We need exactly this logic for autonomous AI agents that increasingly make bookings, write code, access data and trigger workflows independently. In this talk, we apply intelligence service mechanisms to AI governance: which permissions agents receive, where the four-eyes principle applies, why agents should understand objectives rather than merely follow instructions, and how licences are revoked. Participants will take away a practical framework for permission models, approval and escalation processes, and kill-switch mechanisms, designed so that governance enables innovation rather than slowing it down.

AI is accelerating cyber attacks exponentially. Conventional response processes, fixed patch cycles and identity as the last perimeter are outdated. What worked yesterday no longer protects us today. It is time to challenge every defensive assumption and redesign security for the speed of AI.

- Cyber readiness must be proven, not assumed.

- Hands-on practice builds knowledge, skills and abilities that training alone cannot.

- License to operate: Prove people and AI agents are ready.

What if an AI could challenge your own applications, including mobile apps, APIs and AIs at the push of a button – always from the attacker’s perspective?

Tenzai finds not only IT vulnerabilities but also exploitable weaknesses in processes.

These days, exploits emerge within hours, and AI applications are growing faster than security teams can keep up with. This session demonstrates how AI agents enrich findings with context and suggest fixes that your team can review and roll out. The machine does the work; humans make the decisions.

- AI is rapidly changing how organisations build and operate in the cloud. 

- This session explores the evolving AI security landscape and how teams can both secure their use of AI and use AI to improve security outcomes. 

- Learn how Wiz delivers visibility, context-driven risk insights, and protection across the AI application stack, with a live demo included.

- Operational resilience through continuous application protection, straightforward, documented testing, and automatic recovery

- Modernize your infrastructure with confident, testable migrations

- Keep AI data available, secure, and consistent

Attackers now move in seconds. Recovery still takes weeks. AI widens this gap from both sides. This session shows how companies build resilience across data, identities and AI agents.

- Why recovery speed now determines the cost of an incident, rather than prevention

- Which questions must be answered in advance so recovery can begin immediately when needed

- Why identity is the central target of attacks and why you must not have to choose between "clean" and "up to date"

- Why agentic AI is the next gap in resilience: seeing what agents do, controlling what they are allowed to do, and undoing what goes wrong

The vulnerability nobody really wants to acknowledge

Identify and assess risks

Speed as a guarantee of security

Mathias Espeloer (Head of IT at Heuking) and Thomas Esslinger (Regional Sales Manager at Horizon3) explore how to focus on what matters in the age of AI and the vast number of vulnerabilities.

- Goals and wishful thinking in IT/OT/IS 3rd Party Risk Management

- Challenges concerning timeliness, confidentiality, availability ..

- TPRM as part of ERM and nothing more?

- Our approach at Schmitz Cargobull AG and day-to-day practice

Copilot and similar tools, together with autonomous AI agents, create new requirements for security, transparency and governance.

But policies alone are not enough: what matters is how they are translated into concrete guardrails for data access, prompts and AI agent behaviour.

This session shows how companies can develop AI security from a governance concept into effective technical controls, and how they can trace which data AI accesses, what it is supposed to do and what it actually does. This creates an end-to-end chain of control and evidence for people, data and AI agents.

AI agents are changing the rules: they no longer simply analyse information, but independently access data, applications and systems. This creates entirely new attack surfaces, from shadow AI and prompt injection to uncontrolled actions with far-reaching consequences.

But how do you protect systems that act independently? And how can you enable AI innovation without losing control of security and data?

Discover why AI security must go far beyond conventional governance, and the role that visibility, real-time monitoring and AI Detection & Response play in the secure use of autonomous AI.

Using advanced technologies to strengthen resilience against cyber attacks.

Developing innovative, sustainable strategies to protect critical production systems.

- A leaked credential trips no alarm. It just works, with every permission it carries.

- Your vault, scanners and EDR each guard a slice. The credential layer runs across all of them.

- How credential layer security finds, prioritizes and closes exposure across that layer.

Digital sovereignty without sacrificing functionality.

Reduce the attack surface to relieve IT teams.

Insights from the current threat landscape: how attackers deliberately exploit trust, identities and context – and why classic defenses based on IOC/IOA (Indicator of Compromise/Attack) miss these attacks.

How behavior-based, AI-native security identifies and stops new and future attack patterns before damage occurs.

Practical lessons: can behavior-based AI security solutions be integrated quickly and simply, ready to use without complex rules?