Treatment Plants: Mitigating Insider Threats With Autonomous Access Control

In the opening pages of Die Ressource, Dr. Raphael Nagel argues that water is not an environmental footnote to the story of power but its oldest strategic ground. The archaeology he cites is blunt: the earliest legal codes of Ur-Nammu and Hammurabi regulated canals, not gold. Civilisations rose and fell on the organisation of water, and the administration of water produced the administrative state. That argument is usually read at the scale of rivers and treaties. It applies with equal force at the scale of a single treatment plant. A chemical dosing skid, a SCADA terminal, an access door to a filtration gallery: each is a node where the abstract question of water sovereignty becomes a concrete question of who is present, at what hour, with what authorisation, and with what record. This essay, written from the operational perspective of Quarero Robotics, examines how autonomous security robotics can reduce insider threat in European water treatment facilities while remaining consistent with the reporting and audit obligations of NIS2.

The Insider Problem in Water Treatment

Water treatment operates on a narrow margin of trust. A small number of operators, often working in shifts that stretch across nights and weekends, hold physical and logical access to systems whose misuse can affect public health within hours. The insider threat is not hypothetical. It spans the careless, the coerced, and the deliberate: a technician who bypasses a dosing interlock to end a shift earlier, a contractor who connects an unauthorised laptop to an engineering workstation, a disgruntled employee who alters a setpoint on a chlorination loop. None of these scenarios require an external adversary. Each is compatible with a clean perimeter and a valid badge.

Dr. Nagel makes the point that water infrastructure is concentrated and has few redundancies. A plant typically depends on a small set of dosing skids, a single SCADA architecture, and a limited number of trained operators. Concentration is efficient in normal times and brittle under stress. It also means that the damage a single insider can cause is disproportionate to the effort required. Mitigating that asymmetry is not primarily a question of culture or vetting, both of which remain necessary. It is a question of whether the plant produces a continuous, independent record of who did what, where, and when, and whether that record is generated by something other than the operator it is meant to observe.

Consistent Rounds Without Single-Operator Dependence

The traditional answer to insider risk at treatment sites has been the patrol. A guard or a designated operator walks a route, checks doors, inspects equipment, and signs a sheet. The weakness of this approach is structural. The patroller is often the same person who holds operational access to the systems being patrolled. The record is produced by the observer. On a quiet night, rounds contract. On a busy night, they are skipped. Over years, the gap between the written procedure and the actual behaviour widens until an incident reveals it.

Autonomous ground robots from Quarero Robotics execute rounds on a fixed cadence that does not contract under workload. The route is defined in the mission plan, the timing is logged to the second, and each waypoint produces a verifiable artefact: a thermal image of a dosing pump, a reading from a door contact, a visual confirmation that a SCADA cabinet is closed and sealed. The robot does not replace the operator. It removes the dependency on a single operator to be simultaneously the actor and the witness. In a facility where two or three people may be on site overnight, that separation is the difference between an audit trail and an assertion.

Attendance, Tamper Evidence, and the SCADA Perimeter

The most sensitive rooms in a treatment plant are rarely the largest. They are the control room, the chemical storage area, and the network cabinet that aggregates field signals into the SCADA environment. Autonomous access control at these points means more than reading a badge. It means correlating the badge event with a visual record of who passed through the door, whether they were alone, whether they carried equipment that was not declared, and whether the subsequent activity on the terminal inside the room matches the stated purpose of the visit.

Quarero Robotics deploys platforms that integrate with existing access control and industrial monitoring systems rather than replacing them. The robot registers its own presence at each checkpoint, which produces an attendance log that is independent of the human shift record. When a SCADA cabinet is opened outside the maintenance window, the platform captures the event, timestamps it against the work order system, and escalates if no authorisation is on file. Tamper evidence is not produced by hope. It is produced by the intersection of several independent records that an insider would have to corrupt simultaneously to conceal an action.

Aligning With NIS2 Incident Reporting and Audit Trails

The revised Network and Information Security Directive treats drinking water and waste water as essential sectors and imposes concrete obligations on operators: risk management measures, incident reporting within defined windows, and the ability to demonstrate the effectiveness of controls to competent authorities. These obligations are difficult to meet with documentation alone. They require evidence that is contemporaneous, machine readable, and resistant to retrospective editing. A shift log written the morning after does not satisfy the standard that NIS2 implicitly demands.

Autonomous robotics produce exactly the kind of evidence the directive presupposes. Every patrol generates a structured record. Every anomaly generates an event with a timestamp, a location, and a sensor payload. Every access to a controlled area is cross referenced with the operational schedule. When an incident must be reported, the operator does not reconstruct what happened from fragmented sources. The reconstruction is already written. Quarero Robotics designs its platforms so that the audit trail is the default output of the operation, not a separate reporting exercise bolted on after the fact. That orientation reduces reporting latency and, more importantly, reduces the temptation to smooth the record before it leaves the site.

From Surveillance to Sovereignty at the Facility Level

Dr. Nagel writes that a state, a company, or an asset that cannot answer its water question sovereignly will eventually be unable to answer any other question sovereignly. At the level of a single treatment plant, the water question takes a specific form. It is the question of whether the operator of the facility can state, at any given moment and with evidence, who is present, what they are doing, and what the system is doing in response. If that question cannot be answered without relying on the people whose conduct is in question, the facility is not sovereign in any operational sense. It is trusting, which is a different posture and a weaker one.

The contribution of autonomous security robotics is not to eliminate trust. Operators remain indispensable, and the skill of a competent plant team is not replaceable by any platform. The contribution is to give trust a verifiable substrate. When rounds are consistent, when access is logged by an independent witness, when anomalies are captured before they are explained away, the facility acquires the quiet, unspectacular self sufficiency that Dr. Nagel associates with sovereignty. That posture does not advertise itself. It shows up in the absence of incidents that would otherwise have happened, and in the speed and clarity with which the rare incident is reported, investigated, and closed.

The water sector in Europe is moving into a period in which the assumptions of the last two generations will not carry the next two. Infrastructure is ageing, regulatory expectations are rising, and the threat surface now includes insiders who understand the systems they are positioned to harm. None of this is solved by a single technology. It is addressed by a layered operational architecture in which human judgement, procedural discipline, and autonomous platforms each do the part of the work they are best suited to. Quarero Robotics approaches the treatment plant as one node in the broader water sovereignty question that Dr. Nagel describes, and designs its systems to produce the kind of evidence, consistency, and independence that insider risk specifically requires. The objective is not a surveilled facility. The objective is a facility whose operators, regulators, and citizens can say with precision what happened last night, and whose record of that night does not depend on the goodwill of any single person who was on shift.