Care: closing roster gaps without spending your evening on the phone
It is 19:30, tomorrow's early shift has a gap, and the ward lead is still on the phone after her own shift trying to close it. That part of the job can be handed over: an AI phone agent calls the standby pool in parallel, clarifies qualification and rest period and reports back who can come.
Phoning is the wrong work for the ward lead
Roster gaps appear at short notice: illness, a flu wave, a child who needs looking after. What follows is the same in many institutions – somebody in a leadership role calls down the list one by one, often outside their own working hours, often without success, because half the pool is on shift themselves and does not pick up.
The problem is not the pool. It is the sequence: a human can only hold one conversation at a time and loses a minute on every person who does not answer. An agent dials twenty numbers at once and handles the retries on its own.
What that changes in practice: the ward lead starts the run and goes home. Twenty minutes later the roster tool shows who can take the early shift – including from when until when.
Ask for qualification, not just for availability
In care a confirmation is not enough. It has to be the right confirmation. Sensible fields in the call guide:
- Qualification level: registered nurse, healthcare assistant, care aide, assistant with a Red Cross certificate, student placement
- Specialty: acute care, long-term care, psychiatry, rehabilitation, support services
- Shift type: early, late, night, weekend – and whether there is night shift experience
- Rest period: when the last shift ended
- Contract level: whether the person wants to work beyond their agreed hours
- Travel: reachability at the hour in question, especially for night and early shifts
What a call sounds like
AI: «Good evening Ms Steiner, this is the digital assistant of the Muster care centre. We are one qualified person short on ward 2 for tomorrow's early shift, 07:00 to 16:00. Could you step in?»
Ms Steiner: «Tomorrow morning – I can only do it until three, I have to pick up my child.»
AI: «Until 15:00 is helpful, I will note it that way. And when did your last shift end?»
Ms Steiner: «Yesterday, late shift until ten.»
AI: «Thank you, I will pass that on to the roster planning. I will register you for tomorrow 07:00 to 15:00 on ward 2; the confirmation follows by text. Have a good evening!»
The agent does not decide anything it is not allowed to decide. It collects availability and conditions – the actual scheduling stays with the roster planning, with all the information on the table instead of on scraps of paper.
Data protection: the call needs no patient data
This is the point hospitals and care homes rightly examine closely. A staffing call run contains no sensitive health data: it is about shift times, wards and qualifications of employees, not about patients. Processed are name, phone number, qualification and answer – on Swiss servers, ISO/IEC 27001:2022 certified, revFADP compliant and on request as an on-premise deployment in your own infrastructure. A data processing agreement and audit documentation are included in the enterprise package.
Escalation levels instead of a blanket call
Good roster planning does not call everyone at once. The agent therefore works in defined levels – it only moves to the next one if the previous one produced nothing:
| Level | Who gets called | When |
|---|---|---|
| 1 | the ward's internal standby pool | immediately |
| 2 | part-time staff who want extra hours | after 10 minutes without a confirmation |
| 3 | the institution-wide pool, other wards | after 20 minutes |
| 4 | external contacts and the staffing agency | if nothing works out internally |
That keeps things fair within the team: nobody is called at 20:00 while somebody internal is still available.
What it costs
Example: one roster gap
45 contacts in the pool, around 25 talk minutes × CHF 0.39 ≈ CHF 10 per run
Platform from CHF 1'980 per year with 100 free minutes per month, see Pricing.
The real gain is not in that calculation: it is the leadership hour that no longer goes into phone calls – and the shift that does not run short.
Frequently asked questions
Does the AI hear any patient data?
No, and it does not need any. The agent calls employees and talks about shifts, times, wards and qualifications. Patient-related information does not belong in this call and is not provided for in the call guide.
How does it make sure rest periods are respected?
It asks when the last shift ended and marks confirmations where the rest period would be too short. The scheduling itself stays with roster planning – but you see the conflict before you put anyone in.
Does this replace our internal standby pool?
On the contrary, it makes it usable in the first place. Many institutions have a pool that nobody calls through completely under time pressure. The agent works through it in the defined order: internal first, then part-time staff wanting extra hours, then external contacts.
How do employees react to an AI call?
Transparency is decisive: the agent introduces itself as the institution's digital assistant and asks matter-of-factly about availability. Many find that more comfortable than a call from their manager – saying no to a machine is easier and distorts the answer less.
In short
In care, short-notice staffing is above all a burden on the people who already have the most to do. An AI phone agent takes over the calling, keeps to the order and the calling windows, captures qualification and rest period in a structured way – and gives the ward lead her evening back.
Related reading: The main article · The solution · Staffing agencies
Try a call run yourself
Create an outbound agent, upload a test list and listen to the calls before you point it at your real pool.
Start for free Book a demo