Summary

  • Ofcom closed its compliance-remediation period with British Telecommunications PLC on 11 September 2026. The regulator reported changes to BT’s telecare-migration controls, but made no determination that BT had contravened a rule.
  • The public test now spans three hand-offs: identify the customer and alarm before an order, restore service after a migration fault, and evaluate whether the policy works. Ofcom says most measures are in place and will monitor effectiveness; its notice publishes no migration or alarm-test results.

Analysis

A closed file records a remedy, not a migration result

A working personal alarm depends on more than the device beside a customer’s chair. Its status must be recognised in the provider’s records; the line and alarm must remain compatible through a network change; and a failure must reach someone able to restore service. A migration can be marked complete while one of those links is still uncertain.

That is why Ofcom’s September notice deserves a close reading. The case concerned an outbound-calling campaign within BT’s wider IP-migration programme. Monitoring raised concerns about possible agent error when checking whether a customer used telecare, delays or inconsistencies when reconnecting a user after a service loss, and possible gaps in BT’s published vulnerability policy. Ofcom says BT entered compliance remediation in February 2026, a process intended to resolve concerns without formal enforcement.

The regulator closed the matter on 11 September and published its account on 14 September. It says the changes sufficiently addressed the concerns it identified. It also makes the legal boundary explicit: the agreement and Ofcom’s acceptance do not amount to a finding of non-compliance, and Ofcom did not determine whether BT had contravened a regulatory requirement. This is neither a formal breach decision nor a finding that every customer’s migration has been tested successfully.

Three controls sit on different sides of the order

The remedies cover three operational moments. Before an order progresses, BT added system alerts and agent training intended to verify whether a migrating customer uses telecare. After migration, the company strengthened welfare-restoration processes: prompts for internal escalation and service restoration, checks on telecare users after the change, and proactive review or escalation when a line fault appears. It also updated the vulnerability policy to say more clearly how its impact and effectiveness are monitored and evaluated.

These measures are related, but they are not substitutes for one another. Better identification can prevent an unsupported migration from proceeding. A post-migration check can detect a connection that did not behave as expected. A restoration path can reduce the duration of a fault. Policy evaluation can reveal whether those controls work across more than the cases agents happen to notice. A training session alone cannot prove that a customer record is accurate; a completed order cannot prove that an alarm reached its monitoring centre.

Ofcom’s General Condition C5.2 requires regulated providers to establish, publish and comply with clear, effective policies for fair treatment of consumers whose circumstances may make them vulnerable. C5.3 also addresses recording needs, contact channels, and monitoring policy effectiveness; C5.5 covers staff awareness and training. The source documents connect policy, records, people and evaluation. The migration controls make that relationship operational rather than rhetorical.

What Ofcom published—and what remains outside the record

Ofcom says BT had implemented the majority of the measures by 14 September. Final system and policy changes were due by the end of November, and the regulator said it would monitor both progress and effectiveness. It will also continue to scrutinise BT and other providers during the remaining stages of Openreach’s Public Switched Telephone Network (PSTN) switch-off.

The notice does not give the number of telecare customers screened, the share successfully reached, the number of alarms tested after migration, fault or restoration times, or a sampled error rate. That is a limit of the public notice; it does not show that BT or Ofcom hold no such evidence. Nor does the notice establish that a failed alarm occurred in this BT case. Its account concerns risks and process weaknesses identified through monitoring, followed by remedial measures and a continuing commitment to oversight.

The distinction matters because “most measures implemented” is an input status. Effectiveness is an outcome question. For a user who relies on an alarm, the meaningful result is whether a compatible, functioning route to help remains available after the change—not whether a prompt was added to a screen or a policy paragraph revised.

A comparator, not a shortcut

Ofcom’s January 2026 decision on Virgin Media shows why the migration safeguards matter. In a separate investigation covering 23 August 2022 to 18 December 2023, Ofcom found that Virgin Media had failed to comply with its own vulnerability policies under C5.2. Its findings included weaknesses in identifying and recording telecare customers and the disconnection of about 3,500 known users; Ofcom imposed a £23.8 million penalty.

That decision is context for the risk and for one formal enforcement route. It is not evidence about BT. The provider, period, evidence and legal outcome differ. Treating Virgin Media’s findings as proof of BT misconduct would erase the very distinction the September notice draws. Treating BT’s remediation closure as a clean bill of health would make the opposite mistake.

The transition has more than one owner

The UK government’s Telecare National Action Plan describes a target outcome: a known telecare user should not move to digital service without confirmation that a compatible and functioning solution is in place. It estimated about two million telecare users in the UK in its 2025 publication and placed the broader landline migration on a January 2027 horizon. Those are a time-bound estimate and programme expectation, not a current census or a BT-specific deadline.

A March 2026 checklist for non-voluntary fixed-network migrations adds practical steps such as pursuing local-authority data-sharing agreements, checking for calls to alarm-receiving centres, offering support and prioritising restoration where technically feasible. It also advises beginning with smaller cohorts. These documents describe a multi-party safety framework, not a new finding against BT.

Local authorities and alarm providers can help identify users and validate equipment; customers and advocates can expose gaps the provider’s records miss. That participation supplies evidence and warnings. It does not transfer responsibility for BT’s systems to the customer, or regulatory judgment from Ofcom to a stakeholder forum. Heng Lu’s Note 73 makes the same institutional distinction between input and mandate. Here it is an editorial lens, not a legal source: BT operates the controls, while Ofcom must explain what its oversight establishes and what remains to be checked.

Sources