Dot's Tuesday, minute by minute.
The call
Dot · demo clientKate · the engine
voice into actionAround Dot
nurse · family · mealsSimulated demonstration. No real call, text, order or delivery happens on this page. Dot, her family and the nurse are made-up demo characters.
One engine, many hands.
Mary is the friendly voice on the phone. Kate is the engine behind her, turning what is said into the next right action, and handing the decisions that matter to people.
- 1Mary listens, Kate understands.Routine check-in calls are where small worries surface. Kate tags what was said and screens for red flags straight away.
- 2A nurse decides, not the software.Kate opens the case and alerts the on-call nurse with a short summary. The nurse approves what is safe to order.
- 3The client says yes, every time.Mary asks, in plain words, before telling the family and again before anything is ordered. Each yes is recorded.
- 4The provider gets a clean order.Dietary notes, allergies, delivery window and funding come through as one signed request.
- 5Mary keeps checking.She calls back the same evening and again after the delivery, and the case closes only when food is actually being eaten.
Care before convenience.
A meal order is the easy part. The design work is in everything around it.
Red flags go first
Dizziness, chest pain, confusion or trouble swallowing change the whole path. The client is told to call triple zero, a nurse and family are alerted at once, and the meal order is held.
No yes, no order
Telling family and ordering meals each need a spoken yes from the client. Nothing is sent on a guess, and the consent is time-stamped on the case.
A nurse approves the choices
Texture, portion size and dietary needs are set by the nurse. The provider only ever receives choices the nurse has cleared.
A bridge for tonight
Meals arrive tomorrow, so Kate does not stop there. The family is asked to look in this evening, and Mary calls back to check Dot has eaten.
Every step on the record
What was heard, what was decided, who said yes and who acted is logged on one case. It is ready for the client, the family, the care manager and an auditor.
Caps and funding respected
The order stays within the budget the care manager has set, so there are no surprise charges for the client or the family.
What the order looks like on arrival.
One structured request per decision. It carries everything needed to deliver safely, and nothing that is not needed. This is a simulated example for discussion.
// meal_order_request (simulated) { "event": "meal_order_request", "simulated": true, "client": { "first_name": "Dot", "suburb": "(demo)" }, "pack": { "duration_days": 7, "texture": "soft", "portion": "small", "allergies": [] }, "delivery": { "window": "Wed 10:00 to 12:00", "access": "key safe, code released at dispatch" }, "authorised_by": { "client_yes": "by voice, 2:24 pm", "nurse_approved": "2:21 pm" }, "funding": { "source": "client budget", "cap": "set by care manager" }, "signature": "signed by careplans AI" }
- →Accept, change the window, or decline.The provider replies with one of three outcomes, and Kate tells the client, the family and the nurse.
- →Status events as the order moves.Accepted, out for delivery and delivered keep the family and the care team informed without phone tag.
- →One conversation to start.A capped, manually fulfilled pilot is enough to prove the flow before anything is automated.
- →The same rail for medications.When a medication needs to be taken with food, the same flow can add a meal to the plan.
A small sentence, handled well.
The goal is simple: when someone quietly mentions they have not eaten, the right people hear about it within minutes, and the next meal is already on its way.
Talk to Andrew