Course · Training · Workshop

Defense in the Age of Agentic Attacks

One day for leadership: protect conventional systems and business processes against GenAI-assisted attacks, secure your own AI use and prioritize action.

Generative artificial intelligence (GenAI) and agents based on large language models (LLMs) change the tools available to attackers. They can support research, phishing and vulnerability discovery. Organizations that do not use artificial intelligence (AI) themselves are also affected: email, identities, web applications, cloud services and internal networks remain targets. This course focuses on the form of attack: how do you defend existing systems and business processes against GenAI-assisted and agentic attacks? Live demos show capabilities and limitations. You then prioritize measures for identity, vulnerability management, detection and recovery, and clarify ownership, budgets and decision paths. We also cover protecting your own LLM applications and agents: prompt injection, data leakage and excessive tool permissions. You leave with a prioritized roadmap for your organization and a crisis exercise behind you.

Course content details ↓

In-house course

Tailored for your team.

Request

No suitable date?

Request a new public date.

Request a date

What participants say

* Voices from this or related courses.

These customers booked courses in the same topic cluster.More customers →

Content

This training is for leaders and decision makers who want to protect conventional systems, business processes and their own AI applications against GenAI-assisted and agentic attacks. It combines live demos with concrete grounds for decisions; using AI yourself is not a prerequisite.

The course consists of the following topics and can be extended or adapted to the audience. For in-house courses we align examples and the tabletop exercise with your industry.

The new threat landscape

What GenAI and LLM agents change about attacks — and what they don’t
Live demo: deepfake call, GenAI phishing and agentic vulnerability discovery
Current cases and threat reports (NCSC, ENISA, vendor reports)

Assessing the risk

Attack paths targeting email, identities, web applications, cloud services and internal networks
Which attack scenarios are relevant for your organization, regardless of your own AI use
Crown jewels, dependencies and supply chain
Workshop: top 5 risks of your own organization

Protecting conventional systems and business processes

Map exposed services, prioritize vulnerabilities and shorten patching times
Identity: phishing-resistant MFA, privileged access
Processes: approvals with out-of-band verification against CEO fraud
Visibility, detection and when LLMs in the SOC pay off
Resilience: segmentation, backups, recovery

Securing your own GenAI use

Shadow AI, data leakage and policies people actually follow
Prompt injection and data leakage in LLM applications and agents
LLM agents with system access: least privilege, approvals, control and accountability
Governance with NIST CSF 2.0, NIST AI RMF and regulatory requirements (Swiss FADP, FINMA, reporting obligation)

Organization and people

Security awareness that keeps pace with GenAI attacks
Enabling teams: skills, roles, sourcing
Metrics leadership can steer by

Crisis readiness

Incident response plan and decision paths
Tabletop exercise: deepfake fraud and ransomware under time pressure
Communication with customers, regulators and the public

At the end of the day you have a realistic picture of the threat, a prioritized roadmap for your organization and a crisis exercise behind you.

The actual course content may differ from the above depending on the trainer, delivery, duration and the composition of participants.

Goal

Participants can assess GenAI-assisted and agentic attacks on their organization and prioritize countermeasures for conventional systems and business processes, even without using AI themselves. They can justify investments, assign responsibilities and prepare their organization for incidents. They also understand risks to their own LLM applications and agents and can guide their protection and governance.

Duration

1 day

Format

Compact workshop with live demos, short inputs, discussion and a tabletop exercise. Plenty of room for the participants' concrete questions and context.

Target group

CIOs, CISOs, CTOs, heads of IT, tech leads, team leads, security officers and executive board members responsible for IT and risk. No deep technical knowledge needed. For hands-on implementation in the team there is the course Agentic Blue Teaming.

Prerequisites

Experience in a leadership or responsible role in an IT environment. No security or LLM knowledge needed.

Preparation

Every participant receives a short questionnaire about their organization and threat landscape after registration. We align examples and discussions accordingly.

Request this course in-house

Request a public date

No suitable public date? Register without obligation — once there is enough interest we schedule a new public date and let you know first.

More about cyber defense in the age of agentic attacks

GenAI-assisted attacks also target conventional infrastructure without AI features. The course connects technical controls with organizational decisions: which systems are exposed, which identities are critical, who can approve payments or access, and how quickly can services recover? Securing your own AI systems builds on this foundation.

Further resources:

History and development

Cybersecurity was long a topic for the IT department. With waves of ransomware and regulatory requirements such as NIST CSF 2.0 (2024, with the new "Govern" function) and the Swiss reporting obligation for cyberattacks from 2025, it became a leadership responsibility.

GenAI accelerates this: since 2023 security agencies and vendors document the use of LLMs by attackers, deepfake fraud causes losses in the millions and LLM agents automate entire attack chains. At the same time organizations are adopting GenAI themselves — and need governance that brings security and innovation together.