Unattended fuel retail changes the supervision model, but not the underlying operational, environmental and compliance responsibilities. The challenge is to replace permanent on-site presence with connected systems, governed human response and evidence that survives the incident.

By Hamed Taghvaei | Vertex Pulse

Unmanned and unattended fuel sites are becoming increasingly relevant to Australian petroleum retail. Remote payment, improved connectivity, automatic tank gauging, equipment telemetry, CCTV, environmental sensors and automated controls make it possible to operate selected sites without a permanent attendant. For some networks, the format can support lower-cost locations, extended operating hours and more flexible network design.

But removing the attendant does not remove the fuel, the hazardous materials, the customers, the vehicles, the underground infrastructure or the obligation to respond when something goes wrong. It removes one layer of observation and immediate judgement. Those functions must be deliberately reassigned.

The more useful question is therefore not simply whether a site can operate without staff. It is: how are abnormal conditions verified, prioritised, acted on and evidenced when nobody is permanently present?

The risk remains on site

A staffed forecourt is not risk-free, and human presence is not a substitute for engineering controls. Even so, an attendant can notice context that a disconnected alarm cannot: a vehicle impact, an unusual smell, unsafe behaviour, a spill, a customer in distress or equipment operating abnormally. The unmanned model therefore needs a structured way to replace observation, escalation, communication and evidence creation.

• Fuel-system risk — dispenser faults, tank-system anomalies, unexpected wet-stock variance, equipment degradation and possible loss of containment still require interpretation and follow-up.
• Environmental risk — surface spills, subsurface releases and containment failures may require investigation, escalation and evidence, even when they are not immediately visible.
• Safety risk — vehicle incidents, drive-offs, forecourt hazards, unsafe customer behaviour, fire and medical events can occur regardless of staffing model.
• Operational risk — communications loss, equipment offline conditions, payment failure, power interruption and access issues can make a nominally open site unavailable or unsafe.
• Compliance risk — scheduled inspections, emergency plans, corrective actions, records, maintenance evidence and audit readiness remain necessary.

This control principle is already reflected in Australian industry guidance. ACAPMA has highlighted that unmanned-site design and operation need active controls for prompt incident response, emergency-service contact, evacuation and on-ground response. AS 1940:2017 also addresses unsupervised self-service systems and includes an emergency shutdown function that stops the dispenser pump and transmits an alarm to a person or organisation capable of responding.

Monitoring is necessary — but monitoring alone is not the operating model

Traditional monitoring commonly follows a straightforward path:

SIGNAL  →  ALERT  →  NOTIFICATION  →  CUSTOMER DECIDES WHAT HAPPENS NEXT

The gap sits after the alert. Someone still needs to determine whether the event is genuine, understand the site context, assign priority, select the approved procedure, initiate the next action and confirm that the event has actually been closed.

More sensors can create more visibility without creating better control. At network scale, fragmented response across alarms, phone calls, email, CCTV systems, maintenance contractors and spreadsheets can become the new operational burden.

A managed-operations layer closes the gap

A practical managed-operations sequence for unmanned fuel retail is:

TRIAGE  →  VERIFY  →  PRIORITISE  →  ACT  →  EVIDENCE & CLOSE

Triage. Connected systems surface exceptions from sources such as ATG and wet-stock data, dispensers, payment systems, environmental sensors, CCTV analytics or loss of site communications. The first task is to identify the site, asset, source and potential severity.

Verify. A trained operator reviews available context. Camera views may confirm a surface incident; tank and equipment data may help establish whether an alarm is credible. CCTV is a verification input, not necessarily a bank of screens watched continuously.

Prioritise. The event is classified against a customer-approved incident matrix so that a communications outage, equipment fault, suspected release and active emergency do not receive the same response.

Act. The approved response is commenced. Depending on the event and site, this may include stakeholder notification, emergency escalation, dispatch of an authorised responder, or a remote isolation/shutdown procedure where the installed equipment, integration, customer authority and engineering controls support it.

Evidence and close. The event should end with a structured chronology: what was detected, how it was verified, who made or authorised decisions, what actions occurred, which communications were made, what evidence was retained and what final status was confirmed.

This is the distinction that makes the model scalable: connected systems provide continuity; people provide judgement. Human attention is concentrated on exceptions and consequential decisions rather than continuous observation of every site.

Where SiteAccord fits

SiteAccord was developed around a simple premise: an unmanned fuel site should not become an unmanaged fuel site.

It is designed specifically for petroleum retail to provide the managed operational layer between the physical site, connected technology, customer-approved procedures, human response and compliance evidence. Within an agreed service scope, SiteAccord is designed to support:

• exception triage and operator verification;
• incident prioritisation against agreed procedures and authority levels;
• commencement and coordination of approved response pathways;
• structured event records, communications and evidence;
• network-level operational reporting and recurring exception visibility.

SiteAccord is not intended to be unrestricted remote control, a replacement for emergency services or generic CCTV monitoring. The value is in the governed connection between detection and response.

Delivered on a petroleum compliance system of record

SiteAccord is delivered on Vertex Pulse, a petroleum-specific compliance system of record. This allows site records, approved procedures, operational communications, corrective actions and event evidence to form one permanent history rather than being dispersed across separate systems.

The resulting evidence chain is simple: event → response → evidence → permanent system-of-record history. That matters when operations, HSE, risk, legal or an auditor later needs to understand what the organisation knew, which procedure applied and what action followed.

What good looks like before a site goes unmanned

Before an unattended operating model is scaled, the network should be able to answer seven questions clearly:

• Which site systems are considered critical, and is loss of visibility itself treated as an exception?
• Which events require human verification before action?
• Are incident priorities, authority limits and response procedures agreed before the event occurs?
• Which remote actions are technically supported, tested and explicitly authorised?
• How are emergency services, field responders and maintenance contractors brought into the response?
• How are decisions, approvals, communications and evidence captured while the event is happening?
• Can head office see unresolved events, recurring failures and control gaps across the network?

Unmanned should mean differently managed — not unmanaged

The transition to unmanned fuel retail is not simply a change in labour model. It is a change in control architecture. The organisations that scale the format well will be those that deliberately connect engineering controls, site data, human judgement, response authority and permanent evidence.

Technology can provide the visibility. The operating model determines what happens next.

For more information see:  https://siteaccord.com.au

Regulatory note: Australian legal and regulatory requirements vary by state and territory and can also depend on site approvals, licence conditions and individual circumstances. Australian Standards should not be treated as legislation unless their legal application has been established for the relevant site and jurisdiction.

Technical references for editorial review: ACAPMA, “Unmanned fuel sites – getting it right” (28 November 2025); AS 1940:2017, clause 7.3.6 (unsupervised self-service systems).

RELEASE