Beat the rush and join us early for complimentary barista-made coffee and breakfast.

AI has shifted from assistants that make suggestions to autonomous agents that can read files, execute commands, call APIs, and modify systems on their own. That change expands the attack surface from prompt injection to full system compromise, lateral movement between agents, and persistent access through memory and tooling.

‍

This session explores how autonomous agents are reshaping the threat model, what early adopters are discovering in practice, and the questions AppSec teams must confront as AI systems gain more autonomy and more potential for harm.

‍

The speaker will cover:

• New risks from code-executing prompt injection to agent-to-agent lateral movement
• How teams are designing permissions, audit trails, sandboxing, and monitoring agent behavior
• Examples of agents being manipulated to exfiltrate data or modify configurations
• The security shifts required to safely deploy autonomous agents in the next year

Modern applications run on layered platforms, third-party extensions, and AI assisted development and tooling, each introducing dependencies that traditional supply chain controls struggle to track. Even with SBOMs and automated scanning, teams are seeing supply chain risk surface in production through transitive packages, platform abstractions, and components that weren’t visible at build time.

‍

This panel explores how supply chain risks are actually surfacing in real environments and what effective control looks like when dependency sprawl is structural, not accidental.

‍

We'll Cover

• Where SBOMs help in practice and where they still fall short
• How layered platforms, third-party extensions, and AI era tooling introduce new blind spots
• What transitive dependencies, dormant packages, and platform abstractions mean for real-world risk
• Practical approaches for regaining control without killing development velocity

In this innovative session, attendees will be faced with a series of scenarios that they may face in their roles. Attendees will discuss the possible courses of action with their peers to consider the ramifications of each option before logging their own course of action.

‍

Results will be tallied and analysed by our session facilitator and results will impact the way the group moves through the activity.

‍

Will we collectively choose the right course of action?

The rise of AI is introducing supply chain–style risks we’ve seen before in open source. The way developers pick, use, and secure components has direct parallels to how organisations will need to think about models, training data, and AI tooling. This keynote connects the dots between securing the software supply chain and what it means to secure AI.

We’ll cover:

• Parallels to open source: How today’s AI risks echo the dependency, trust, and governance challenges seen in software supply chains.
• Where AI helps (and misleads): The promise and pitfalls of using AI to guide security choices and code fixes.
• Applying the lessons: Practical steps organisations can take from supply chain security to get ahead of AI-specific threats.

‍

AI is changing how software gets written. Code is no longer produced line by line by a single developer. It is generated, refactored, and stitched together by AI tools at a speed traditional review processes were never designed to handle. Yet many AppSec programs are still relying on the same manual reviews, static rules, and approval gates built for a pre-AI era.

This keynote explores why secure code review is breaking down as AI becomes a core part of development, where existing practices create false confidence, and what needs to change to keep risk under control without slowing teams to a crawl.

‍

The speaker will cover:

• Why AI-generated code shifts risk from individual lines to system-level behaviour
• Where traditional code review and SAST fail in high-velocity, AI-assisted pipelines
• How leading teams are redesigning review around intent, context, and ownership
• Practical ways to evolve secure code review for AI-native development in the next 12 months

Modern engineering teams move fast, and AppSec teams are constantly negotiating when to block, when to slow down, and when to accept risk to keep delivery on track. As exceptions, waivers, and temporary approvals become part of everyday workflows, many organizations struggle to understand what risks they’ve accepted, why they accepted them, and whether those decisions are still defensible months later.

‍

This panel explores how high-performing teams balance speed with security, how they document and monitor accepted risk, and the frameworks that keep fast-moving environments accountable.

‍

The panel will cover:

• How teams decide when risk acceptance is justified and when it isn’t
• Practical approaches to tracking exceptions, waivers, and approvals over time
• Techniques for documenting context so decisions remain defensible later
• How AppSec and engineering collaborate to keep velocity without losing control

Put your knowledge to the test in this fast-paced quiz covering real-world trivia, key concepts, and emerging trends. Compete for bragging rights - and a travel voucher - as the top scorer takes the crown.

‍

Most security programs have matured around AI-generated code risks, but a critical gap remains: the injection surface introduced by autonomous AI agents. Gaurav Vikash, Head of Security and Risk, Apac at Axon presents hands-on testing across Claude Desktop with MCP tooling, browser-based agents, and document processing pipelines, with demos highlighting where trust models fail and where they still hold.  

Key Takeaways:

• Understand how second-order prompt injection emerges when agents retrieve and process external content.  
• See concrete examples of where trust boundaries break across common AI agent deployments.  
• Learn how to evaluate and strengthen trust models to mitigate these emerging risks.  

‍

AppSec teams sit at the center of fast-moving engineering organisations, yet there’s still no consensus on how they should be structured, what they should own, or how much authority they should have to slow things down.

‍

This interactive session puts those debates on the screen literally. The audience votes live on five core questions covering team design, ownership boundaries, blocking power, developer experience, and how AI is reshaping the AppSec operating model. We explore the results, debate the trade-offs, then vote again to see if perspectives shift in real time.

‍

This session will cover:

• How structure and ownership shape AppSec’s influence
• When blocking authority helps or harms engineering velocity
• How AI is forcing teams to rethink traditional operating models
• What leading organisations are learning about building developer-first AppSec

‍

Unwind with your peers for a couple of drinks on us!