The Record Comes Home

The winding-down of VIQ Solutions’ Australian operations in 2026 has been reported as the failure of a transcription contractor. The more important story sits one layer deeper, and has gone almost unremarked: the software that records Australia’s courtrooms and parliaments, together with the transcription services above it, is now very largely foreign-owned, and some of it was Australian in origin before it was sold abroad. This paper traces how that came to be, identifies the part of the problem the current debate has missed — the ownership of the recording software itself — and sets out what a genuinely sovereign arrangement would require. It is written to inform the design of the arrangements now being built to replace what has failed.

01What is actually at stake

An accurate transcript is the mechanism by which a hearing can be revisited after it has ended. It allows an appeal court to check what a witness said, it gives a self-represented litigant the only complete account of a day that may have changed their life, and it fixes, in a form that can be examined later, what a court decided and why. Beneath the transcript sits something more fundamental. The audio recording is the primary record. The transcript is a rendering of it, and where the two differ it is the recording that governs. Control of the audio, and of the system that captures it, is therefore control of the court record at its source.

For most of the past two decades these facts have been obscured by the simple circumstance that the arrangements worked well enough to be ignored. The recording happened, the transcripts arrived, and almost no-one asked who owned the machinery that produced them or where the data went. The events of 2026 have removed that comfort, and the lesson is not the narrow one that has so far been drawn.

02How the Australian court record changed hands

The winding-down in 2026 is usually treated as the beginning of the story. It is closer to the end of one that runs back more than two decades, through the privatisation of a public function and the gradual movement of the resulting businesses, and the technology beneath them, into foreign ownership.

Court recording and transcription in Australia was once performed in-house by the state; Queensland’s work, for example, was done by its State Reporting Bureau. Across the 1990s and 2000s, jurisdiction by jurisdiction, that in-house capacity was contracted out to private firms. Two Australian companies came to lead the field that resulted.

The first was Auscript, a Brisbane firm founded in 1921 that had recorded Australian courts for close to a century. Struggling by the early 2000s, it was bought out of voluntary administration in 2004 and rebuilt into the largest court recording and transcription provider in the country. The second, and in the longer run the more significant, was For The Record, or FTR: a digital court-recording company, founded in 1993, whose software became the international standard for capturing the spoken record of a courtroom, with development operations in Brisbane and Perth. This is the part of the history most Australians would be surprised by. The software that records courtrooms in roughly two-thirds of the United States, and in the majority of courtrooms and parliaments in Australia, across dozens of countries, was an Australian creation, and for a period one of the country’s notable technology exports.

In 2013 the Brisbane holding company that owned Auscript acquired FTR as well. For several years a single Australian-owned group held both the largest transcription service in the country and the dominant court-recording software in the world. At that moment the commanding heights of the court record, in service and in software, were in Australian hands.

They did not remain there, and the transfer offshore happened in stages. In 2015 a large United States public-sector software company took a minority equity position in the Australian group. In 2020 a United States private-equity firm made a recapitalisation investment in the group that then held both FTR and Auscript, and by late 2020, as disclosed to the Queensland Parliament, Auscript’s holding company was substantively owned by that United States firm.

The other leading Australian name had gone offshore much earlier. Spark & Cannon, a long-established reporting firm, had been acquired in 2004 by a Canadian company that later became VIQ Solutions Inc., listed on the TSX Venture Exchange, and it operated as that company’s Australian arm before being rebranded to the VIQ name in 2020.

The consolidation then completed. In late 2021 VIQ Solutions acquired Auscript, bringing the two historic Australian transcription businesses under a single Canadian owner, and Auscript was itself rebranded to VIQ Solutions in 2023. The recording software, For The Record, was not part of that sale, and travelled a separate path described below.

Two points follow from this history. The first is that the apparent diversity of suppliers had become smaller than it looked: by the early 2020s the two principal names in Australian court transcription shared a single foreign owner. The second is that ownership of both the services and the software had moved offshore, so that a critical function of the state depended on companies headquartered and controlled elsewhere.

03Queensland: the failure rehearsed

Much of what happened at the Commonwealth level in 2026 had already occurred, on a smaller scale, in Queensland a decade earlier, and was documented at the time. The Queensland experience is worth setting out, because it shows the same pattern failing once before, on the public record.

Queensland’s court record was for many years produced in-house by the State Reporting Bureau, staffed by public-service reporters. In February 2013 that arrangement was ended. The state government of the day, then pursuing a program of public-sector cost reduction, outsourced court recording and transcription to a single provider, on the stated basis that the change would save the state up to $6 million a year, and the State Reporting Bureau was closed. The courts recorded the transition; the District Court of Queensland noted in its annual report the phasing out of the State Reporting Bureau and its replacement, with the teething troubles that had been expected.

The provider’s account was that the change reduced the cost of a transcript to the state and shortened delivery times. The Queensland Audit Office, examining the outsourcing in a report tabled in the state parliament in 2015–16, found a more complicated result. It concluded that the procurement had been rushed and poorly executed in order to meet imposed budget timeframes, that it had drawn limited interest from the market, and that it had a “lasting detrimental impact” on the service. The department could not demonstrate that it had obtained the best value for money, and there was no documented analysis supporting the decision to rely on a single provider at all. The savings that materialised were about half of those promised. And the model shifted cost from the state on to court users, with an estimated 119 per cent increase for litigants in the civil jurisdiction; the lower price paid by the state was achieved in significant part by transferring the cost to the people using the courts.

The audit was equally direct on quality. When the department tested transcript accuracy over a sixteen-month period, more transcripts failed to meet the contract’s minimum standard than met it, and the department was not independently verifying the provider’s own reporting of how much had been recorded. The judiciary noticed. The President of the Court of Appeal described the transcripts of the period as often poor, in language recorded in the Supreme Court’s own annual report, noting that matters were sometimes transcribed incorrectly or not at all, and that judges were having to check the transcript against the original audio when accuracy mattered to an appeal.

The sequel is the part that matters most now. When a new government took office, it had the audit, the judicial criticism and the cost complaints all pointing the same way, and it moved to change the arrangement. It did not simply restore the in-house bureau. It built a state-owned ordering and data platform, taking the ordering system and the data back into public hands, and replaced the exclusive contract with a panel of transcription providers working beneath that platform. That structural instinct, public ownership of the platform and the data with plurality of providers beneath it, is the right one, and this paper returns to it. Its limitation, examined below, is that the panel appointed was itself composed of foreign-owned subsidiaries, so the reform secured the data layer without securing the ownership of the providers.

The elements are all there: a single point of dependence, an absence of independent verification, the transfer of cost and risk on to the people the courts exist to serve, and a procurement driven by short-term savings rather than by the long-term integrity of the record. The audit that documented it was public and was tabled in a parliament, a decade before the Commonwealth’s own single-provider arrangement collapsed.

04The layer the debate has missed

The Senate Estimates hearings of May 2026 established, in unusually candid terms, that the exclusive outsourcing of transcription services to a foreign-owned provider had failed, and that the future lay in a panel of onshore providers with in-house capacity and Australian control of the data. That is right, and this paper endorses it. But the discussion, in the committee room and in the reporting that followed, concerned the transcription services. It did not reach the recording software. And the recording software is where the deepest exposure now sits.

The system that captures the audio in an Australian courtroom is, in the great majority of cases, For The Record. Queensland’s state-owned ordering platform sits above For The Record software running across 248 courtrooms; federal courtrooms and parliaments rely on the same family of products. This is not a peripheral vendor. It is the instrument that produces the primary record, the audio against which every transcript is ultimately measured.

In February 2026, in the same period that VIQ’s Australian transcription business was being wound down, For The Record was acquired by Tyler Technologies, a United States software company listed on the New York Stock Exchange, for a cash consideration of approximately US$212.5 million. For The Record is now headquartered in the United States and folded into Tyler’s courts and justice division, where its recording and speech-to-text technology is being combined with Tyler’s other court software. The Australian-invented standard for capturing the court record has passed into United States corporate ownership.

This did not feature in the sovereignty discussion at Estimates, and its absence is the central point of this paper. A jurisdiction may bring its transcription entirely in-house, employ only local staff, hold every transcript on Australian servers, and still be capturing its primary record through software owned and controlled abroad, updated on a foreign company’s schedule and increasingly bundled with that company’s artificial-intelligence products. The question that has not been asked in public is a plain one: who owns the software that records Australia’s courts, and what follows from the answer lying overseas?

05Why a panel is necessary but not sufficient

The natural response to the failure of a single foreign monopoly is to introduce several suppliers on a panel. That is the right instinct, and it is what the Federal Court has adopted for the period after 30 June 2026, and what Queensland adopted some years earlier. The Queensland experience carries a caution, though, that should temper any confidence that a panel by itself solves the problem.

When Queensland ended its exclusive contract, it built a state-owned ordering and data platform and appointed a panel of transcription providers beneath it. When the Queensland Parliament asked whether the new providers were foreign-owned, the department confirmed that both were Australian subsidiaries of larger global companies. One had a Canadian parent; the other is the Australian arm of a United States group. In ownership terms, the panel replaced a single foreign-owned provider with a small number of foreign-owned subsidiaries. It solved the single-point-of-failure problem, which was real, and it brought the platform and the data into public hands, which was a genuine advance. It did not, by itself, address the sovereignty of ownership, because on that measure the change was modest.

This is the trap to avoid as the Commonwealth panel is assembled. A panel is necessary, but a panel is not sufficient. If membership is decided on price alone, foreign-owned and offshore operators may be admitted on that basis, and the exposure of the past will reappear distributed across several names rather than concentrated in one. The risk is not hypothetical; the June 2026 Senate Estimates record itself refers to an overseas entrant already advertising in Australia for court reporters. The word that the Federal Circuit and Family Court’s chief executive accepted at Estimates as a dealbreaker, sovereign, has to mean sovereign ownership, and not merely onshore data.

06The current position, plainly stated

It is worth assembling in one place where ownership now rests, because the picture is rarely set out whole. The transcription services that served the federal courts have been wound down, after the Canadian parent declined to support the Australian business, a failure that affected a reported 600 staff and some 200 courtrooms and left the courts arranging continuity at short notice. The recording software beneath those same courts is now owned by a United States corporation. The other established national panel provider is the Australian subsidiary of a United States group owned by North American private-equity and pension capital. The historic Australian names in the field no longer exist as independent businesses, having been consolidated into the company whose Australian operations have just closed. The arrangements that came through all of this intact are the in-house ones: the High Court’s own reporting, the Victorian Government Reporting Service, and Queensland’s state-owned platform. These kept control close to home.

Set out this way, the position is stark. The infrastructure of the Australian court record, in both of its layers, is now very largely foreign-owned, and the independent domestic capacity that once existed has been consolidated away. The firms that remain genuinely Australian-owned and independent are a minority in their own country’s justice system. That is the context for any serious discussion of the future.

07What genuine sovereignty requires

Rebuilding should aim higher than replacing one foreign supplier with several. Sovereignty over the court record has a number of components, and they need to be pursued together.

Ownership, not merely residence

Data residence and Australian incorporation are not the same as Australian ownership. A meaningful panel would give real weight to genuinely Australian-owned providers, so that the entities holding the nation’s court data answer to Australian owners and Australian law, and cannot be recapitalised, sold or wound down at the decision of a foreign parent.

Plurality of real, local capacity

Resilience requires several providers that exist as independent businesses. Consolidation has left few of them, and rebuilding that capacity, together with supporting the independent Australian firms that survived it, is part of the task rather than a byproduct of it.

Corporate memory as a selection criterion

A dependable record depends on accumulated, tacit knowledge: familiarity with a jurisdiction’s members and counsel, its conventions and protocols, its suppression orders, and the specialist vocabulary of the matters heard. This knowledge lives in long-serving practitioners and the firms that retain them. It cannot be assembled quickly, and it is worn away rather than preserved by repeated changes of ownership.

Control of the software and the audio

This is the component the current debate has missed, and it needs concrete measures rather than a statement of principle. Four are available. First, where the recording layer remains foreign-owned in the near term, it should be procured as though it were regulated infrastructure, with Australian data residency for court audio, source-code escrow, guaranteed continuity of supply on the insolvency or withdrawal of a foreign parent, a binding prohibition on training artificial-intelligence models with the record, and a right to migrate. Second, and most powerfully, the courts should mandate an open, published, court-owned format for courtroom audio and its metadata, to which any capture system must be able to export. That single requirement turns a proprietary layer into a contestable one, because once the record is portable, jurisdictions can change vendors and Australian entrants can compete on equal terms. Third, court-record capture should be treated as critical infrastructure, so that foreign acquisitions of it are subject to national-interest review rather than passing unremarked. Fourth, the development of Australian-owned recording and speech-to-text tools should be actively supported, so that the country is not indefinitely dependent on foreign-owned software for the capture of its own primary record. A nation that once built the international standard for court recording is capable of owning the tools that record its own courts.

Standards and formats owned by the courts

Conventions and file formats that belong to the court, and that are open and able to be audited, keep the record consistent across a panel and prevent any single provider’s proprietary system from becoming the next concentration of risk.

Verification, not assurance

The recurring failure, from the 2015–16 Queensland audit to the events of 2026, is the acceptance of a provider’s own assurances in place of independent checking. Vetting, security accreditation and prohibitions on offshoring are worth nothing unless they are audited independently and regularly.

08Artificial intelligence, and who controls it

Artificial intelligence runs across all of this, and it is entering the court record at both layers, in the transcription services and, through the recording software, in the capture layer as well. The framing that pits human reporters against automated transcription is a distraction that suits vendors selling replacement. The real question is one of ownership and control.

There is a use of the technology that supports the making of the record, and a use that draws value from it for other purposes. Questions have been raised, and are before the Privacy Commissioner, about whether court transcripts and recordings were used to train artificial-intelligence models without the knowledge or consent of litigants or courts. As the recording software passes to a large foreign company that is combining it with its own artificial-intelligence products, the question of what is done with the Australian court record, and where, becomes more pressing rather than less.

An approach that serves the record has characteristics that follow from sovereignty. Processing should run on Australian infrastructure, or on the practitioner’s own equipment, so that court audio and text are not sent elsewhere to be handled. The proceedings of a court should never be treated as training data, and a responsible provider can give a binding undertaking to that effect. The technology belongs in the role of a rapid first-draft assistant working under a qualified reporter who reviews, corrects and certifies the result, and not in the role of the certifier. The integrity of testimony must be treated as absolute: formatting may be standardised, but spoken words are not to be altered, and where the audio is the only reliable source the tool should mark the uncertainty for a human to resolve. And the court should be able to see how any transcript was produced, and to inspect the process end to end.

Tools built to these standards already exist, and they tend to be developed by practitioners rather than by large platforms, because the people who carry professional responsibility for the record understand, from the work itself, why each of these constraints matters.

09Principles for what comes next

The following principles are offered to the courts, departments and procurement officers who will design the arrangements that replace what has failed.

1. Prefer Australian ownership, not merely Australian data residence. Ownership is the variable that determines whether the record can be sold, recapitalised or withdrawn by a foreign parent, and it should be weighted accordingly.
2. Build genuine plurality of local capacity. A panel is only as sovereign as its members, and rebuilding independent Australian providers is part of the task, not a side effect of it.
3. Pay for corporate memory. Reward longevity, vetted personnel and institutional continuity, because these protect the accuracy of the record and cannot be created at short notice.
4. Take the software layer seriously. Procure any foreign-owned recording layer as regulated infrastructure, mandate an open court-owned audio format, treat court-record capture as critical infrastructure, and support Australian-owned recording and speech-to-text development.
5. Own the standards and formats. Court-owned, open, interoperable conventions keep the record consistent and prevent a new proprietary dependency.
6. Verify rather than assume. Independent, repeated auditing of vetting, security and offshoring obligations is the safeguard whose absence has caused this failure more than once.
7. Adopt artificial intelligence on the profession’s terms. Local processing, a prohibition on training models with proceedings, human certification, the integrity of testimony, and an inspectable account of how each record was produced.

None of these is exotic, and together they describe a system more resilient, more accountable, and more genuinely Australian than the one it replaces.

Coda

The winding-down of 2026 has been told as the failure of a transcription company. It is better understood as the moment it became visible that, over more than two decades, the infrastructure of the Australian court record had passed almost entirely into foreign hands, some of it built in Australia before being sold abroad. The transcription services closed, and the recording software passed into United States ownership in the same period. The independent Australian firms that once did this work were consolidated close to out of existence, and those that remain are a minority in their own jurisdiction.

The decision to build a panel of onshore providers with in-house capacity and Australian control of the data is a sound beginning, and it deserves to be built with the whole of this history in view. The more important and more difficult work is to treat sovereignty as a matter of ownership and control at both layers of the record, the service and the software alike, and to rebuild the local capacity that makes such sovereignty possible. Australia built the tools that record the world’s courts. It is well within the country’s capacity to own the tools that record its own.