Key points
- At a quaternary children's hospital in Queensland, median time to finalise outpatient letters fell from 7.9 days to 14 minutes across 131 clinicians using an AI scribe.1
- Check attribution first. UK GPs using ambient scribes reported errors most often in multiparty consultations (38%).5
- Record consent from the parent and from a child competent to decide. Under Gillick, treatment can proceed for a competent under-16 even when parents disagree.4
- Keep safeguarding content verbatim. NICE asks clinicians to record exactly what is observed and heard, from whom and when, and why it is of concern.3
- wai drafts the note and the family letter, and the paediatrician reviews and signs, with role-based access enforced in the database.6
Summary
An AI scribe gives a paediatrician the whole consultation back for the child. Set the template to record who said what, say growth figures aloud so they land as numbers, record consent from the parent and any competent child, and keep safeguarding content verbatim. The paediatrician reviews and signs every note.
The child gets your full attention and the family leaves with the letter already drafted. A paediatric room has a parent giving the history and a child adding their own, so the template has to keep their accounts apart and the clinician checks that first.
See it in your clinicThirty minutes on your own patients and doctors.
Request a demoSee a consultation with a parent and child drafted into a note that records who said what, in fifteen minutes with a clinician who uses wai daily.
| Element | What the draft records | What the paediatrician checks |
|---|---|---|
| History given by | Parent, child and any other speaker, each account kept separate | Every statement attributed to the right person |
| Growth | Weight, height, head circumference where relevant, centiles as read aloud | Each figure against the growth chart |
| Development | Milestones by domain, parent report and clinician observation apart | Nothing inferred that was not said or seen |
| Consent and chaperone | Who consented, to what, child's understanding, chaperone name and role | Competence assessment recorded where the child consents |
| Safeguarding | Exact words in quotation marks, who said them and when, reason for concern, actions | No paraphrase, no concern copied into a family letter |
| Family letter | Addressed to parents, copied to the family doctor, a plain paragraph for the child | Tone, accuracy and what is safe to share |
Who said what
A parent's account and a child's own description are different evidence, and the note should say which is which. UK GPs using ambient scribes reported errors most often in multiparty consultations, at 38%, ahead of complex histories and non-English consultations.5
Name the speaker aloud when it matters, for example by asking the parent about the first episode and then asking the child directly. Give the template a field for who gave the history, with room for the child's own words, and check attribution before anything else.
Growth and development as numbers
Say each measurement and centile aloud so the draft captures weight, height and, in infants, head circumference. Plotting stays on the growth chart, and every figure in the draft is checked against it before signing.
Milestones read best as a short list by domain (gross motor, fine motor, speech and language, social), with parent report and clinician observation kept apart. The plan states what will be measured at the next visit and when.
Consent from parents and competent children
Gillick competence assesses whether a child under 16 has the maturity to make their own decisions, and where a child is competent, treatment can proceed even if parents disagree.4 The record shows who consented, to what, and how the child's understanding was assessed.
In a UK survey of 235 paediatricians and paediatric endocrine nurse specialists, most regarded pubertal staging as an intimate examination, yet only 38.2% always documented consent, and many used a parent as the chaperone, which GMC guidance does not recommend.2 A template field for consent, chaperone name and role makes the record complete every time. Consent to the scribe itself is asked of the parent and, where the child can understand, of the child, as set out in patient consent for AI scribes.
Safeguarding notes, verbatim and attributed
NICE guidance on child maltreatment asks clinicians to record exactly what is observed and heard, from whom and when, and why it is of concern, and then every action taken and its outcome.3
Keep the child's and parent's exact words in quotation marks, check nothing has been paraphrased or softened, and add the reason for concern in your own words. Check family letters separately so a concern is never copied into a letter sent home where sharing it could put the child at risk.
On wai, staff access is role-based and least-privilege, enforced in the database, and every draft, approval and signature is written to an exportable log.6
A worked example
Take a seven-year-old who attends with their mother after three months of tummy aches. The mother describes the pattern and school absences, the child points to the area around the belly button and says it hurts most before school, and the examination is normal.
The draft records the mother's and the child's history separately, weight and height with centiles as read aloud, the normal examination, and the plan agreed with the family, a four-week symptom diary, stool and blood tests and a review in six weeks. The letter goes to the parents, copied to the family doctor, with a short plain paragraph the child can read. The paediatrician checks attribution, figures and plan, then signs.
What children's services have measured
At Children's Health Queensland, across 131 clinicians (89 medical, 36 allied health and 6 nurse practitioners), median time to finalise outpatient letters fell from 7.9 days to 14 minutes, document quality scored a median 4 out of 5 in 9 of 10 domains, and more than 90% of 333 patient and family responses reported better clinician interaction.1 It is one service and one tool, and it sets the size of gain worth measuring for.
Where to start
Start with general paediatric follow-up clinics. Build the template with a history-given-by field, a growth block and a consent and chaperone block, and run it for two weeks. Measure time from the end of clinic to a signed letter, and count attribution corrections.
Agree a written rule for safeguarding content before go-live, then widen to new patient clinics. The accuracy guide sets out how to sample and score drafts in the first month.
Bring a paediatric template you use today, and watch a consultation with a parent and child fill it.
Frequently asked questions
Who should consent to an AI scribe in a paediatric consultation?
The parent, and the child too where they can understand what is being asked. Record both. A child assessed as Gillick competent can make their own decisions about their care.4
Can the scribe tell the parent's voice from the child's?
Check it every time. Multiparty consultations are where users report the most errors.5 Name speakers aloud when it matters and review attribution first.
Should a safeguarding concern be recorded with an AI scribe?
The clinician decides. If it is, keep exact words in quotation marks with who said them and when, as NICE expects, and check that the draft has kept them word for word.3
Is a parent a suitable chaperone for an intimate examination?
GMC guidance, as reported in a UK survey of paediatric clinicians, does not recommend using a parent as the chaperone.2 Record the chaperone's name and role.
How much time does an AI scribe save in paediatrics?
At one Queensland children's hospital, median time to finalise outpatient letters fell from 7.9 days to 14 minutes.1 Measure your own before and after over two weeks.
Who can see sensitive paediatric notes?
On wai, staff access is role-based and least-privilege, enforced in the database, and every action is logged and exportable.6
Sources
- 1Journal of Paediatrics and Child Health, Kapoor and colleagues, 2026Improving clinical efficiency using artificial intelligence scribe for documentation (ICE-AID study)Read 14 October 2026Read source
- 2Archives of Disease in Childhood, Moon and Davies, 2023Confidence, consent and chaperones for pubertal staging examinations, a national surveyRead 14 October 2026Read source
- 3National Institute for Health and Care Excellence, 2025Child maltreatment, when to suspect maltreatment in under 18s (CG89), using this guidanceRead 14 October 2026Read source
- 4NSPCC Learning, 2022Gillick competence and Fraser guidelinesRead 14 October 2026Read source
- 5BMJ Health and Care Informatics, Blease and colleagues, 2026Ambient AI in primary care, an exploratory mixed methods survey of UK general practitionersRead 14 October 2026Read source
- 6wai, 2026Security at waiRead 14 October 2026Read source
Your eyes on the child, and the letter drafted by the time the family reaches the door.
Walk through a paediatric consultation of your own with a clinician who uses wai daily, in fifteen minutes, and hear back within a working day of booking.
Partner with us
See it on your own patients, then build it with us
Thirty minutes on your own patients, doctors and rules, in your market. We reply within one working day, and the clinics that partner with us early shape what we build next.
See what wai does between the visits