Domiciliary Care Rota Template: When Software or AI Fills the Gaps

A domiciliary care rota template needs one row per visit with client code, start and end time, tasks, required skills and a named carer, plus a separate travel line between visits and a continuity flag. AI can suggest cover when a gap opens and draft the messages, but a coordinator approves every allocation.

The rota is the product in domiciliary care. Clients experience the agency as who turns up, when, and whether they have met them before. Gaps, late runs and a revolving cast of strangers are the commonest complaints, and they usually trace back to a rota that hides its assumptions. A good template exposes those assumptions so a person, or a tool, can check them.

Follow me on Instagram@sagnikteaches

AI's genuine jobs here are narrow: suggesting candidates, drafting cover messages and summarising the week, all against rules you have written down first. It does not replace rostering software, and where your software already does a job, use that. For the software side, see what the AI features in domiciliary care software actually do.

Connect on LinkedInSagnik Bhattacharya

The template: one visit per row

Build the rota as a table of visits, not as a grid of carers by days. A visit table can be sorted by carer, by client, by time or by run, and it is the shape any scheduling tool expects when you import or export. Use these columns.

Subscribe on YouTube@sagnikteaches

Keep a second table for carers: staff code, contract hours, availability, transport (car, bike, bus, walking), skills held with expiry dates, and any restrictions agreed with them. A third, small table lists the rules, which is where the template earns its keep.

Writing the rules a rota must obey

Most rota problems come from rules that exist only in the coordinator's head. Put them in a sheet your team can read, and number them so a message can say "this change breaks rule 4".

1. Skills: no visit goes to a carer without every listed skill and a current training date.
2. Double-ups: both carers must overlap for the full visit, and neither can be pulled onto another call during it.
3. Travel: the gap between visits must be at least the stated travel time at that hour, plus a buffer your agency chooses.
4. Time-critical visits: flagged calls (for example time-specific medication) move only with coordinator approval and a documented reason.
5. Continuity: a client with a regular-carer flag is offered their regular carer first and a named alternative second. The second choice is always someone who has visited before.
6. Hours: contracted and maximum hours per week and rest periods, as set by contracts and working-time rules your adviser confirms.
7. Preferences and restrictions: carers who cannot work with a certain client for agreed reasons are never offered that visit.

The exact limits (maximum hours, breaks, buffers) come from carers' contracts, commissioner requirements and employment advice, so treat any figures in this tutorial as examples to replace.

Travel time: where templates quietly fail

The commonest hidden fault is travel time that is a flat guess. A flat 10 minutes between every call is wrong for a rural run and also wrong for the same carer on a congested road at 08:30. Record travel as its own line, tag it by area and time band, and review it against a sample of actual arrival times every month.

A quick sum shows why. Say a carer has six visits in a morning run, each with a 10 minute assumed gap, so 50 minutes of travel allowed across five gaps. If real travel averages 14 minutes, the run is 20 minutes behind by the last call. That pushes the last client's call past the window and risks a late medication prompt. Nothing is wrong with the carer. The template's number is wrong.

A filled-in morning run

Here is an imaginary single-carer morning run, with codes in place of names. Times are illustrative.

Run the rules against it. The 08:51 visit is a double-up with a hoist, so the second carer must be hoist-trained and available from 08:51 to 09:36. If she is on a different run until 09:00, rule 2 is broken. This is the sort of conflict a person finds on a good day, and a missed one on a bad day, which is exactly what a rule-check prompt is for.

Where AI helps with a rota, and where it should stop

Separate checking, suggesting and deciding. AI is useful for the first two when the data is anonymised and your care system or spreadsheet is the source of truth.

• Checking: feed it the day's visits and your numbered rules and ask which visits break which rules.
• Suggesting: when a gap opens, ask for ranked candidates with reasons, using staff codes and skill lists.
• Writing: draft the message to each candidate, the message to the client's family if a time changes, and the end-of-day summary.
• Deciding: never. A coordinator makes the allocation and records it.

If your scheduling software already offers suggestions, use them first because they read live data. A general AI tool works from whatever you paste, which can be stale within the hour. Tools differ more than their marketing suggests. As of November 2026, Birdie's rostering page lists an auto-assign tool that looks for matches for unallocated visits, scheduling by continuity, travel time visuals and conflict detection. CareLineLive's rostering page describes carer-client matching on skills and continuity, travel time calculated between visits and conflict warnings, with the manager allocating by drag and drop. Check which of these jobs your own system already does before you add a separate AI step.

A rule-check prompt, with a sample result

This prompt uses a variation on the run above: the journey before the hoist call is 15 minutes, and the partner carer's skills are listed.

You are checking a home care rota against written rules. Use
only the data below; if something is not stated, say "not stated".
For each visit, say OK or list the rule numbers it breaks and why.
Do not suggest who to allocate. Do not use names.

RULES: (1) required skills held; (2) double-ups overlap fully;
(3) travel gap at least the travel minutes listed;
(4) time-critical visits are not moved.

CARER K-07: skills = medication, hoist; availability 07:00-15:00.
VISITS:
2411-0346 08:12-08:42 medication prompt, travel before 9 min
2411-0347 08:51-09:36 double-up hoist, travel before 15 min,
partner carer K-12 (skills: medication only)

An illustrative response:

2411-0346: OK.
2411-0347: breaks rule 1 (partner carer K-12 does not hold hoist).
Travel rule 3: K-07 leaves 08:42, travel 15 min, arrives 08:57,
which is later than the 08:51 start. Breaks rule 3.

That is a good catch of two separate issues. What you would check: the model has read the travel line correctly (it should add 15 minutes to 08:42). Where figures drive a decision, check the arithmetic yourself or use a spreadsheet formula, since a language model can slip on times.

Filling a gap step by step

A gap arises, for instance when a carer calls in sick at 06:40 with five visits on her run. A stable routine keeps panic from wasting the first 20 minutes.

1. Mark the visits and sort by risk

Change the status of every affected visit to "gap" and sort them: time-critical medication first, then double-ups, then personal care, then welfare and social visits. Reassign or rearrange the lowest-risk visits last, and agree with the client or family if a visit must move.

2. Ask for candidates, not an answer

Visits needing cover (anonymised):
V1 07:30 single, personal care, hoist required, C-112
V2 08:12 single, medication prompt, C-087
V3 08:51 double-up hoist (other carer K-12 fixed), C-203

Available carers (codes, skills, current location, finishing time):
K-03 medication, hoist, finishes run at 07:20, 5 min away from C-112
K-09 medication, finishes 08:00, 20 min from C-087
K-15 hoist only, off until 08:30

Rank who could cover which visit, with reasons. Respect skills
and travel. Flag any visit where continuity is lost.
Do not allocate; the coordinator decides.

A useful answer ranks K-03 for V1 and V3 (hoist, near, but note V1 and V3 are an hour apart so check the intervening run), suggests K-09 for V2, and flags that C-112 and C-203 lose their regular carers. The coordinator then decides. The useful part is the list of trade-offs surfaced in seconds, not the ranking itself.

3. Draft the messages

Once the coordinator picks, ask the model for a short message to each carer and, if a time changes, to the client or family. Use placeholders and keep client information out of group chats.

Hi [first name], can you cover visit 2411-0345 (07:30-08:00)
this morning? It's a single call with a hoist. Please reply YES
or NO by 06:55. If you can't, tell me what you can do. Thanks,
[your name]

For the client, a short, honest note beats a vague one: who is coming, the new time if it has changed, and a number to ring. The wording should come from your agency, and a coordinator reads it before it is sent.

4. Close the loop and log it

Record who covered, any time change, and whether the client or family was told. If cover could not be found for a non-critical visit, the escalation steps in your policy apply, and that is a call for the manager, not a prompt.

Continuity, the rule most easily lost under pressure

When gaps are filled fast, continuity is the first thing sacrificed. Prevent that by storing it as data. Each client's record should hold a regular carer, an acceptable second and third, and any reason continuity matters strongly (for example dementia, distress with strangers or an intimate-care preference). When the model ranks candidates, tell it to prefer previous visitors and say when it cannot.

Each week, count the number of visits to clients flagged "regular carer" that were delivered by someone else. A small table makes the trend visible.

These figures are invented to show the format. A spike like week 3 points to a cause, such as two carers on leave in the same area, and gives the owner a reason to look at caregiver retention or recruitment rather than only at cover.

Keeping client data out of the wrong places

A rota is full of personal and health data. The safest habit is to keep identifiable fields in your care system and use codes in anything else. A message that says "C-112 at 07:30" is fine in a team chat. One that includes a name, address and "hoist for stroke patient" is not. If you plan to use any AI tool with rota data, check the tool's data terms and your own policies first, and anonymise as a default. The agency's wider plan for adding AI sensibly is in the 90-day plan for a home care agency owner.

Overnight, sleep-in and weekend variants

The one-visit-per-row template stretches to other patterns with small changes, and it is worth deciding them in advance instead of improvising on a Friday night.

• Waking nights: one row per client per night with a start and end time, and the checks required (for example hourly checks or two repositioning visits). The skills column matters as much as for days, and the carer's rest before and after is a rule.
• Sleep-in: a long row with a base address, the expected disturbance level and a flag for how any disturbances are recorded. How sleep-ins are paid and counted towards hours is a contract and employment matter. Ask your adviser, and put the answer in the rules sheet rather than in someone's memory.
• Weekends and bank holidays: copy the weekday template, change availability, and mark visits that are lower priority if staffing falls short. Decide which those are before the weekend, with the manager, not at 06:30 on Saturday.
• Short-notice new packages: a new client arriving from hospital needs a row set before the first visit. Add a "new package" status so it is visible and the first visit goes to someone experienced.

A week of hours, with the sums

Take an imaginary carer on a 30-hour contract who is offered extra shifts to cover gaps. The template should show the running total so nobody has to count in their head. These numbers are illustrative; your own limits come from contracts and employment advice.

Accepting Friday takes this carer to 34.5 against a 30-hour contract. That may be fine, or it may trigger overtime rules or a maximum-hours check. The coordinator sees it in the table before offering the shift. A model can do this addition, but a spreadsheet formula does it more reliably, so let the sheet total the hours and ask the model only to flag anyone over a threshold you give it.

Publishing the rota to carers

A perfect rota that carers cannot read in time does not help. Decide when the next week's rota is final (say Thursday 14:00), how carers receive it (an app, a shared view or a message with their own runs only), and how they confirm. AI can produce a short personal summary for each carer from their rows: start time, first client code, any double-ups and any visits where continuity differs from usual. Carers should see only their own runs. A message that includes other carers' runs or client details beyond need is a data problem, not a convenience.

For changes after publication, use a single place and a single wording style. "Change: visit 2411-0346 is now 08:30-09:00. Reply OK." A carer can take that in at a glance before setting off.

Client and family requests to change a visit

Requests to move a visit arrive by phone and often midweek. Treat each as a mini-gap with its own steps: check whether the visit is time-critical, check the carer's route, check whether the change affects another client, then confirm back. A short prompt can help the coordinator think of the knock-on effects. Give it the visit, the requested new time, and the neighbouring visits on the run (as codes) and ask what else would change. The answer should be a list of effects, such as "moves the next visit by 25 minutes, which breaks the travel rule". The coordinator decides, and tells the client plainly what can and cannot be done.

A monthly review of the rota itself

Once a month, review the rota as a system, not just as a daily task. Use five questions, each answered with a number from your records:

1. How many visits began as a gap, and how many were filled by a regular carer?
2. How many visits ran more than 15 minutes late, and on which runs?
3. How often did travel time on a run exceed the template's figure?
4. Which clients saw the most different carers in the month?
5. Which rule was broken most often, and was the rule or the practice wrong?

Ask the model to summarise the month's data in a table and name the three runs that need attention, but check its numbers against your system. The review is a place for the owner and coordinator to agree a change, perhaps a rewritten run or a recruitment push in one area. The monthly review is also where spot checks join in: patterns of lateness and rushed visits are things to look for when you observe care directly, as set out in the spot check template with AI summaries and action lists.

Coordinator review: the control that keeps this safe

Whatever the suggestions, a named coordinator should do a short review before each day's rota is final. A five-point review takes about ten minutes for a small agency:

1. Every visit has a carer or an explicit gap status.
2. Every double-up has two carers with the right skills and overlapping times.
3. Every time-critical visit sits within its window.
4. No carer is over their hours or short of a rest period.
5. Continuity exceptions are listed and the clients or families have been told where that matters.

Sign off in the system, with the time. If you have an out-of-hours phone line that creates last-minute changes, AI answering for home care agencies shows how to route those into your gap routine rather than a personal mobile.

Testing the whole routine before you rely on it

Run a rehearsal. Take a quiet day's rota, remove one carer at random, and time how long it takes the coordinator to produce a covered rota with and without the suggestion prompt. Check three things: did the final rota break any rule, did anyone lose a time-critical visit, and did the messages go to the right carers? Do it twice with different carers removed, and keep the prompts that worked in a shared document. If the model suggests something your rules forbid, add the missing rule to the prompt text rather than trusting yourself to spot it later.

The aim is not an automated rota. It is a rota where the rules are written, gaps are found early, cover suggestions arrive in seconds, and a coordinator still makes the call. That combination reduces the scramble without reducing responsibility.

Rota questions from care coordinators

Can AI build a care rota on its own?

It can propose one, but a coordinator should approve every change. Rotas depend on things a model cannot fully know: a carer's real travel, a client's changing needs, competence for specific tasks and last-minute sickness. Use AI for options and messages, not final allocation.

Should I use rota software or a spreadsheet?

A spreadsheet suits a very small agency with a handful of carers and stable visits. Once you pass roughly 15 to 20 carers or need live changes, dedicated care scheduling software usually earns its cost. Some include auto-assign or matching features, so test those before adding another tool.

How should travel time be handled?

Record it as its own line between visits, not as slack inside a visit, and base it on realistic journeys at the hour of travel. Check a sample of real journeys each month. If carers regularly arrive late, the travel figure is wrong, not the carer.

Can I paste client details into ChatGPT to find cover?

Avoid it. Use anonymised visit codes and general needs ('double-up morning call, hoist') rather than names, addresses or health detail. Check your agency's data rules and the tool's terms first, and keep identifiable information inside your care system.

Further reads

• Care Home Rota Template: Cover, Skill Mix and Gap Checks Using AI — The residential equivalent, with cover and skill-mix checks.
• Caregiver Recruiting Strategies: AI for Ads, Screening and Follow-Up — Fill structural gaps with recruitment, not only cover.
• Home Care Agency Software Pricing: What the AI Features Add — What care software costs and what its AI features add.
• Caregiver Onboarding Checklist: Pre-Start Checks and Induction With AI — Get new carers rota-ready faster and safely.
• Home Care Agency Employee Handbook: Draft With AI, Then Review — Put rota expectations in the handbook carers sign.
• Supported Living Software: Rotas, Support Plans and Where AI Fits — How supported-living software links rotas to support plans.
• Home Care Agency Policies and Procedures: A Review-First AI Method — Start from the regulator's standard, draft with AI, and put a named person through a ten-point review before any policy reaches a carer.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Birdie rostering page; CareLineLive rostering features page; general home care scheduling practice. Check your commissioner and regulator requirements, your carers' contracts and your care software vendor's own documentation.