AI Automation That Gives Therapists
More Time With Every Child
Agix built a 9-step AI intake and coordination pipeline for Kites Children's Therapy, automating referral capture, smart triage, NDIS eligibility checks, therapist matching, and family communications, so every clinician can focus on what matters most: the child in front of them.
Every child. Any challenge.
Kites Children's Therapy is a leading Australian paediatric allied health provider delivering early intervention and school-age support for children living with disability or developmental challenges. Their multidisciplinary team offers Speech Therapy, Hydrotherapy, Online Therapy, and the unique Dog-Assisted Therapy program, all underpinned by evidence-based practice and deep family partnership.
As demand for NDIS-funded children's services grew rapidly, Kites faced an operational challenge: a manual intake and coordination process that consumed clinician time, created family wait anxiety, and left NDIS funding eligibility checks prone to delay. They needed AI to handle the administrative complexity, without losing the warmth that defines their care.
Four specialist pathways. One AI-coordinated intake.
Each therapy type requires different eligibility criteria, therapist profiles, and session logistics. The Agix AI engine routes every child to the right service from day one, no manual sorting required.
A growing caseload. A team drowning in admin instead of therapy.
NDIS demand for paediatric services doubled in three years. Kites was adding therapists faster than they could build the intake infrastructure to support them. Every new referral required manual data entry, phone-based eligibility checks, and ad-hoc therapist allocation, a process that took days and consumed hours of clinical coordinator time per family.
From referral to first session, fully automated.
Nine intelligent steps that replace manual coordination, running in the background, 24 hours a day.
An end-to-end intake engine built for paediatric allied health at NDIS scale.
Agix built a modular AI pipeline connecting referral capture, clinical triage, service matching, NDIS eligibility, scheduling, and family communications into a single automated flow, with full coordinator visibility at every step.
AI Intake Capture
A unified AI intake form intelligently captures referrals from any source, web form, phone transcription, or email parsing. The system extracts child details, concerns, documents, and family preferences into a structured clinical record automatically, eliminating manual re-entry and reducing intake time from days to hours.
Smart Triage Engine
An AI triage layer analyses the child's documented needs, developmental concerns, and urgency signals, categorising each case by priority and flagging children who require expedited assessment. Clinicians review and confirm; the AI handles the sorting logic across every referral simultaneously.
Service Matching
Based on triage output, the AI routes each child to the right therapy pathway: Dog-Assisted Therapy, Hydrotherapy, Speech Therapy, or Online Therapy, or a combination. Matching logic accounts for the child's age, diagnosis category, geographic access, and family preferences documented at intake.
NDIS Funding & Eligibility Automation
The system automatically validates NDIS plan categories, available support budgets, and registered provider alignment, cross-referencing the NDIA portal data against each family's submitted documents. Eligibility results are returned within minutes, not days, with clear flags for plans requiring coordinator review.
Automated Scheduling & Therapist Assignment
Once eligibility is confirmed, the system matches the child to the optimal therapist across Kites' team, considering specialty, caseload capacity, location, and child-specific needs. It then automatically proposes appointment slots to the family, books the session, and notifies the assigned therapist, no coordinator phone tag required.
Family Communication & Progress Tracking
Automated, personalised communications keep families informed at every stage, appointment confirmations, session preparation notes, therapist introductions, and post-session outcome summaries. A progress tracking layer compiles therapist notes into family-readable summaries and surfaces next-step recommendations automatically.
Less admin. More therapy. Better outcomes for families.
Measured against pre-deployment baselines across Kites' intake and coordination operations.
Before this, our coordinators were spending half their week on intake admin, phone calls, NDIS checks, finding the right therapist. Now the system handles all of that. Our team is doing what they trained for: supporting children and families. The difference has been immediate and visible.
Automation that supports clinicians, it doesn't replace their judgment.
The core design principle was that every automated step should be transparent and interruptible by a clinician. The AI does the pattern-matching and data integration; the coordinator reviews and confirms. This preserved clinical ownership of every intake decision while eliminating the hours of administrative work that surrounded those decisions.
NDIS funding automation was the single biggest unlock. Eligibility was the largest delay in the prior process and the most error-prone. Automating the validation step, while surfacing exceptions clearly for human review, compressed intake timelines by two-thirds and reduced claim rejection rates significantly.
Human-in-the-loop by design
Every AI recommendation is visible to coordinators and reversible. Automation handles volume; clinicians own the exceptions, which are surfaced proactively, not buried.
NDIS complexity as the unlock
NDIS eligibility validation was the single biggest manual bottleneck. Automating it with NDIA portal integration immediately cut intake time by two-thirds, making every other step faster too.
Matching logic that scales
Therapist assignment used to rely on coordinator memory. A rule-based matching engine, drawing on specialty, caseload, and location, makes optimal assignments consistently at any referral volume.
Family experience as a metric
Automated communications were designed to feel warm and personal, not robotic. Personalised prep notes and therapist introductions meant families arrived at their first session informed and confident.
What this system doesn't replace.
Process automation is powerful, but it has limits worth understanding before building something similar.
Clinical Assessment Is Still Human
AI triage identifies urgency signals from documented referral information, it does not replace an initial clinical assessment. Every child still receives a human-led intake assessment before therapy commences; the automation handles the coordination work that surrounds that assessment.
NDIS Portal Dependency
Eligibility automation relies on NDIA portal API availability and data accuracy. When NDIS plans are under review, have outdated budget information, or involve complex plan management arrangements, automated eligibility checks surface the case for manual coordinator review rather than auto-approving.
Complex Family Situations Require Human Contact
Families navigating a new disability diagnosis, requesting urgent support, or with complex cultural or communication needs are flagged by the system for direct coordinator outreach, not automated messaging. The system knows what it can't handle well and routes accordingly.
Data Quality Shapes Matching Quality
Therapist matching accuracy depends on keeping therapist profiles, specialty tags, and availability data current. Stale or incomplete therapist records produce suboptimal matches, so process discipline around data maintenance is a prerequisite for the system performing as designed.
Is this the right build for your organisation?
What powers this system.
AI Process Automation
End-to-end workflow automation for intake, triage, eligibility, and coordination, removing administrative drag across complex multi-step service delivery processes.
Conversational AI
AI intake capture via conversational interfaces, gathering structured information from families across web, phone, and messaging channels without manual data entry.
Healthcare AI Solutions
AI built for the specific operational and compliance requirements of NDIS providers, allied health practices, and disability support organisations.
Predictive Analytics AI
Progress tracking models that surface intervention signals early, supporting therapists with outcome data and flagging children who may benefit from a change in approach.
Custom AI Development
Bespoke AI automation systems designed around your specific workflows, funding frameworks, clinical governance requirements, and patient or client populations.
Coordinator Dashboards
Real-time operational visibility across your full caseload, intake pipeline status, NDIS eligibility queue, therapist capacity, and family communication history in one interface.
Common questions about AI for allied health intake.
The AI triage layer works on documented referral signals, not clinical examination. It identifies urgency indicators from referral notes, reported diagnosis codes, reported developmental concerns, and the child's age and NDIS plan type. The output is a prioritisation recommendation that helps coordinators focus attention, not a clinical diagnosis. Every child still receives a human-led initial assessment before any therapy commences. Triage accuracy improves over time as the system learns from coordinator override decisions.
The eligibility automation layer connects to the NDIA myplace provider portal via API, validating the child's NDIS plan category, support budget line items, and registered provider alignment against Kites' service registrations. The system cross-references the submitted NDIS plan number and date of birth against portal records. Where plan data is current and the support category matches, eligibility is auto-confirmed. Where there are discrepancies, expired plans, or complex plan management arrangements, the case is surfaced to the coordinator queue with a specific exception flag, so they know exactly what to check rather than reviewing the entire plan from scratch.
For a system of this scope, covering intake capture, AI triage, service matching, NDIS eligibility, automated scheduling, and family communications, typical build and deployment time is 10–14 weeks from project kickoff to live operations. The first 3 weeks are discovery and workflow mapping; weeks 4–10 are build and integration; weeks 11–14 are parallel running (AI alongside existing manual process) and staff training before full cutover. Simpler intake automation builds with fewer workflow steps can be live in 6–8 weeks.
Yes, the core intake automation architecture (referral capture, AI triage, service matching, scheduling, and family communications) is funding-framework agnostic. The NDIS eligibility automation module is specific to Australian NDIS providers, but an equivalent integration can be built for other funding schemes, Medicare, private health insurance, UK NHS referral pathways, or US Medicaid. The eligibility layer is modular; it can be replaced or supplemented without changing the rest of the pipeline. Contact us to discuss what your specific funding and referral management context looks like.
Ready to give your clinical team their time back?
Most projects go from kickoff to deployed AI system in 8–16 weeks. Let's talk about what's possible for your organisation.