AI medical scribes vs AI report writers for ADHD and autism

An AI medical scribe turns one recorded conversation into a well-drafted clinical note. An ADHD or autism diagnostic report is a different document: it pulls the interview together with parent and teacher questionnaires, school reports and developmental history, and every claim in it has to be right, because it follows the patient into school, work and medication decisions. Done well, that draft can still be fast: on Cerebric pilots, drafting the report fell from around two hours to under 30 minutes.

On the surface the two look like similar technology, and they get sold into the same budget, which is why they are so often confused. The documents they produce are not the same, and buying for the wrong one is expensive. This article sets out what a scribe does well, what a diagnostic report needs, and how the two compare on the job they both claim: drafting the document you sign.

What an AI scribe does well

An AI scribe (also called an ambient scribe, or AVT) listens to a consultation, transcribes it, and drafts the clinical note in your chosen template while you keep your attention on the patient. For the job they were built for (the short, high-volume consultation typical of GP and primary care), the leading tools do it well.

Heidi Health has real NHS traction, listing NHS Dudley, Hertfordshire and West Essex, Walsall Healthcare and WWL Teaching Hospitals among its users, alongside dictation, template switching and coding support. Accurx Scribe, from the company behind much of NHS primary care messaging, reports documentation time per appointment falling from around seven minutes to just over one, with one-click saving into EMIS and SystmOne.

Their sweet spot is the short encounter: a ten-minute appointment, a follow-up, a titration review, where the note is essentially a faithful account of one conversation. So, put simply: if your bottleneck is turning conversations into notes (follow-ups, titration reviews, general psychiatry clinics), a scribe can be useful.

A neurodevelopmental assessment is a different shape of job, and that is where the two tools stop being interchangeable.

What a diagnostic report actually requires

An ADHD diagnostic report is a multi-source clinical document, not a long consultation note. NICE guideline NG87 is specific about what sits behind a diagnosis: a full clinical and psychosocial assessment, a complete developmental and psychiatric history, and observer reports alongside a mental state examination. It restricts diagnosis to specialist psychiatrists, paediatricians and other appropriately qualified professionals with training in ADHD, and, critically for anyone hoping to automate the whole thing from a transcript, states that a diagnosis "should not be made solely on the basis of rating scale or observational data".

NG87 also requires that symptoms cause at least moderate psychological, social, educational or occupational impairment, present in two or more settings. Two or more settings means two or more informants. In practice a defensible report weaves together:

• the clinical interview (a structured instrument such as DIVA for adults, or ADOS-2 and ADI-R for autism pathways)
• parent, teacher or partner questionnaires (NICE names Conners' rating scales and the Strengths and Difficulties Questionnaire as useful adjuncts)
• school reports, occupational evidence and prior clinical records
• a developmental history that usually predates the appointment by decades

Autism assessments make the same point more forcefully. An ADOS-2 and ADI-R pathway runs longer, often across two clinicians and two appointments, and the report has to reconcile direct observation, a parent interview about early development, and school or nursery evidence. ADHD is the shorter example here; everything below applies at least as strongly to autism.

A scribe is built around one of those sources, the conversation in the room. The rest can be got in, up to a point: Heidi and most of its competitors accept document uploads alongside the transcript, so a school report or a scored questionnaire can be attached by hand. What that does not change is who does the handling. Someone still chases the teacher, scores the questionnaire, saves the result as a file and uploads it to the right note, for every source and every assessment.

How each source reaches the draftWith an AI scribeWith an assessment report writerClinical interviewDIVA, ADOS-2, ADI-RRecorded and transcribed in the roomRecorded and transcribed in the roomInformant questionnairesConners, SDQ, SRS-2Chased by email, scored by hand, saved as a file, then uploadedSent, completed and auto-scored in the platformSchool reports and prior recordsOccupational evidence, past lettersUploaded file by file, for every assessmentAttached to the case at intake, read with everything elseDevelopmental historyOften taken decades after the eventsRetyped or dictated into the noteCaptured in structured intake, carried into the draftBoth tools accept uploads, so the difference is handling, not access. A scribe takes the sources one file at a time from the clinician; a report writer inside the platform already holds them, because the questionnaires and intake were completed there. The clinician still makes every clinical decision and signs the report.Both tools take uploads. The difference is who does the handling.

On an assessment platform that work happens upstream and stays in one place. The questionnaires are sent, completed and scored in the platform, the intake history is captured there, documents are attached at referral, and the assessment is recorded there too. By the time the report is drafted, the sources are already in context, so there is nothing to assemble first. That gap is why clinics using a scribe for assessments still report substantial write-up time: the scribe compresses the transcription-and-prose step, not the gathering or the synthesis. (For UK turnaround norms, see how long an ADHD report should take.)

The verification gap

AI makes drafting a report easy, but the effort moves downstream, to verification: the step where a clinician confirms that every claim in a draft is true before signing it. For a neurodevelopmental report, long and drawn from several sources, checking that thoroughly is a new piece of work the clinician now owns. It is also where a scribe and a report writer built for assessments come apart.

Checking a consult note is quick, and not only because you were in the conversation ten minutes ago. The note is also short. Scribe notes are concise by design, which is what makes them fast to read and also where they tend to lose clinical nuance. Ask for that nuance back and the note gets longer, and the checking grows with it.

A diagnostic report is long because it has to be. The draft makes claims about a teacher's questionnaire from six weeks ago, a school report from 2014, and something said in minute 97 of a 150-minute interview. Checking those from memory is far more complicated and time-consuming than reading back a page of consultation notes, and when it gets rushed the risk is a claim that misrepresents what an informant actually said, or a clinical detail that is plainly wrong, in a document that follows the patient for years.

Where one sentence of the report comes fromChildhoodDevelopmental historyTaken from a parent, decades after the events2014School reportScanned into the record6 weeks agoTeacher questionnaireCompleted and scored before the appointmentMinute 97Of a 150-minute interviewOne answer, deep inside the transcriptOne sentence in the draft“His mother described him losing school equipment from age seven, his 2014 school report records the same, his teacher rated inattention in the elevated range, and he gave his own examples at interview.”A summary of what the sources say, each one checkable beside the draftThe evidence and the draft sit on different planes: the sentence is written today, the evidence behind it was gathered over years. The draft restates what each source says and nothing more. Whether that evidence amounts to impairment, and what it means diagnostically, is the clinician’s judgement, written by the clinician. Checking a summary from memory is slow and error-prone, which is why each sentence carries a citation the clinician can open next to the draft before signing.Four sources, gathered years apart, behind one sentence of the report.

This is the problem a purpose-built report writer is designed around. In Cerebric, every sentence in a drafted report is traceable to its source (a transcript timestamp, a questionnaire answer, or an uploaded document) and reviewed in a Split View that shows the statement and its evidence side by side. What the draft does is restate and cite the evidence the clinician has gathered. It does not interpret that evidence: whether the picture amounts to impairment, what the formulation is, and whether the criteria are met are clinical decisions, written and signed by the clinician. Sections can be regenerated individually rather than re-rolling the whole document, every change is tracked, and QA checklists gate the report at approval. Verification becomes a systematic pass rather than a re-read, which is how the whole draft-to-signature pass drops to under 30 minutes.

Clinicians notice the difference where the structure lives. As one consultant psychiatrist using Cerebric put it: "I had my report, recommendations and all, finished by 5pm. I love how quick and easy it makes filling in all the DIVA boxes. I am over the moon." The DIVA boxes are the point: a diagnostic report is mostly structure and evidence, and prose is the easy part. We cover the record-keeping side in more depth in making AI clinical reports audit-defensible.

Scribe draft vs report-writer draft

Put the two on the same job, drafting the document you sign at the end of an assessment, and the comparison is like for like:

If the document you sign is a consultation note, a scribe is the right purchase: Heidi and Accurx are strong options with real NHS footprints. If the document you sign is a diagnostic report, compare scribes against report writers rather than against each other. The scribe is drafting a shorter document from a fraction of the evidence, and that shows up in rewriting and checking time rather than in the demo.

The takeaway

Match the tool to the document you sign. Short consultations suit a scribe. ADHD and autism diagnostic reports suit a report writer built for them, one that already holds the questionnaires, history and recording, drafts into your report template, and shows you the source behind each sentence.

The test takes one assessment. Draft the report both ways and time what happens after the draft appears: how much you add because a source was never in the tool, how much you rewrite because the draft is pitched at note length, and how long it takes to satisfy yourself that every claim is true. That last step is where the time goes on a report, and it is the one a citation-backed draft actually shortens. Ask any vendor the same two questions: where did this sentence come from, and how long does it take me to check the whole document?