2:30 pm

Registration

3:00 pm

Welcome & short introduction from Maria Pisa, Partner Technology Consulting - Cyber Security at EY

3:05 pm

Marco Cuomo, Managing Director at Cuomo ITConsulting GmbH and Partner at ACHILLES Consortium

Charting the Path to Trustworthy AI: Security is a Fundamental Pillar
Every AI solution must be Trustworthy! But what does that mean ? How is security related? Trustworthy AI (TAI) is built on a set of principles, many you already know and practice. 

We will share the security specific tools and techniques being researched in the ACHILLES EU Innovation Project demonstrating how TAI principles can be mapped to the Security Domain.

This enables you to determine where existing measures serve the purpose and where TAI demands something new.

3:35 pm

Jonathan Sinclair, Head of Cyber Security for Pharma Technical Operations at Roche

AI Governance? How Digital Identities serve as the "Central Nervous System" for your Automated Workforce
While AI adoption accelerates, governance frameworks struggle to keep pace, creating significant enterprise risk. This session delivers a synthesized roadmap of current deployment models before pivoting to the next frontier: Agentic Enablement and the critical role of Digital Identities in managing automated fleets.

We will challenge conventional wisdom and confront a sobering reality: current leadership often lacks the frameworks required to govern autonomous systems. Join us for a provocative analysis of how to bridge the gap between rapid technological evolution and robust, executive-led oversight.

4:10 pm

Mark Barwinski, Founder at CyberOps Advisory

From Patch Cycles to Continuous Remediation: Operating Security at Machine Speed
For two decades, vulnerability management has run on a cadence: discover, triage, schedule, deploy. AI-assisted vulnerability discovery has broken that rhythm. Anthropic's Claude Mythos Preview and Project Glasswing mark the inflection point: zero-days surface in volume and exploitation now follows disclosure in minutes. The window for scheduled patching has closed. In this short talk, Mark Barwinski reflects on what it means to suddenly find ourselves operating at machine speed, where real-time exploitation forces real-time remediation, and on the questions CISOs should be asking as the operating model compresses around them. Drawing on cyber operations leadership at UBS, Siemens, and the NSA.

4:30 pm

Presentation from a sponsor, if we are able to find one. Otherwise, we will be finished half an hour earlier.

5:00 pm

Coffee Break

5:30 pm

James Packer, Cybersecurity Transformation Lead at EY

AI Platform Security in Reality (Case Study, Lessons Learned)
The AI era is no longer emerging — it is here. Organizations worldwide are rapidly adopting AI platforms, yet the real security challenges often remain hidden beneath the surface.

In this session, EY will share insights from real AI platform security assessments, revealing what we actually see in live environments. Through selected case studies, you will discover lessons learned, unexpected security gaps, and overlooked back doors that traditional security approaches fail to address.

This is a behind‑the‑scenes look at AI security as it truly exists today — not in theory, but in reality.

6:00 pm

Marco Bertenghi, Senior Penetration Tester at Oneconsult

Hacking the Agentic Enterprise — A Penetration Tester's Field Report
Most AI agent penetration tests surface the same shape of problem

The findings split cleanly into two columns. On one side: agents with long-lived static credentials, no role-based access control, no just-in-time scoping for sensitive operations, and memory shared across every agent in the system — the kind of identity hygiene we would never tolerate for a human employee, somehow waived for autonomous software acting on the human's behalf. On the other side: no agent gateway in front of the agents or MCP gateway behind them, no AI firewall, no input or output guardrails, no centralised policy enforcement, no observability on tool calls — the kind of perimeter we would never ship a microservice without, somehow optional for systems that take untrusted input and call sensitive tools.

Most engagements find one column missing. Some find both.

This talk shows what each missing layer looks like, why the same pattern keeps appearing across different customers and stacks, and the checklist your team can apply to any agent before deployment.

6:30 pm

Apéro Riche (more like a dinner) & Networking till open end

Mark Barwinski
Founder at CyberOps Advisory

Marco Bertenghi
Senior Penetration Tester at Oneconsult