Summary

  • Between November 2023 and September 2026, Australia's emergency-calling accountability passed through four instruments: an enforceable undertaking and a $1.5 million penalty aimed at Telstra, a regulator report on failures involving three carriers, industry-wide resilience rules, and a Senate committee recommendation to move the Emergency Call Person function to a statutory authority.
  • The instrument texts themselves were not obtained in this research, so what each obligation actually required, from what date, and with what verification remains unverified; the documented sequence is reported here, and the gaps are stated as evidence boundaries rather than filled in.

A question that keeps outliving the answer

Every emergency-calling failure produces the same public question: who is accountable for a call that never reached a police, fire or ambulance service? The answer given by a regulator is usually an instrument — an undertaking, a penalty, a rule. The answer given by a legislature, eventually, is a structure. Australia's public record between late 2023 and late 2026 shows both answers being tried in sequence, on the same function, within three years.

That sequence is the subject of this article. It is not an account of what went wrong inside a network on any particular day, because the incident-level detail — dates, durations, call volumes, affected locations — was not obtained in this research. It is an account of what the public record shows institutions doing about emergency calling, in what order, and what that order implies about whether operator-level enforcement can settle a question that outlives every individual operator.

The four instruments, in the order they appeared

The documented arc has four dated steps.

November 2023: an undertaking, and a penalty for what came before it. The Australian Communications and Media Authority secured an enforceable undertaking from Telstra under which the company committed to improve its emergency call services following failures in that service (ACMA, November 2023). In the same period the regulator required Telstra to pay a $1.5 million penalty in connection with Triple Zero outages, and its investigation found that the company had failed to comply with emergency call service rules (ACMA enforcement record).

The ordering matters, and it is the detail most likely to be lost in retelling. The outages to which the penalty related occurred before the undertaking was given. That forecloses the tidy narrative in which a company promised to fix something and then failed to honour the promise. What the record actually supports is narrower and more interesting: the regulator had a set of past failures, and it chose a forward-looking, negotiated instrument alongside a monetary penalty, rather than a single enforcement route.

May 2024: the unit of analysis widens from one firm to the industry. The ACMA published a report on emergency call service failures that examined incidents affecting Optus, Telstra and TPG (ACMA, May 2024). The significance is structural rather than numerical. Once three separately owned carriers appear inside one regulator's failure report, the failure class can no longer be described as a defect of one operator's engineering culture. It becomes a property of how emergency calls are carried, handed over and verified across an industry — and an undertaking addressed to one company cannot, by its own terms, bind the others.

June 2024: rules replace undertakings as the instrument of choice. The ACMA introduced new industry rules intended to strengthen the reliability and resilience of emergency call services, applying across carriers rather than to Telstra alone (ACMA, June 2024). This is the pivot point in the sequence. An enforceable undertaking is a bilateral instrument: it binds the party that signs it. A rule binds everyone who provides the service. Moving from one to the other is an implicit admission that the earlier form could not reach far enough.

September 2026: a committee proposes to move the function itself. On 21 September 2026 the Senate Environment and Communications References Committee handed down its final report on failures of the Triple Zero emergency call service, a report described consistently across three news organisations (ABC News; news.com.au; The Australian). The committee recommended that the Australian Government consider establishing a statutory authority to assume responsibility for the Emergency Call Person function — a change that would remove Telstra as custodian of the Triple Zero network (DataCenterDynamics report).

Read carefully, that is a recommendation to consider a structural transfer. It is not a decision, not a bill, and not a transfer that has happened. Reporting on the committee's findings also describes recommendations to mandate mobile roaming for emergency calls following an Optus outage, and a set of eleven recommendations covering modernised legislation and reliability standards (news.com.au; DataCenterDynamics report). The inquiry itself is reported as having been established after a major Optus network outage in September 2025, and as having gone on to examine further issues including Telstra network outages (ABC News). Reporting on the committee's findings further describes Telstra as having held the Emergency Call Person role since 1961, a claim carried by a single source and repeated here as attribution rather than as established institutional fact (DataCenterDynamics report).

What each instrument can and cannot do

The four steps are not four attempts at the same thing. They are four different theories of change, and distinguishing them is the practical value of the sequence.

A monetary penalty prices a past failure. It creates a financial consequence and a public record, and it does nothing directly to change the configuration of a network. Its deterrent value depends on the penalty being large enough to alter a capital-allocation decision, which a $1.5 million figure against a national carrier's network budget plainly is not. What it does buy is an authoritative finding that rules were not complied with — the fact that makes everything after it possible.

An enforceable undertaking buys a forward programme. It is a negotiated instrument, and its strength is entirely a function of its terms: what is promised, by when, who verifies it, and what happens if a milestone is missed. This research did not obtain the undertaking's text, so none of those four questions can be answered here. That is the single largest gap in this article, and it is stated rather than papered over.

Industry rules bind competitors symmetrically. They are the appropriate instrument when the failure mechanism is a shared practice rather than a single firm's error, and they are the appropriate instrument when a regulator has concluded that bilateral negotiation produced a result it cannot generalise.

A structural transfer changes who bears the duty. It is the heaviest instrument available, because it concedes that the arrangement — not the incumbent's performance under the arrangement — is the problem. A committee recommending that a government consider moving the Emergency Call Person function into a statutory authority is making exactly that concession, conditionally.

The control surface: who holds the emergency line

The control surface here is not a network element. It is the allocation of the Emergency Call Person function itself, together with the instruments that condition it. Whoever holds that function determines how an emergency call is accepted, how it is routed, and how it is handed to a response agency. The instruments attached to the function — an undertaking, a penalty, a rule, a statutory transfer — are the only recorded means by which a regulator or a legislature can compel or verify a change in that behaviour.

That framing exposes a circularity worth naming. Where the custodian of the emergency line is also the regulated party, a large part of the assurance about emergency-call performance is evidence the regulated party produces about itself. A regulator can audit, investigate and penalise, but the underlying telemetry — how many calls arrived, how many were answered, how many were handed over, how quickly — is generated inside the operator's systems. Nothing in the public record reviewed here suggests misconduct in that reporting.

The point is structural: an assurance model that depends on self-generated evidence has a weaker failure-detection property than one that does not, and the weakness is invisible precisely when performance is good.

Why widening happened

The sequence is best read as an accumulating judgement that the failure class outlived the remedy.

If a single carrier's configuration caused a national emergency-calling failure, then an undertaking plus a penalty is a proportionate response: it addresses the party, records the finding, and buys a remediation programme. If a second and third carrier then appear in the same regulator's failure report, the proportionality argument weakens, because the remedy's scope no longer matches the problem's scope. If industry-wide rules follow, the regulator has effectively said so in the only language it has.

And if a parliamentary committee then recommends moving the function out of the incumbent entirely, the argument has moved from conduct to architecture.

There is a second, less flattering reading that the record cannot exclude. Structural recommendations are cheap to make and expensive to implement; a committee recommendation to consider a statutory authority may reflect genuine engineering judgement, or it may reflect the political utility of being seen to act decisively after repeated failures. This article cannot distinguish those motives, and does not try. What it can say is that the recommendation exists, is dated, and is conditional.

What the record does not establish

BTW's standard here is that an evidence boundary is a finding, not a gap to be smoothed over. Seven boundaries apply to this article, and a reader should treat each as load-bearing.

No ACMA document was read verbatim in this research. Every regulator source arrived as a provider summary or claim candidate, with no publisher-supplied publication timestamp, so the regulator's own wording is paraphrased throughout and dates are taken from document paths rather than from document text. The enforceable undertaking's operative obligations, commencement date, duration, reporting requirements and any independent review or audit condition were not obtained. No incident-level dates, durations, affected call volumes or affected-location detail for any Triple Zero failure were obtained.

No Telstra statement, submission or incident report was read, so the operator's own account is absent from this article except where a regulator summary describes it. The Senate committee report itself was not read; its findings are known through three news reports and one provider-reported excerpt. The basis of the $1.5 million penalty, any infringement notice or court filing, and any subsequent variation or withdrawal of the undertaking were not obtained. And no independent network measurement or technical assessment of any emergency-calling failure was obtained.

The practical consequence is that this article can establish sequence, instrument type and stated purpose. It cannot establish obligation-level detail, causation, or the durability of any repair. Readers who need those should treat the underlying instruments — the undertaking, the May 2024 report, the June 2024 rules and the committee report — as the primary record and this article as an index to them.

What the sequence implies for the people who carry the duty

For a carrier's board, the lesson is uncomfortable and specific: the instrument that follows a failure is chosen by someone else, and the choice escalates. A company that resolves an incident internally and communicates nothing should expect the next instrument to be wider than the last one — from undertaking, to rule, to a recommendation that the function leave the company altogether. The cheapest moment to influence that trajectory is before the first instrument exists.

For a regulator, the sequence is a reminder that a bilateral instrument has a scope problem built in. An undertaking can be an efficient way to buy a remediation programme from a cooperative incumbent. It cannot, by construction, bind the second and third carriers who appear in the next failure report. If a regulator expects recurrence across an industry, the rule is the honest instrument from the start.

For the institutions that depend on emergency calling — hospitals, transport operators, municipal services, payment systems — the relevant fact is that the function's custodian has changed hands in policy discussion at least once, and that a transfer, if it ever happens, will be a multi-year engineering programme rather than a signing ceremony. Continuity planning that assumes a stable custodian should carry an explicit assumption about how long a transition would take.

The question to keep asking

The most useful question after any emergency-calling failure is not how long the outage lasted. It is: what did the last instrument oblige, was it verified by anyone other than the party it bound, and did the failure class recur afterwards? On the public record assembled here, the first part cannot yet be answered for Australia's November 2023 undertaking, the second is unverified, and the third appears to have recurred across multiple carriers — which is precisely why the instruments kept widening.

That is an uncomfortable place to leave a reader, and it is the honest one. A sequence of four escalating instruments is not proof that any of them failed. It is evidence that the people with the authority to choose instruments kept concluding that the previous one was too narrow.

Sources