Secure AI Voice Infrastructure for Public Programs
Empathetic, clinically intelligent voice calls across aged care, veteran services, NDIS, and maternal health — built for population scale, with data stored in Australia.
Enterprise-grade by default
Australian Data Residency
Data stored in AWS Sydney (Australia) with end-to-end encryption; AI processing currently runs in the US (Anthropic/Hume), with zero-data-retention in progress. Essential Eight Maturity Level 3 controls implemented; ISO 27001:2022 aligned, certification in progress. Aligned to the Australian Privacy Act and APPs.
Voice-First, No App Required
Reaches every citizen via a simple phone call — no smartphone, no app download, no digital literacy barrier.
Population-Scale Analytics
Real-time dashboards across thousands of participants with automated risk scoring and trend detection.
Multi-Program Support
One platform powers aged care, veteran support, NDIS, postnatal screening, and carer wellbeing programs simultaneously.
Audit-Ready Compliance
Every call, task, and escalation is timestamped and logged — ready for regulatory review and accreditation.
11 Languages, All Regions
Culturally sensitive conversations in English, Italian, Arabic, Mandarin, Cantonese, Spanish, and more.
Kate is the agentic voice engine behind every persona.
Not a persona. Not a chatbot. Kate is the intelligence layer that makes every call purposeful, every insight actionable, and every outcome actionable. She is the reason this is more than an AI that makes phone calls.
What Kate does on every call
- Schedules the call at the right time for each person
- Selects the right persona (Mary, Jennie, Bloom, or any other)
- Loads full context from every previous conversation
- Configures the conversational pathway with branching logic, conditions, and guard rails
- Orchestrates the conversation in real time
- Receives emotional signals so the conversation adapts to mood in real time
- Triggers mid-call API actions: check a database, book an appointment, update a CRM record, send an SMS
- All while the person is still on the line
- Transcription and recording
- Routine analysis on cost-optimised local models
- Advanced reasoning through frontier models
- Emotional analysis (tone, sentiment, mood)
- Generates WHO-5 wellbeing scores, structured JSON data, and compliance evidence
What makes Kate different
Escalation and action
Escalation alerts to the care team. Family notifications through nonni.ai. CRM updates. Webhook pushes. Compliance evidence packs. Follow-up call scheduling.
Longitudinal intelligence
Kate maintains memory across calls. She tracks wellbeing trends over weeks and months. She detects a gradual mood decline over three weeks, increasing mentions of pain, growing isolation.
Multi-model architecture
The right model for each task. Routine analysis on cost-optimised local models. Advanced reasoning through frontier models. Emotional analysis through specialised voice models. Cost-optimised without compromising accuracy.
Cross-call memory
Every persona remembers. Margaret mentioned her daughter last Tuesday. John has been sleeping poorly for two weeks. Context carries forward automatically.
Conversational pathways
Node-based conversation flows with branching logic, conditions, and guard rails. The AI follows the pathway. Hallucination-proof. Safe by design.
Continuous compliance
Every call maps to the relevant quality standards. Evidence generates as a byproduct of caring, not as an admin task. Always audit-ready.
Request a policy briefing.
Federal and state health teams: tell us a little about your program and what you'd like to know. We'll come back with a tailored conversation — population scale, evidence, governance, data sovereignty — whatever fits.