It’s quite a challenge to be a CISO, in this era of AI-driven change. Organizations we're talking to know things could go south quickly. Having lived through ransomware waves, they know the next generation of vulnerabilities is ripe for exploitation — and they're not sure what to do about them. Risk assessments are fine. Best practices help. But neither quite captures the scale of what's in front of them.
Challenges are coming from all sides: externally, new threats such as prompt injection while internally, software is being built faster than ever, based on increasing levels of automation. At a recent security event, one vendor we spoke to put it well. They were in conversation with a CISO who admitted that AI applications were "coming in through the windows, never mind the doors."
Shadow AI is rife, deployments are ungoverned, tools are being adopted by development teams before anyone in security has had a chance to blink. The exposure surface has already grown — and continues to grow. So what can enterprises actually do, right now? The answer, is to address what we’re calling the burning platform of platformization.
Most enterprises we research are running fragmented security environments that were never designed to work together: point tools inherited from years of best-of-breed purchasing, bolt-on compliance layers, and acquisitions that arrived with their own integration layers and dashboards. Each was the right decision at the time, but together, they create blind spots, integration debt, and operational drag.
Platformization — consolidating to fewer, deeper, more integrated security platforms — has been a discussion point for a couple of years now; but it is fast moving from "strategic direction" to "urgent necessity." When AI is expanding the attack surface and accelerating attack sophistication, organizations cannot afford the coordination overhead of a fragmented stack. CISOs need coherent visibility, fast response times, and the efficiencies that a smaller number of tools can bring.
Time is not on any organization's side, which brings to the “burning platform” concept: the stacks upon which our security teams depend are becoming a risk in their own right, as their fragmented nature undermines responsiveness and effectiveness. So, what can CISOs do? From experience, this comes down to treating platformization as a must-do item on the budget, based on the very real threats emerging from AI.
“Must-do” is more than a clever phrase. No CFO wants to spend money on cybersecurity: it tends to appear lower down the budget, below infrastructure refresh, cloud costs and application licensing. “Do we have to do this?” is a familiar question from the money people: the answer has to be a categorical “Yes – and this is why.”
With AI threats looming on the horizon, there is little time to lose; but it won’t stop there, as quantum technologies continue to make breakthroughs (driven in part by AI research) while simultaneously increasing risk to existing encryption. While organizations cannot respond to threats until they happen, they can use this time to get their houses in order.
A first step is an audit of tools in place: surveys suggest most organizations have between 50 and 100, which is both too many, and an achievable number to catalog. Many tools will have increased in capability since they were acquired: inevitably there will be quick wins to be had in deduplication, whilst keeping security operations teams on board.
The trick is to move from security as a collection of tools, to security as an architecture. Good looks like an integrated stack based on a managed set of event data repositories, standardsization around event management and automation, and dashboards according to role. Get things right and you will save money in the short term, even as you reduce longer term risk – what CFO could ever push back on that?