How ambient documentation lands across allied health disciplines, where clinicians and patients recorded strong acceptance, and where the real work sits.
Most published evidence on ambient AI documentation comes from medical specialties (Kanaparthy et al., 2025). Allied health documents differently: physiotherapy, occupational therapy, speech pathology, dietetics, social work and psychology each carry their own note structures and terminology, with more focus on narrative functional reporting and progression tracking.
Whether ambient AI scribes suit that work has been largely untested. Gold Coast Hospital and Health Service, one of Australia's largest public health services, published a mixed-methods evaluation of clinician and patient acceptance of an AI scribe across its allied health services.
This article summarises those findings and extracts practical learnings for teams considering or implementing these tools.
The study measured acceptance and perception rather than objective note quality, so these are best read as a strong signal of fit and willingness to adopt, not a measure of output accuracy.
Clinician survey, n = 97. Ryan et al., Digital Health (2026).
Collectively, the results show strong clinician acceptance and clear perceived benefits across efficiency, documentation, patient care and wellbeing, alongside usability and patient-confidence considerations that require attention.
of allied health clinicians wanted to keep using the scribe, and 89% would recommend it to a colleague. Strong acceptance from a workforce whose documentation looks nothing like general medicine.
These reflect Lyrebird's interpretation of the GCHHS findings, informed by implementation experience. They are written for allied health specifically.
Clinicians found the scribe versatile across settings, but named discipline-specific failure modes: terminology that didn't translate, occasional wrong-speaker attribution, and spelling errors in specialist vocabulary. The relevant question is not whether it performs well in general, but whether it fits the documentation demands of a given discipline.
Those errors are also why clinician review remained central. Clinicians described review as "an integral part of safe practice," and noted that ambient capture is not appropriate in every encounter, for example with a patient experiencing persecutory delusions involving surveillance.
The tool is evaluated against each discipline's own notes, the details that matter are fast to verify and correct, and clinicians retain discretion over whether to use it at all.
Customisation was the most frequently cited source of friction. Clinicians described it as time-consuming, and only about half found it easy. Yet the clinicians who invested in it were those who reported being "surprised" by the accuracy of the output. Setup effort and output quality were closely linked.
Teams plan for an upfront settling-in period and support template setup actively, rather than leaving it to busy clinicians. The goal is to make customisation faster and better supported.
Reduced documentation stress, greater job satisfaction, and relief from after-hours documentation were prominent, with one clinician suggesting they might have left their role without the tool. At the same time, clinicians raised a longer-term risk: over-reliance could erode documentation skill or flatten an individual's clinical voice.
Implementation captures the wellbeing benefit while preserving judgement and individuality, through review habits, customisation that keeps each clinician's style, and training in critical evaluation.
Patients who responded were broadly positive about clinician attention, but a meaningful share were unsure about safety, privacy and future use, and the patient sample was small and hard to recruit. The authors are candid that this limits what can be concluded about the patient perspective.
Patient experience and consent are actively measured, not assumed. How the tool is explained matters, and patient voice, especially among vulnerable groups, is captured deliberately.
This study is useful because it examines real-world acceptance across allied health disciplines rather than a single medical specialty.
How easily can templates be tailored to each allied health discipline, and how much setup effort is realistic? What support is provided during the settling-in period, and the long tail thereafter?
How well are speaker attribution and specialist terminology handled, and how easily can clinicians verify and correct the details that matter before finalising a note?
Given that customisation is the main barrier to value, what onboarding gets clinicians productive quickly, and does training build critical-evaluation habits as well as tool skills?
Do the benefits persist as novelty fades and customisation matures?
How is allied health documentation quality best measured objectively, given the limited validated instruments for these disciplines?
What does the patient perspective look like at greater depth, including among vulnerable groups?
How safe and appropriate are AI scribes in acute mental health contexts specifically?
Because Lyrebird's ambient scribe was used in this study, its findings translate directly into how we build and support the platform, particularly around template customisation, discipline-specific accuracy, safe-use controls, and supporting clinician review across very different allied health workflows.
We understand that human judgement remains the gold standard for assessing the quality of a clinical note. Our framework for clinical note quality evaluation reflects that. We do not claim to have solved customisation or eliminated every error; our focus is on making setup easier, review faster, and quality issues easier to surface and fix.
This analysis was prepared by the clinical and research leadership team at Lyrebird Health, who are committed to objective interpretation of research findings and transparent discussion of both benefits and limitations.
Lyrebird is the clinical AI platform for Australian clinicians. Ambient scribing is one core feature, shaped to the way each discipline actually documents, with clinician review built in.
