Coming Soon

Ember

Disaster Recovery

Emotional-support check-in calls for communities recovering from natural disasters — for the long years after the cameras leave.

Ember will provide regular emotional-support check-in calls to people recovering from bushfires, floods and other natural disasters — a warm voice that listens for distress, isolation and unmet practical needs, and quietly surfaces anything that should reach a person on the recovery team. Ember offers emotional support and connection. It is not clinical care, therapy or diagnosis, and it is not a crisis service.

We are seeking partners in emergency management and disaster recovery. If your organisation works in this space, we would welcome the conversation.

The recovery is longer than the response

A decade after Black Saturday, 22% of people in the worst-hit communities still lived with PTSD, depression or severe distress.1

Recovery after a major disaster takes five to ten years — but support and attention fall away in the second year, when "the eyes of the world move on".2 Distress doesn't fade on that timetable. In the world-leading Beyond Bushfires cohort, probable adjustment disorder actually rose from 15% at five years to 19% at ten years,3 and sub-clinical adjustment difficulty carried a roughly five-fold risk of escalating to a severe disorder at the next wave.3

The strongest predictor of later PTSD and depression isn't the fire itself — it's the ongoing life stressors of recovery: housing, insurance, finances.4 After the 2017 Lismore flood, people displaced more than six months had 24 times the odds of probable PTSD.5 That long, quiet stretch — roughly months six to twenty-four, after the volunteers and media have gone — is the gap Ember exists to cover.

22%

of people in the worst-hit Black Saturday communities still had PTSD, depression or severe distress 10 years on1

1 in 5

Australians with high Black Summer bushfire exposure met the clinical screening cut-off for PTSD6

24×

the odds of probable PTSD for people displaced more than 6 months after the 2017 Lismore flood5

5–10 yrs

how long recovery takes — yet formal support recedes in the second year2

How it works

A warm call that stays alongside, long after the response ends.

Scheduled check-in calls

Ember phones on a phase-mapped cadence through the recovery — lighter while agencies are still on the ground, then steadier through the months when formal support recedes, with extra calls around disaster anniversaries. The approach is modelled on Australia's long-running Telecross welfare-call precedent.7

Psychological-First-Aid conversation

Calls follow the Look–Listen–Link approach of Psychological First Aid8 — a wellbeing check, a talk through practical stressors like housing and insurance, and a social-connection prompt. No debriefing, no forced retelling of the event, no clinical screening. Emotional support, not treatment.

Unmet-needs & distress detection

Ember listens for the signals the evidence flags as load-bearing: persistent or worsening distress, prolonged displacement, isolation, financial or insurance crisis, and anniversary distress. It surfaces a clear, plain-language summary for a person on the recovery team to review — Ember never makes the call itself.

Engineered human escalation

Distress or self-harm cues trigger a deterministic, rehearsed warm handover to people, with Australian support numbers built in — never left to the model's discretion. Independent testing found 0 of 29 mental-health chatbots responded adequately to escalating suicide-risk messages,9 which is exactly why Ember's escalation path is engineered, not assumed.

Designed from the evidence, honest about what's proven

Modelled on what already works — human recovery outreach.

Why scheduled calls

Proactive scheduled welfare calls are an accepted, valued Australian disaster-recovery modality — Red Cross Telecross has run daily welfare calls with unanswered-call escalation for decades.7 Ember's cadence is modelled on those human outreach protocols. Ember itself is a support-and-connection tool informed by that evidence — not a treatment — and we treat evaluation as ongoing.

Why the long tail

The Royal Commission into National Natural Disaster Arrangements found mental-health impacts persist long after the event and recommended governments plan for services following disasters.10 Distress is delayed and rising at ten years.3 Ember is built to own that 6–24-month window.

Why human-in-the-loop

There are no trials of AI voice check-ins in post-disaster populations, and general-purpose chatbots mishandle crisis.9 So Ember stays firmly non-clinical: it extends reach and connection, and every concern that matters reaches a person. It does not replace Red Cross, government outreach or case support.

Powered by Kate

Every call Ember will make is orchestrated by Kate, the intelligence engine behind all CAREPLANS AI companions. Kate manages scheduling, emotional analysis, unmet-needs detection and the engineered escalation to people across every persona and every vertical.

Safety, privacy & what Ember is not

Careful by design, in a domain where care matters most.

Emotional support — not clinical care

Ember does not diagnose, treat or prevent PTSD, depression, anxiety or any disorder, and it does not provide counselling or therapy. It offers emotional support, listens, and flags unmet needs for people to act on. Clinical decisions always rest with qualified humans.

Not a crisis service — distress goes to people

Ember is not a crisis line. Urgent concerns are priority-flagged to a person on the recovery team through a deterministic, rehearsed path, and anyone in distress is warmly guided to Lifeline 13 11 14, the Lifeline disaster line 13 HELP 13 43 57, 13YARN 13 92 76, or 000 in an emergency.

Safe, careful language

All Ember copy and call scripts follow Australia's Mindframe guidelines for safe communication about suicide, mental ill-health and traumatic events11 — non-stigmatising, person-first, no method detail, always pointing to help.

Data handling & security

Data is stored in AWS Sydney (Australia). AI processing currently runs in the United States (Anthropic and Hume), with zero-data-retention in progress; we never train on customer data. Essential Eight Maturity Level 3 controls implemented; ISO 27001:2022 aligned, certification in progress. Built on Claude and Hume EVI.

If you or someone you know needs support now: Lifeline 13 11 14 · Beyond Blue 1300 22 4636 · In an emergency call 000.

Partner with us

We are seeking emergency-management and disaster-recovery partners to bring Ember to disaster-affected communities. If you work in this space, we would welcome the conversation.

andrew@careplans.io

Sources

  1. Bryant R.A. et al. (2021). Mental health and social connectedness across the bushfire-affected communities. Australian & New Zealand Journal of Psychiatry — 22% of adults in high-impact Black Saturday communities still had PTSD, depression or severe distress 10 years on.
  2. Australian Red Cross, long-term flood recovery reporting (2024): recovery takes 5–10 years, but "in the second year fatigue sets in, the eyes of the world move on" as formal programs wind down.
  3. Pacella et al., BJPsych Open (2024), Beyond Bushfires 10-year wave: probable adjustment disorder rose from 15% (5 yrs) to 19% (10 yrs); sub-clinical adjustment disorder predicted ~5× risk of escalating to a severe psychiatric disorder by 10 years.
  4. Bryant R.A. et al. (2017), Beyond Bushfires: ongoing life stressors — not the fire itself — were the strongest predictor of later PTSD/depression (ORs 2.1–3.2 at 5 years), supporting recovery-phase support.
  5. Longman et al. (2023). After the 2017 Lismore flood. Australian & New Zealand Journal of Psychiatry (UCRH cohort, n=2,530): displacement >6 months carried adjusted odds of ~24.4 for probable PTSD.
  6. Australian National University, Bushfire 2021 Survey Report: 1 in 5 people with high Black Summer (2019–20) bushfire exposure reported symptoms at the clinical cut-off for PTSD.
  7. Australian Red Cross Telecross / Telecross REDi: daily volunteer welfare calls 365 days a year with an escalation procedure if unanswered; REDi activated on declared extreme-weather events — the established AU precedent for scheduled welfare phone calls.
  8. World Health Organization (2011), Psychological First Aid: Guide for Field Workers — Look–Listen–Link: practical, calming, connecting support without debriefing or forced retelling of trauma; core of Red Cross and Phoenix Australia disaster practice.
  9. 2026 C-SSRS-based independent evaluation of 29 mental-health/companion chatbots: 0 of 29 gave adequate responses to escalating suicide-risk messages (14 inadequate, 15 marginal). Ember's design assumes AI alone is not safe for crisis moments.
  10. Royal Commission into National Natural Disaster Arrangements (2020), Ch. 15: mental-health impacts persist long after the event; aftermath stressors prolong trauma; governments should plan for delivery of mental-health services following disasters with consistent outcome metrics.
  11. Everymind, Mindframe national guidelines for safe communication about suicide, mental ill-health and traumatic events, and the "Our words matter" language guide (2023).

These statistics describe population research on disaster recovery, not Ember's own outcomes. Ember is modelled on human recovery-outreach evidence; its effectiveness as an AI voice check-in is unproven and under evaluation. Ember provides emotional support and unmet-needs detection — it is not clinical care, therapy, diagnosis or a crisis service.