Clinical intake with no AI where a client could meet it

• Built a free clarity check using three validated screening instruments, scored by arithmetic against published clinical cutoffs
• Automated the administrative edge only: the emailed report, the practitioner notification, and a follow-up sequence on day three and day seven
• Built twelve self-help tools and a crisis page with real helpline numbers
• Kept every result explicitly non-diagnostic, in the interface, in the email and on every condition page

• Enquirers get a structured, plain-language report within minutes
• The practitioner sees the screening result in the notification subject line, so triage happens before the email is opened
• Follow-ups go out on their own, and the practitioner is copied on every one

• There is no AI anywhere a client can touch it. No chatbot, no model-scored assessment, no triage by a model.
• The only AI in the project is a separate content agent for the blog. Its validator blocks any unsourced statistic, bans diagnostic claims, and forces a crisis link on any article that touches crisis terms.
• The highest-risk screening question is stored in its own column so it can never be lost in a score.

Administrative minutes per new enquiry, counted by the practice before and after. Enquiry-to-booking rate tracked in the pipeline.

The practice, the practitioner, client numbers, and anything at all about any individual client.