General

Traditional approaches to cybersecurity are no longer enough. We must move beyond passive defence and actively prepare for this new era of cyberwarfare.

Red teaming, or adversary simulation, emulates real-world attackers using physical, social and technical techniques to test an organisation’s defences. Unlike standard penetration tests, these covert exercises emphasise persistence, privilege escalation and detection avoidance to reveal weaknesses across people, processes and technology.

For years, businesses have struggled to fill cybersecurity roles.The assumption that automation or AI tools would help hasn’t played out in practice. And that brings the conversation back to people - how to spot potential in places that haven’t traditionally been part of the recruitment pipeline. One such place is the gaming world.

There’s nothing inherently wrong with bots. In fact, when they’re used well, they’re one of the most powerful tools in a business’s digital toolkit — streamlining repetitive tasks, freeing up human time, and delivering consistent results. The issue isn’t the bots themselves. It’s how carelessly we’ve started deploying them.

The current market conception of agents as LLMs with function calling represents a fundamental mischaracterization of true autonomous systems. 25% of companies that use GenAI will launch agentic AI pilots or proofs of concept in 2025, yet over 40% of agentic AI projects will be canceled by the end of 2027 due to escalating costs, unclear business value, or inadequate risk controls.

"Agentic AI" is the newest thing in AI. It promises systems that can talk to people and things around them, plan, and learn. But the truth is more complicated than what all the ads say. We're creating a world where the line between human and artificial interaction quickly disappears.

We're living through the biggest shift the internet has seen since its creation. It's not just that bots exist online now. It's that they're on track to become the majority. And we're running out of reliable ways to tell who's real.

The current conversation around agentic AI has been focused on autonomy and decision-making. But the potential impact of agentic AI is bigger than that. Since agents have the ability to plan their own actions, use enterprise tools, interpret environmental cues, and refine their behavior based on feedback, they pose exponential risks that users need to consider.