Dr Kola Tytler (MBBS MBA MRCGP) | 15 July 2026 | 12 min read
Health Service Journal reported on 15 July 2026 that Heidi Health has been selected as sole supplier for a Midlands NHS ambient voice technology (AVT) framework, with reporting suggesting the arrangement could eventually cover around 70,000 clinicians across 15 NHS trusts. That is a significant regional development, and it deserves to be read carefully rather than compressed into a headline about 70,000 clinicians suddenly using an AI scribe.
The announcement, stated precisely
A framework is a procurement vehicle. It authorises participating organisations to buy a named product on agreed terms, at agreed prices, without each trust running its own tender. It is not, by itself, evidence that every eligible clinician is using the product, or even that every trust in scope has decided to deploy it yet.
So there are three separate things worth distinguishing here:
- A framework has reportedly been established, with Heidi as sole supplier.
- Individual trusts within the region will still need to choose to activate it locally.
- Individual clinicians will still need to be onboarded, trained and, in practice, decide to use it consultation by consultation.
The 70,000 figure describes the workforce that the framework could theoretically make Heidi available to, not a confirmed current user count. Until trust-by-trust deployment data is published, the accurate description is that the framework makes ambient scribing available to, or is intended to support, up to that number of clinicians, not that 70,000 people are already dictating notes through Heidi. This distinction is not pedantry. NHS technology reporting has a long history of conflating contract scope with actual usage, and the gap between the two is often where the real story sits eighteen months later.
What is publicly known, and what still needs clarifying
What appears established from the reporting: the region, the number of participating trusts (15), the sole-supplier status, and the scale of the potential workforce reach. What remains genuinely unclear at the time of writing includes the contract's duration and total value, whether licences are centrally funded or drawn from individual trust budgets, whether participation is mandatory or opt-in for each trust, whether general practice and community providers sit inside or outside the framework's scope, and whether the rollout spans acute, mental health, community and ambulance settings equally, or is weighted towards specific care settings first.
There is also a set of questions that tend to matter more in practice than in the initial press coverage: who owns the relationship with each trust's electronic patient record supplier for integration purposes; whether Heidi or the ICB is contractually responsible for ongoing clinical safety monitoring once a trust goes live; and what happens contractually if a trust within the framework later decides the product is not right for a particular specialty and wants to pause or opt out locally. None of this is unusual for a large NHS framework agreement, but it is exactly the kind of detail that separates an optimistic headline from a working deployment.
Buyers and clinicians in the region should expect these details to surface through trust board papers, ICB communications and Heidi's own commercial announcements over the coming weeks and months, rather than assume the framework answers them by default. Anyone advising a trust board on whether to activate the framework locally should be asking for the underlying contract schedule, not relying on the HSJ summary alone.
Why a regional framework matters more than another local pilot
The UK has had ambient scribe pilots since roughly 2023. Most of them were small, locally negotiated, and repeated the same procurement and governance work at every site. A regional framework changes the shape of that process in several ways. It can standardise procurement terms across trusts that would otherwise negotiate separately, each with its own legal review and commercial terms. It can create shared information governance and data protection impact assessment materials rather than fifteen bespoke versions drafted by fifteen different information governance teams. It can align clinical safety documentation and reduce duplicated review by clinical safety officers who would otherwise be assessing substantially similar risks from scratch at every site. It can also give Heidi a single large reference deployment to draw evidence from, across specialties and settings, rather than assembling a patchwork of smaller case studies collected opportunistically.
In procurement terms, this is the difference between a supplier having to win fifteen separate arguments and winning one argument that then has to be operationalised fifteen times. The operational work does not disappear. It moves from the sales and legal stage into the implementation stage, where it arguably matters more, because that is where clinical safety and actual adoption get decided.
How this compares with the wider pattern of AVT procurement so far
Most prior UK AVT deployments, including Heidi's own growth story, have followed a bottom-up pattern: individual clinicians or single practices adopt a tool, generate positive anecdotal or evaluated results, and that evidence then supports a slightly larger deployment. The Modality Partnership rollout, still one of the most cited primary care AVT evaluations in the UK, followed that pattern closely: a defined evaluation of 47 GPs, expanding only after results were reviewed, eventually reaching over 200 clinicians across 53 surgeries.
A 15-trust regional framework is structurally different from that pattern. It is closer to how the NHS has historically procured infrastructure such as pathology systems or picture archiving software: a central decision made at regional level, with local deployment following afterwards rather than preceding it. That is not automatically worse. It can be faster and more consistent. But it does mean the evidence base clinicians are relying on when the framework activates locally may be thinner, at the point of go-live, than clinicians are used to seeing before they trust a new tool with their documentation. Trusts activating early under this framework are, in effect, contributing to the evidence base rather than only benefiting from an existing one.
What clinicians in the framework may receive
Based on Heidi's existing UK product, clinicians onboarded through the framework can expect ambient capture of consultations, draft structured clinical notes, draft letters and follow-up correspondence, configurable templates by specialty, and the option of integration with local electronic patient record systems where supported. Notes are drafted for clinician review rather than filed automatically, which remains the standard expectation across the UK ambient scribing category and is consistent with Heidi's current Class I intended purpose.
What trusts still have to do locally
None of this removes the implementation work that sits with each trust. Local teams will still need to complete their own DCB0160 clinical safety case, appoint or confirm a named clinical safety officer, complete or adopt an appropriate data protection impact assessment, agree patient information and consent wording for their own population, configure access controls and single sign-on, train staff on when and how to use the tool, and establish local error-reporting and post-deployment monitoring routes. A framework can supply templates and reference material for most of this, and Heidi's published NHS compliance documentation already offers deployment packs intended to speed this work up. It cannot substitute for a trust actually doing it, and trust boards should treat framework membership as the start of a governance process, not the end of one.
Why this matters for Heidi specifically
For a company whose UK reputation was built substantially on individual clinician adoption and a widely cited primary care partnership with Modality, a regional secondary and primary care framework of this scale is a different kind of validation. It signals movement from clinician-led, bottom-up uptake towards institutional, board-level procurement. It gives Heidi a large, potentially multi-specialty NHS reference site to draw evidence from, at a scale that would take years to accumulate through opportunistic single-trust deployments. And it arrives at a moment when competitors including Tandem and TORTUS have each secured independent Class IIa medical device certification, making regulatory credibility and institutional scale a live part of how NHS AVT suppliers are compared in tender documents and board papers alike.
What to watch over the coming months
A handful of concrete developments will tell us more about the real shape of this deal than the initial announcement can. Whether Heidi publishes or is required to publish the underlying framework documentation, including scope and financial terms. Which of the 15 trusts activate first, and in which clinical settings. Whether general practice and community providers are confirmed as in scope. Whether Heidi's Class IIa certification process, referenced in its own NHS compliance materials, concludes before or after significant trust-level activation. And whether independently reviewed adoption and safety data, as opposed to vendor-reported figures, is published within the first year.
A balanced read
The Midlands framework is a genuine commercial and procurement milestone. Its long-term significance will depend on what happens next: how many of the 15 trusts actually activate it, how quickly clinicians are onboarded and how many keep using it beyond the first few months, whether meaningful EPR integration follows, and whether independently reviewed safety and productivity data eventually gets published rather than only vendor-reported figures. Until then, "could cover" and "makes available to" are the accurate ways to describe the 70,000 figure. "Is already used by" is not.