Heidi does not need to answer Tandem's funding round with a matching announcement to make a convincing competitive response. A stronger case would show that clinicians return to its tools, that organisations can deploy them successfully, and that the combined experience solves recurring problems without creating new administrative work.
Strategic analysis, 14 September 2026. Suggested priorities are editorial judgements, not a report of Heidi's internal plans. Product availability is described from public documentation checked on this date.
The Midlands procurement is a route to use, not a use count
Heidi's announcement dated 15 July 2026 in its body describes a Midlands procurement covering 70,000 clinicians, with Heidi selected through a process led by The Dudley Group and Sandwell and West Birmingham. The current page displays a later July publication label; the announcement's stated event date is the relevant one.
That coverage is commercially significant because it establishes an organisational route to deployment. It is not evidence that every covered clinician is an active user, that every organisation has completed implementation, or that the entire Heidi product range is available through the arrangement.
Calling Heidi only a consumer-style app would therefore miss part of its position. Calling the procurement figure an active-user total would make the opposite mistake. The interesting question lies between those claims: how effectively can the company turn an available route into repeated, useful clinical work?
Tandem's 14 September funding announcement makes that execution question more pressing, but did not create it.
Individual preference and organisational deployment are different tests
An individual clinician can decide that a tool helps with a particular letter or consultation. An organisation must also decide how people access it, which workflows are supported, who deals with problems and how the output fits its records. A product can succeed at the first test and still require considerable work at the second.
The two routes could reinforce one another. Familiar users may help colleagues understand a new system; organisational support may make an individually useful tool easier to use consistently. These are plausible mechanisms, not measured explanations of Heidi's growth.
There are also reasons the routes might diverge. Someone who likes personal dictation may not want a different documentation template imposed across a department. A service may need standardisation that an individual user would not choose. A credible implementation recognises the tension rather than treating every request for flexibility as resistance to change.
For Heidi, this suggests a more precise ambition than "more adoption": preserve useful individual preferences while making team-level deployment dependable.
The evidence opportunity has an access boundary
Heidi's current product site describes scribing, dictation and evidence tools. It is reasonable to ask whether reference questions could give clinicians another reason to return between documentation tasks. It is not reasonable to assume that the resulting retention benefit has already been demonstrated.
The access details are consequential. Heidi's UK and EU guidance, dated 14 July 2026, distinguishes standalone Evidence from in-session Evidence. As checked on 14 September, it says in-session Evidence and patient/session linking are unavailable to UK/EU accounts, and NHS UK email accounts are excluded from Evidence.
Those restrictions do not mean Heidi Scribe disappears. They mean an NHS documentation procurement should not be described as automatically delivering the same evidence experience advertised to every other market or account type.
A competitive strategy based on combined products must therefore begin with the combination a particular user can actually access. A missing feature caused by account eligibility needs a different response from a poor feature, and neither should be confused with lack of user interest.
Measure a useful return, not another login
Consider a fictional outpatient department planning a documentation rollout. The proposed evaluation follows a consultation from preparation to a checked record. It separately records whether a clinician returns on a subsequent eligible clinic day. This is a suggested study design, not a description of results from a Heidi deployment.
A useful return means the tool completes a relevant job. An extra login prompted by an error or an unanswered support request is not the same kind of engagement as a voluntarily completed consultation. Product analytics need enough context to separate them.
The department could record activation, continued use, abandoned attempts, substantive corrections and successful filing. It should also ask non-users what stopped them: no relevant task, unsuitable equipment, confusing access, output problems or a preference for another method. These categories point towards different remedies.
Reporting only the most enthusiastic users would make the evaluation less useful. So would treating a trainee rotating away from the service as equivalent to a clinician abandoning the product because it failed. A fair adoption analysis should preserve the reason for discontinuation where it is known.
Test the retention hypothesis without rewarding unnecessary activity
The hypothesis worth testing is specific: does access to a useful evidence workflow increase continued use or willingness to pay among otherwise comparable, eligible users? It is not simply whether users with more features click more buttons.
A proposed evaluation could distinguish documentation-only use, evidence-only use and use of both. It should define account eligibility first, record relevant roles and compare similar observation periods. Without that groundwork, a difference could reflect geography, clinical setting or a paid-plan restriction rather than the value of combining products.
Where feasible, a controlled rollout would provide a clearer comparison than a retrospective chart of successful accounts. Short of that, transparent cohort definitions would still be an improvement over presenting total queries as a retention measure.
The company should also look for the opposite effect. An additional panel might distract clinicians, duplicate an existing reference subscription or complicate onboarding. A strategy is stronger when its evaluation can detect that possibility rather than only count benefits.
What Heidi should prioritise before another long feature list
The most valuable next improvement may differ by setting. A community service could need dependable audio capture in its actual rooms. An outpatient department could prioritise templates and record transfer. A clinician moving between organisations could need clearer information about what their account can use.
These are not findings that Heidi has specific defects in those areas. They are proposed discovery questions. Each should be tested with the people doing the work before it becomes a roadmap commitment.
A practical prioritisation rule is to favour changes that remove a repeated obstacle from a supported task. Record the obstacle, the people affected, the proposed correction and the evidence that would show improvement. That is more actionable than the instruction to "become an AI care partner", even where the broader ambition is commercially sensible.
Reliability should also be evaluated after upgrades and across supported settings. A strong average experience is not sufficient if a particular clinic cannot complete a routine workflow. The relevant unit may be the clinic session or task type, rather than the company-wide user count.
A clinician does not need every task inside one brand
This article is published by iatroX and includes its learning tools in the comparison. The iatroX learning offer, checked on 14 September 2026, includes question-specific Socratic tutoring. That serves a different purpose from generating a clinical note.
For a doctor preparing for an examination, the useful question is whether a learning tool helps them work through a misunderstanding and practise again. The organisation's choice of scribe need not settle that personal learning decision. Equally, a clinician who only needs documentation should not buy an educational subscription simply because it also uses AI.
This leaves room for complementary choices without suggesting an existing Heidi-iatroX integration. The products can be selected independently for different tasks.
The competitive verdict depends on the buyer
For a healthcare organisation, Heidi's argument should rest on the supported deployment, the reviewed work completed and the experience after initial enthusiasm fades. For an individual clinician, the test is whether the available tools repeatedly remove a real problem at an acceptable cost.
For a learner, documentation adoption is largely beside the point: compare the teaching, practice and feedback needed for the learning goal. These are different buying decisions, even when the same person participates in all of them.
Heidi's most persuasive response to a well-funded competitor would be evidence that people keep choosing its available products for good reasons. Another funding round could support that work, but cannot substitute for it.
Frequently asked questions
Does Heidi's Midlands procurement mean it has 70,000 active NHS users?
No. The July 2026 announcement describes procurement coverage, not a verified active-user count.
Does a Heidi NHS scribing account automatically include Evidence?
No. The UK/EU access guidance checked on 14 September 2026 explicitly excludes NHS UK email accounts from Evidence, and feature eligibility should be checked separately from scribing access.
Would adding evidence search necessarily improve Heidi's retention?
No. It is a plausible hypothesis that needs evaluation among eligible users, with attention to their roles, available features and reasons for returning.