The 5 Questions to Ask Before You Ship an Agent in Europe
What EU AI Act compliance actually requires when AI agents move from development into production
Your agent just called a third-party agent, which used an over-provisioned credential to pull data it should never have seen.
Now you have a security incident, a potential compliance problem, and a question nobody seems able to answer. Who's liable? The model provider, the agent vendor, or you?
And can you even reproduce what happened?
If you don't have an answer, you're not alone.
The EU AI Act was written with AI systems in mind. But AI agents are different. They call other agents, access tools, use credentials, and take actions with consequences that can be difficult to predict or reconstruct after the fact.
That makes EU AI Act compliance a whole different challenge when AI agents are running in production. Who's responsible when an agent goes off-script? What does meaningful human oversight look like when agents act autonomously? And what happens when an agent-related security incident triggers regulatory reporting obligations?
Depending on the infringement, EU AI Act penalties can reach 3% of global annual turnover, or up to 7% for certain prohibited AI practices. While implementation timelines have evolved, the underlying questions around AI agent security, accountability, and governance aren't going away.
In this 30-minute expert debate, Andrea Di Bari, Technical Director at Axon Technologies, Marius Poskus, Cybersecurity Leader and CISO, Jon Hammant, CEO of Orbital and former AWS Specialist Lead for UK&I, and Ryan Sheldrake, EMEA CTO at Straiker, tackle the five questions that matter before you put AI agents into production in Europe.
Expect a practical conversation about EU AI Act compliance, agentic AI security, human oversight, liability, and incident reporting, with insights from experts who understand the challenges of securing and deploying AI systems.
Shipping an AI agent is really just the start. Knowing how to secure it, govern it, and account for its actions is what your team needs to be on top of today.
Join us before your agents outrun your compliance plan.
Webinar details
- Date: October 22, 2026
- Time: 11:00 AM CET
- Duration: 30 minutes
- Format: Expert debate and discussion
Meet the speakers
- Andrea Di Bari — Technical Director, Axon Technologies
- Marius Poskus — Cybersecurity Leader and CISO
- Jon Hammant — CEO, Orbital; former AWS Specialist Lead, UK&I
- Ryan Sheldrake — EMEA CTO, Straiker
The 5 Questions to Ask Before You Ship an Agent in Europe
What EU AI Act compliance actually requires when AI agents move from development into production
Your agent just called a third-party agent, which used an over-provisioned credential to pull data it should never have seen.
Now you have a security incident, a potential compliance problem, and a question nobody seems able to answer. Who's liable? The model provider, the agent vendor, or you?
And can you even reproduce what happened?
If you don't have an answer, you're not alone.
The EU AI Act was written with AI systems in mind. But AI agents are different. They call other agents, access tools, use credentials, and take actions with consequences that can be difficult to predict or reconstruct after the fact.
That makes EU AI Act compliance a whole different challenge when AI agents are running in production. Who's responsible when an agent goes off-script? What does meaningful human oversight look like when agents act autonomously? And what happens when an agent-related security incident triggers regulatory reporting obligations?
Depending on the infringement, EU AI Act penalties can reach 3% of global annual turnover, or up to 7% for certain prohibited AI practices. While implementation timelines have evolved, the underlying questions around AI agent security, accountability, and governance aren't going away.
In this 30-minute expert debate, Andrea Di Bari, Technical Director at Axon Technologies, Marius Poskus, Cybersecurity Leader and CISO, Jon Hammant, CEO of Orbital and former AWS Specialist Lead for UK&I, and Ryan Sheldrake, EMEA CTO at Straiker, tackle the five questions that matter before you put AI agents into production in Europe.
Expect a practical conversation about EU AI Act compliance, agentic AI security, human oversight, liability, and incident reporting, with insights from experts who understand the challenges of securing and deploying AI systems.
Shipping an AI agent is really just the start. Knowing how to secure it, govern it, and account for its actions is what your team needs to be on top of today.
Join us before your agents outrun your compliance plan.
Webinar details
- Date: October 22, 2026
- Time: 11:00 AM CET
- Duration: 30 minutes
- Format: Expert debate and discussion
Meet the speakers
- Andrea Di Bari — Technical Director, Axon Technologies
- Marius Poskus — Cybersecurity Leader and CISO
- Jon Hammant — CEO, Orbital; former AWS Specialist Lead, UK&I
- Ryan Sheldrake — EMEA CTO, Straiker
The 5 Questions to Ask Before You Ship an Agent in Europe
What EU AI Act compliance actually requires when AI agents move from development into production
Your agent just called a third-party agent, which used an over-provisioned credential to pull data it should never have seen.
Now you have a security incident, a potential compliance problem, and a question nobody seems able to answer. Who's liable? The model provider, the agent vendor, or you?
And can you even reproduce what happened?
If you don't have an answer, you're not alone.
The EU AI Act was written with AI systems in mind. But AI agents are different. They call other agents, access tools, use credentials, and take actions with consequences that can be difficult to predict or reconstruct after the fact.
That makes EU AI Act compliance a whole different challenge when AI agents are running in production. Who's responsible when an agent goes off-script? What does meaningful human oversight look like when agents act autonomously? And what happens when an agent-related security incident triggers regulatory reporting obligations?
Depending on the infringement, EU AI Act penalties can reach 3% of global annual turnover, or up to 7% for certain prohibited AI practices. While implementation timelines have evolved, the underlying questions around AI agent security, accountability, and governance aren't going away.
In this 30-minute expert debate, Andrea Di Bari, Technical Director at Axon Technologies, Marius Poskus, Cybersecurity Leader and CISO, Jon Hammant, CEO of Orbital and former AWS Specialist Lead for UK&I, and Ryan Sheldrake, EMEA CTO at Straiker, tackle the five questions that matter before you put AI agents into production in Europe.
Expect a practical conversation about EU AI Act compliance, agentic AI security, human oversight, liability, and incident reporting, with insights from experts who understand the challenges of securing and deploying AI systems.
Shipping an AI agent is really just the start. Knowing how to secure it, govern it, and account for its actions is what your team needs to be on top of today.
Join us before your agents outrun your compliance plan.
Webinar details
- Date: October 22, 2026
- Time: 11:00 AM CET
- Duration: 30 minutes
- Format: Expert debate and discussion
Meet the speakers
- Andrea Di Bari — Technical Director, Axon Technologies
- Marius Poskus — Cybersecurity Leader and CISO
- Jon Hammant — CEO, Orbital; former AWS Specialist Lead, UK&I
- Ryan Sheldrake — EMEA CTO, Straiker
Speakers
