Many breaches exploit known flaws: the reason why is organisational

Tanium

By Tim Morris, Chief Security Advisor at Tanium
Tuesday, 17 November, 2026


Many breaches exploit known flaws: the reason why is organisational

The volume of vulnerabilities teams must assess and patch is climbing fast. According to Zero Day Clock, a public scoreboard for vulnerability management and exploitation, at the time of writing, new CVE records published in the last three complete months are running at 315% higher than the same period last year. With pressure on security teams increasing by the day, the conundrum becomes knowing what matters and patching it quickly to reduce exposure, while avoiding rushed changes that could break the system or fail to close the patching loop.

Issuing a patch and knowing it was applied are different things. In October 2025, Victoria’s Department of Education told its schools to apply a critical patch. The Office of the Victorian Information Commissioner later found that not all schools did, and that the department had no adequate process to confirm which ones had. A job can finish and the ticket can close while the machine is still exposed; knowing means checking that the problem the patch was meant to fix is actually gone. Machines that miss the update, fail to restart or never check in stay exposed while the report says the job is complete.

Closing the loop means confirming the fix on the endpoint in real time, then reporting that result back to the team that owns the risk.

Having teams that don’t work seamlessly together, without a shared real-time view of their endpoints, often results in manual reconciliation, missed windows and known flaws left open for weeks. This is no longer enough to protect an organisation; while most security and technology leaders now acknowledge this as fact, this recognition isn’t turning into tangible action.

Automating chaos

Many executives understand the productivity benefits of automating existing, valuable and manual processes that are taking up too much time or resources. They assume that automation only applies to things that are already working, and that you can’t automate chaos. Unfortunately, they’re wrong: you can automate chaos, but you just create more chaos, faster. Before automating anything, ask in order: can we eliminate this step, can we simplify it and can we standardise it. Only then can we automate it. Automating a process nobody has questioned carries its flaws forward at machine speed.

Businesses don’t need to look far beyond their own security and operations teams to see how this could get out of hand. Security is measured on risk, while operations is judged on stability. The two teams are handed different tools and different versions of the truth, and too often the result is a misalignment that can create complexity and increase risks. The industry is well-versed in the security half of this fight, yet the actual fracture is rarely addressed: two teams, two data sets, and one argument that never ends. Organisations need to work towards aligning teams to simplify and streamline processes.

It’s common to see organisations set out on their automation projects only to realise this process is broken, with two teams pitted against each other with separate budgets and KPIs for separate tools and headcounts that don’t communicate seamlessly. When moving to autonomous IT, organisations have to acknowledge the disconnect, audit all tools and asset inventory and connect security and operations so they are talking the same language and working towards the same goal. For example, I have seen a peer review checkbox that nobody was qualified to tick contribute to a major Linux server outage. The review existed as a checkbox, not a check. Automate that process, and the gap repeats every time it runs.

In order to trust the automation, organisations need to trust the data first. This does not require removing every silo: it requires one shared, real-time picture that both teams trust, and someone accountable for the space between them. Good cybersecurity is IT Ops done well.

Stop overcomplicating it

Automation follows instructions. Autonomy pursues outcomes. People should manage the problem, the goal and the outcome, which is the what and the why. Agents should manage the solution, which is the how. When people try to manage the solution themselves, they get stuck in the weeds, which is why it’s important to work towards outcomes, not step-by-step instructions. The real gains will happen when we use AI to reach clear business objectives, rather than ingest problem-based prompts.

Take two ways of asking an autonomous platform to deal with the same problem. The first is a set of instructions: run this query on every server in the finance group, restart the agent service on any host where it has stopped, then export the list to a spreadsheet. The second is an outcome: every server in the finance group should be reporting and running the current agent, without breaching the change freeze. The first tells the platform how to work. The second tells it what good looks like and what it must not break, and leaves the steps to the platform. When the first unexpected host turns up, the instructions have nothing to say: the outcome still applies.

While it takes time to learn how to get the best out of AI, we can’t limit our workforces to the old mantra of “if it’s not done right, do it yourself”, because it’s no longer scalable to do so. Unlike a human, the training is cumulative, easier and more durable. If done right, organisations will begin to see the long-term benefits of automation.

Looking inward at the organisational structure and AI training, processes cannot be skipped on the way to automation. When security and operations share the same live picture, remediation stops being a negotiation and becomes a workflow. Similarly, when teams know how to get the best out of AI tools, the work moves from reconciliation to outcomes.

Image credit: iStock.com/da-kuk

Related Articles

AI concerns shouldn't extend to the enterprise agent

The safety concerns around AI aren't without merit, but there are practical reasons for...

Identity blind spots put business security investments to the test

Identity management tools designed primarily around employees can struggle to provide a complete...

The non-human face of security threats: why identity has become critical infrastructure

As AI becomes more autonomous, identity governance must account for what those systems can...


  • All content Copyright © 2026 Westwick-Farrow Pty Ltd