OT Security Blind Spot: Where IT Security Stops
Most IT security stops at the IT/OT firewall. See how attackers reach the plant floor, what the NCSC now advises, and a practical 90-day plan to close the gap.
When IT Security Stops at the Firewall: The OT Blind Spot
The NCSC's Annual Review 2026 describes a cyber incident at a major UK manufacturer in late 2025 that caused severe and prolonged disruption to production, with around £196 million in direct cyber-related costs. The knock-on effect was big enough to help hold UK GDP growth to 0.1% in the third quarter of 2025. Further down the supply chain, smaller firms that relied on its production schedules, orders and payments dealt with operational problems they had no part in causing.
Most security teams would read that and ask one question first: did the attackers get into the plant?
It's an understandable question, but it misses the point. A factory doesn't need its controllers hacked to stop production. It only needs the business systems it depends on to go dark. Our security programmes rarely see that gap, because most of them were drawn to end at the firewall between the corporate network and the plant floor. We call that gap the OT security blind spot, and it's where a growing share of real operational damage now happens.
So What Is the OT Blind Spot, Exactly?
It's everything on the operational side of the business that your IT security programme assumes is protected but has never tested, inventoried or monitored. In most organisations, that starts right at the IT/OT firewall. Above that line, there are pen tests, vulnerability scans, SOC alerts and board reports. Below it, the usual answer is "that's operations."
Operational technology is the hardware and software that detects or causes a change in physical equipment and processes: the PLCs, HMIs, SCADA systems and safety controllers that keep production lines, pumps and substations running. When IT fails, you lose data or access. When OT fails, things stop moving, and sometimes they move in ways they shouldn't.
You Don't Have to Touch a PLC to Stop a Plant
A modern factory leans on IT far more than its network diagram suggests. Production schedules come from ERP. Work orders flow through MES. Engineers log in with Active Directory accounts. Logistics, quality and supplier portals all sit on the corporate side of the firewall.
When those systems go down, operations teams face an awkward choice. Either they keep running without the data they need, or they stop until someone can confirm the plant is clean. Most choose to stop. That's the right call for safety, but it means a ransomware attack that never crossed the firewall can still halt production for weeks.
The damage rarely stays inside one company either. When an ERP system cannot issue payments or accept deliveries, failures cascade to suppliers, distributors and customers. Attackers have worked this out. A stopped line puts far more pressure on a business to pay than stolen files ever did.
And Now They're Going After the Controllers Too
For years, direct attacks on industrial controllers were treated as rare, state-level events. That's changing.
On 27 August 2026, the NCSC reported "increased targeting of operational technology (OT) systems across multiple sectors globally, including in the UK", with "some limited real-world disruption." The advice that followed included a line every operations director should read twice: organisations "should not assume that their OT is inaccessible from the internet without verifying it."
The state-backed threat is more serious still. The NCSC's annual review notes that an attack on Poland's energy grid in December 2025 was formally attributed to Russia's FSB Centre 16. Had it worked, it "could have caused 500,000 citizens to lose electricity."
AI Is Making Old Shortcuts Expensive
Every OT environment carries a few compromises that made sense at the time. An HMI still running an unsupported operating system because the vendor won't certify an upgrade. One engineering password shared across three shifts and two contractors. A firewall rule added for a commissioning project in 2019 that nobody has touched since.
Those shortcuts used to be fairly safe, because finding and exploiting them took time and skill. That's no longer true. The NCSC now judges it highly likely that AI-powered malware will be used in a ransomware attack on UK critical national infrastructure by 2028. NCSC chief executive Richard Horne warned that vulnerabilities "that have been tolerated for far too long can now be discovered and exploited" at scale. In OT, where those weaknesses are often left in place for years, that shortens the time between a gap existing and someone using it. We looked at what this means for leadership teams in our breakdown of the NCSC Annual Review 2026 and AI ransomware risk for UK boards.
How an Attacker Gets From a Laptop to the Production Line
Very few OT incidents start in OT. They start the same way IT incidents do, and then they follow the paths your firewall was set up to allow.
| Step | What the attacker does | Why it often goes unnoticed |
| --- | --- | --- |
| Get a foothold | Phishes a user or talks the service desk into resetting a password | Spotted and handled as a routine IT incident |
| Borrow an identity | Uses Active Directory that's shared between IT and OT | Nobody has mapped which IT accounts can log in to OT |
| Cross the boundary | Moves through a historian, jump host or old firewall exception | The traffic is allowed, so nothing alerts |
| Reach engineering | Takes over a workstation with PLC programming software | The SOC can't see below the DMZ |
| Cause the damage | Encrypts HMIs and servers, or changes controller logic | Operators notice before security does |
Look at the last column. After the first step, almost everything happens somewhere the security team isn't looking. So the useful question after an incident isn't "did the firewall hold?" It's "which routes through it did we never test?"
What We Keep Finding Behind the Firewall
The NCSC's August recommendations line up closely with what OT assessments tend to uncover. A few issues come up again and again.
The first is internet access nobody knew about. Cellular routers tucked inside control cabinets, OEM support modems and port forwards set up for a remote fix years ago all connect OT to the outside world without passing through the corporate firewall. They don't show up in IT's external scans because IT doesn't know they exist.
The second is credentials. Shared engineering logins and default passwords on HMIs are still common, which is why the NCSC now explicitly calls for individual admin accounts and MFA wherever the equipment supports it.
Then there are controllers left in program mode. A PLC left in program or remote mode can have its logic changed over the network. The NCSC recommends keeping controllers out of that mode during normal running and write-protecting their logic. It's a simple control, but it rarely appears in IT security standards, so it rarely gets checked.
Industrial protocols are another weak point. Modbus and older versions of DNP3 were designed for closed networks and trust every command they receive. Moving to secure versions such as DNP3-SAv5, CIP Security or OPC UA takes time, and legacy hardware often can't support them at all, a problem we explored in how legacy PLCs are becoming modern security liabilities.
Finally, backups. Most organisations back up their servers. Far fewer can restore a PLC's logic, configuration and engineering project files, and fewer still have ever tried.
Why Your IT Incident Playbook Can Make Things Worse in a Plant
When an IT incident hits, the instincts are well rehearsed: isolate the machine, patch the hole, wipe and rebuild. On a plant floor, each of those can backfire.
Pulling the network cable on the wrong segment can leave a process in an unsafe state or trip a safety system. Emergency patching may need a planned shutdown and sign-off from the equipment vendor. An aggressive vulnerability scan, run in a hurry to "see how bad it is", can crash older controllers outright. And the rebuild assumes you have clean copies of everything, including controller logic that may never have been backed up.
None of this means OT incidents can't be handled well. It means they need a playbook written jointly by security and operations, before anyone needs it. Our look at what recent manufacturing incidents teach us about OT resilience goes further into why recovering a plant takes longer than recovering an office.
Regulators Have Stopped Accepting "That's Operations"
The rules are moving in the same direction. NIS2 brings manufacturing, energy, transport and water operators firmly into scope across the EU. In the UK, operators of essential services already measure themselves against the NCSC's Cyber Assessment Framework, and the forthcoming Cyber Security and Resilience Bill will widen the net to more organisations and their suppliers.
What these frameworks share is an expectation that you manage risk across everything that delivers the service, not just the corporate network. If a regulator asks which OT assets support a critical service and how you protect them, "operations owns that" won't be an acceptable answer. Our NIS2 compliance roadmap covers how OT fits into that work.
Where to Start: The First 90 Days
Closing the blind spot doesn't need a three-year programme to get going. It needs honest evidence about what's really out there, gathered without putting production at risk.
In the first month, focus on finding out what's connected. Passive monitoring can build an inventory of PLCs, HMIs, firmware versions and communication paths just by listening to network traffic, so nothing on the line is touched. Asset discovery is the foundation of OT security, and it nearly always turns up connections nobody had written down. At the same time, check from the outside whether any OT device can be reached from the internet. Don't take anyone's word for it.
In the second month, test the boundary the way an attacker would. Start from a compromised corporate laptop and see how far you get: shared accounts, jump hosts, historians, old firewall rules. List every route a vendor uses for remote access, and close any that nobody can justify. Then tighten what's left using network segmentation to reduce operational cyber risk, with IEC 62443 zones and conduits as the framework.
In the third month, give your SOC sight of the plant. That means monitoring tools that understand industrial protocols, a baseline of what normal traffic looks like, and alerts that tell analysts which physical process is affected, not just which IP address. Finish by agreeing a joint incident playbook: who can authorise isolating an OT segment, how a safe shutdown works, and what has to be checked before production restarts.
Five Questions Worth Asking This Week
Put these to your security and operations leads together. If the honest answer to any of them is "we're not sure", you've found your blind spot.
- Can any of our PLCs or HMIs be reached from the internet, and how do we know?
- Which IT accounts can log in to OT systems today?
- When did someone last try to break through the IT/OT boundary, rather than just review the rules?
- Does our SOC get any data at all from our OT networks?
- If a controller's logic was wiped tomorrow, could we restore it, and when did we last prove that?
The Firewall Was Never the Finish Line
A firewall is one control among many. It was never meant to mark the point where security responsibility ends. The organisations hit hardest over the past year weren't short of security tools. What they lacked was a clear view of how a problem in IT turns into a stopped production line.
At Cyber Advisory Services, we help security and operations teams build that view together, without disrupting the plant. Our OT security services include passive asset discovery, IEC 62443-aligned segmentation, OT-safe vulnerability assessment and continuous monitoring of industrial networks. They work alongside our managed detection and response and incident response teams, so containment and recovery account for how a plant actually runs. If you'd like to know what sits beyond your firewall, talk to our OT security team.