MHW.01
Intake & Scheduling
Enquiries, intake forms and session booking for therapy practices, on top of SimplePractice, Jane or WriteUpp. Every message is screened for risk first, and any sign of it goes to a person.
Open pageIndustry
Mental health AI
for the admin, not the care.
Admin only. Any sign of risk goes to a person. AI in mental health, as we build it, is administrative only: intake forms, scheduling and draft notes a clinician reviews and signs. It does not provide therapy, assess anyone or handle a crisis, and any sign of risk goes to a person immediately.
Last reviewed 30 October 2026 · Engineering guidance, not legal, clinical or regulatory advice.
In context
AI in mental health, as we build it, is administrative only: intake forms, scheduling and draft notes a clinician reviews and signs. It does not provide therapy, assess anyone or handle a crisis, and any sign of risk goes to a person immediately.
Where it breaks today
01
Admin eats clinical hours
Progress notes, treatment plan updates, letters and insurer paperwork are written after sessions, often in the evening, by the same clinicians whose waiting lists are already long.
02
Intake is slow and repetitive
Enquiries arrive by phone, email and web form, each one needing the same questions answered and the same forms sent, while the person enquiring may be waiting at a difficult moment.
03
General chatbots are being used as therapists
People already turn to general AI chatbots for emotional support. Professional bodies have warned these tools are not a substitute for care and can respond badly to crisis cues. A practice cannot let its own tools drift that way.
OUR APPROACH
A narrow job, with a person behind every edge
We do not build AI therapy, AI companions, mood assessment, diagnostic screening or crisis lines. That is a decision, not a gap in the catalogue. Therapy depends on a relationship and on clinical judgement; assessment of risk is a clinical responsibility; and the evidence on chatbots handling distress is not good enough to put a person's safety behind one. Some US states, Illinois among them, have legislated to stop AI providing therapy, and the EU AI Act prohibits systems that exploit people's vulnerabilities.
What we do build is the administrative layer for therapists, counselling services and clinics: intake and scheduling, and draft notes for the clinician to correct and sign. Every one of those systems carries the same safety design. It screens for any sign of risk before it does anything else, with the threshold set deliberately low, and hands over to a person with the right contact details shown at once. The clinical determination of what that person does next is the practice's, never the software's.
HOW IT DEPLOYS
PHASE 01
Write the escalation policy first
With your clinical lead: what counts as a risk signal, who is on call when, what is shown out of hours. Nothing else is built until this is signed off.
PHASE 02
Connect the practice system
SimplePractice, TherapyNotes, Jane, Cliniko, WriteUpp or Carepatron for diary and records, read with the minimum access each task needs.
PHASE 03
Clinicians review before widening
Escalations and drafts are reviewed by clinicians from day one, and the escalation test set is re-run whenever anything changes.
What we build
01
Intake and scheduling
Enquiries answered from your published information, intake forms sent and checked for completeness, sessions booked against clinician availability.
SEE INTAKE AND SCHEDULING →02
Clinician admin and notes
Draft progress notes from the clinician's own dictated summary, or from a session recording only with explicit client consent, for the clinician to edit and sign.
SEE CLINICIAN ADMIN →03
Safe escalation design
The risk screen, handover rules and testing that every system in this sector carries, also available as a review of tools you already run.
SEE SAFE ESCALATION →04
Guardrails and evaluation
The technical layer underneath: input and output checks, refusal behaviour and evaluation sets built from the cases that matter most.
SEE GUARDRAILS →Controls we implement
ANY SIGN OF RISK TO A PERSON
NO THERAPY OR ASSESSMENT BY AI
CLINICIAN SIGNS EVERY NOTE
RECORDING ONLY WITH CONSENT
NO TRAINING ON CLIENT DATA
AI DISCLOSED FROM THE FIRST MESSAGE
These are engineering controls we build into the architecture, not certifications. QQuantum.ai holds no security certifications, and the clinical governance and data-protection obligations of your service remain yours — see our GDPR notes.
Mental health use cases
-
-
MHW.02
Clinician Admin & Notes
Draft progress notes and paperwork from the clinician's own summary, or a session recorded only with explicit consent. No clinical content is added; the therapist edits and signs the record.
Open page -
MHW.03
Safe Escalation
Risk screening and human handover for any system that talks to people who may be at risk: low threshold, crisis contacts shown at once, a person alerted, and testing on the hardest cases.
Open page
Questions
Questions about Mental Health & Wellbeing.
How can AI be used in mental health?
Safely, mostly behind the scenes. AI can draft notes for a clinician to review, handle intake forms and scheduling, answer practical questions about a service and route enquiries. Uses that put AI in front of a person in distress, such as chatbot therapy or automated risk assessment, carry far higher risk and are not something we build. Every system we deliver routes any sign of risk to a person immediately.
Can AI replace a therapist?
No. Therapy depends on a trained clinician, a relationship and judgement about risk that software cannot carry responsibility for. Research and professional bodies have warned that general chatbots are not a safe substitute, and some jurisdictions have restricted AI therapy by law. AI can take administrative work off therapists so more of their time goes to clients, which is the only role we design it for.
What are the best AI tools for therapists?
For most private practitioners, the AI features in an established practice management system or a dedicated note-taking product are the sensible start, provided the vendor signs the data agreements your jurisdiction requires. A custom build suits group practices, clinics and services that need their own intake rules, their own escalation policy, several systems joined up, or data kept on infrastructure they choose.
Is AI for therapy notes HIPAA or GDPR compliant?
Compliance is a property of how a practice uses a tool, not something a vendor can declare for it. Mental health data is special-category data under GDPR Article 9, and in the US psychotherapy notes have extra protection under HIPAA, with substance use records covered by 42 CFR Part 2. We design to keep data in your perimeter, with no training on it, but the determination and the agreements remain yours.
What happens if someone in crisis contacts the system?
The system stops its administrative task and hands over. During opening hours that means a person on your team, alerted immediately with the conversation attached. At any hour the person is shown emergency and crisis contacts straight away, such as 112 or 999, 988 in the US, Samaritans on 116 123 in the UK or 024 in Spain. The system does not counsel, assess or try to keep the person talking.
