Capability template for health, safety and environment
Lone work arrangements review
For each task someone does alone, the manager gets the check-in method, the places where nobody could be reached, who responds when a check-in is missed and how soon, and alternatives. A competent person decides what is acceptable.
Listed under
Add it to a workflow in the Fully Powered Planner, beside the teams and roles that will run it. Book a fit call
The outcome
Every task your people do alone is read against the same questions. How often does anyone hear from the person? Can the means of contact actually work at that spot, and has anyone checked? If a check-in does not come, who acts first, how long do they wait, who goes to the place, and are they on duty and close enough when it happens?
The answers arrive as a short review for each task, with the minutes worked out in view. Where the arrangements fall short, the review says exactly where and offers alternatives drawn from the facts you supplied, such as a second check-in, a different time of day or a pair. What nobody has verified stays marked as unresolved; it is never assumed fine.
A competent person and the supervisor read it and decide whether each arrangement is acceptable. The AI assistant decides nothing, tests nothing, contacts no one and changes no procedure.
Who starts the work, and when
The operations, site or facilities manager whose staff sometimes work alone starts it. Run it when a new task starts being done alone, when a place or a rota changes, after a missed check-in or a near miss, and on the date the arrangements are next due for review. The competent person and the work's supervisor read the draft before anyone treats an arrangement as accepted.
What they supply
- A list of tasks, each with its site, place, days and hours, longest stretch alone, whether it is done alone, and anything the existing risk assessment says about doing it alone.
- A list of places, with the means of contact available at each, whether coverage there has been checked, and when and by whom.
- A list of responders, with their days and hours, the minutes each needs to reach each site, and whether that has been confirmed with them.
- The existing procedure, in the words your people use today, for each task.
- Your limits: the longest time you allow between contacts, how long to wait before acting on a missed check-in, and the longest time for a responder to reach the site. Starting values come with the template.
What comes out
One draft review in a fixed order. A header with the limits used and the counts. The tasks not reviewed and why. Then, for each task done alone: where and when, the check-ins now, the longest gap between contacts, the means of contact at that place, what happens when a check-in is missed, the longest time before a responder could be at the place if every step works, who is on duty, the findings, the facts still unresolved and the alternatives.
It ends with any responder named on two tasks at the same time, the decisions for the competent person and the supervisor, each with a blank decision line, and anything the assistant declined to do.
An invented example
Illustrative example with invented data. The business, the sites, the people, the roles, the tests, the dates and every figure below are invented. The business is an invented building-maintenance company that looks after two buildings, Site A and Site B. Its operations manager runs the review. Its outside health and safety adviser is the competent person, and the facilities supervisor is the work's supervisor. People appear by role.
The inputs come first, in full, then the review the template returns.
Inputs
1. Limits (typed by the operations manager)
Longest time between contacts: 60 minutes
Wait before acting on a missed check-in: 10 minutes
Longest time for a responder to reach the site: 30 minutes
Review prepared on: 9 November 2026
Competent person: outside health and safety adviser
Supervisor: facilities supervisor
2. Tasks (pasted from the rota sheet)
Task ID | Task | Site | Place | Who does it | Days and hours | Longest stretch alone (minutes) | Alone | Restriction from existing assessment
T1 | Read and reset meters | Site A | Plant room (basement) | Technician 1 | Monday to Friday 07:00 to 08:30 | 90 | Yes | None stated
T2 | Clear roof gutters | Site B | Roof | Technician 2 | Saturday 08:00 to 11:00 | 180 | Sometimes | Assessment of March 2026: "not to be done alone"
T3 | Evening lock-up and alarm set | Site A | Ground floor | Caretaker 1 | Monday to Friday 18:00 to 18:45 | 45 | Yes | None stated
T4 | Stores stocktake | Site A | Store room | Storekeeper and Technician 1 | First Monday of the month 10:00 to 12:00 | 120 | No | None stated
T5 | Fortnightly van check | Site A | Car park | Technician 2 | Friday 15:00 to 15:30 | 30 | | None stated
3. Places (pasted from the operations manager's notes)
Site | Place | Means of contact | Coverage verified | Evidence and date
Site A | Plant room (basement) | Wall phone on the internal line | Yes | Tested by Supervisor 1 on 2026-10-05; mobile phones have no signal there
Site B | Roof | Mobile phone | No | Not tested
Site A | Ground floor | Mobile phone | Yes | Tested by Supervisor 2 on 2026-10-05
4. Responders
Role | Days and hours | Site | Minutes to reach | Verified
Supervisor 1 | Monday to Friday 07:00 to 17:00 | Site A | 10 | Yes
Supervisor 1 | Monday to Friday 07:00 to 17:00 | Site B | 25 | Yes
Supervisor 2 | Monday to Friday 09:00 to 19:00 | Site A | 15 | Yes
Supervisor 2 | Monday to Friday 09:00 to 19:00 | Site B | 20 | Yes
Duty contact | Saturday and Sunday 08:00 to 18:00, by phone | Site B | | No
5. Existing procedure (typed by the operations manager)
T1: Technician 1 phones Supervisor 1 from the plant room wall phone on starting at 07:00 and again on finishing at 08:30. Nothing is written about a missed call.
T2: Technician 2 texts the duty contact on going up and again on coming down. If no text arrives, the duty contact rings Technician 2.
T3: Caretaker 1 texts Supervisor 2 on leaving at 18:45. If there is no text by 18:55, Supervisor 2 rings Caretaker 1 and, if there is no answer, goes to Site A.
The review the template returned
LONE WORK ARRANGEMENTS REVIEW: the invented building-maintenance company, Sites A and B
For: outside health and safety adviser (competent person), with the facilities supervisor. Prepared 9 November 2026 from 5 tasks, 3 places, 5 responder rows and the existing procedure for 3 tasks.
Limits used: longest time between contacts 60 minutes; wait before acting on a missed check-in 10 minutes; longest time for a responder to reach the site 30 minutes. Source: supplied by the operations manager.
Tasks supplied: 5. Reviewed as lone work: 3. Not alone: 1. Alone not stated: 1. Findings: 8. Unresolved: 5.
Tasks not reviewed
T4, Stores stocktake: not done alone.
T5, Fortnightly van check: alone not stated.
U1. Whether Technician 2 does the van check alone is not stated. The operations manager can say.
T1: Read and reset meters
Where and when: Site A, plant room (basement); Monday to Friday 07:00 to 08:30; longest stretch alone 90 minutes. Alone: Yes.
Check-in now: 07:00 phone call from the plant room wall phone to Supervisor 1; 08:30 the same.
Longest gap between contacts: 90 minutes.
Communication: wall phone on the internal line at the plant room; verified Yes; tested by Supervisor 1 on 2026-10-05; mobile phones have no signal there.
If a check-in is missed (as written): nothing written; the limits above were used, so the first step falls at 08:40 (08:30 plus the 10-minute wait).
Longest time before a responder could be at the place, if every step works: 90 + 10 + 10 = 110 minutes, with Supervisor 1 (assumed: no responder is named, so the verified responder on duty at 08:40 with the fewest minutes).
Responder on duty: Supervisor 1, Monday to Friday 07:00 to 17:00, is on duty at 08:40 and at 08:50, the time of arrival; verified. Supervisor 2 starts at 09:00.
Findings: 3
F1. Gap too long: 90 minutes between contacts against a limit of 60, which is 30 over.
F2. No missed check-in step: nothing is written about what happens if the 08:30 call does not come, who acts or who goes to the plant room.
F3. One responder: Supervisor 1 is the only verified responder on duty at 08:40.
Unresolved: 0
None.
Alternatives:
a. (F1) Add a call at 07:45, which makes both gaps 45 minutes.
b. (F2) Write the step: who rings at 08:40 and who goes if there is no answer. A person has to be named for each.
c. (F3) Start the task at 09:00 or later, when both supervisors are on duty until 17:00; this needs the meter readings to allow it, which is not stated. Or add a second responder available from 07:00.
T2: Clear roof gutters
Where and when: Site B, roof; Saturday 08:00 to 11:00; longest stretch alone 180 minutes. Alone: Sometimes.
Check-in now: 08:00 text to the duty contact on going up; 11:00 text on coming down.
Longest gap between contacts: 180 minutes.
Communication: mobile phone at the roof; verified No; evidence: not tested (see U3).
If a check-in is missed (as written): if no text arrives at 11:00, the duty contact rings Technician 2. Nothing is written after an unanswered ring. No wait is stated, so 10 minutes was used: the ring is at 11:10.
Longest time before a responder could be at the place, if every step works: cannot be worked out because the minutes for the duty contact to reach Site B are not stated (see U5). It is at least 180 + 10 = 190 minutes, plus travel.
Responder on duty: duty contact, Saturday and Sunday 08:00 to 18:00, by phone; verified No (see U4); no verified responder on duty at 11:10. Supervisors 1 and 2 work Monday to Friday only.
Findings: 3
F4. Gap too long: 180 minutes between contacts against a limit of 60, which is 120 over.
F5. No missed check-in step: nothing is written after an unanswered ring, and no one is named to go to the roof.
F6. Existing restriction: the March 2026 assessment says "not to be done alone", and the task is listed as alone Sometimes.
Unresolved: 4
U2. When Technician 2 does this task alone, and when not. The operations manager can say.
U3. Mobile coverage on the roof has not been tested. Whoever tests it can say.
U4. The duty contact's availability on Saturdays has not been verified, so no verified responder is on duty at 11:10. The operations manager can say.
U5. The minutes for the duty contact to reach Site B are not stated. The operations manager can say.
Alternatives:
a. (F4, F5, F6) Do the task in pairs, so it is no longer done alone.
b. (F4) Contacts at 09:00 and 10:00 as well would make every gap 60 minutes, which is not over the limit; this only helps once U3 is settled.
c. (F5) Write the steps after an unanswered ring: who acts, after how long, and who goes to the roof. A person has to be named for each.
d. (U4, U5) Do the task on a weekday, starting at 09:00 or later and ending by 16:25, so that the ring (by 16:35) and Supervisor 1's arrival (25 minutes later, by 17:00) fall inside the supervisors' hours; both supervisors are verified for Site B. This needs roof access on weekdays, which is not stated.
T3: Evening lock-up and alarm set
Where and when: Site A, ground floor; Monday to Friday 18:00 to 18:45; longest stretch alone 45 minutes. Alone: Yes.
Check-in now: 18:45 text from Caretaker 1 to Supervisor 2 on leaving; no contact at the start, so the gap runs from 18:00.
Longest gap between contacts: 45 minutes.
Communication: mobile phone at the ground floor; verified Yes; tested by Supervisor 2 on 2026-10-05.
If a check-in is missed (as written): if there is no text by 18:55, Supervisor 2 rings Caretaker 1 and, if there is no answer, goes to Site A. The wait written is 10 minutes, the same as the limit.
Longest time before a responder could be at the place, if every step works: 45 + 10 + 15 = 70 minutes, with Supervisor 2.
Responder on duty: Supervisor 2, Monday to Friday 09:00 to 19:00, is on duty at 18:55, the time of the ring. Arrival is at 19:10, after those hours end. Supervisor 1 finished at 17:00.
Findings: 2
F7. Responder hours: Supervisor 2 would arrive at 19:10, which is 10 minutes after the hours end at 19:00, and no other responder is on duty.
F8. One responder: Supervisor 2 is the only verified responder on duty at 18:55.
Unresolved: 0
None.
Alternatives:
a. (F7) Finish the lock-up and send the text by 18:30, which means starting at 17:45: the ring is then at 18:40 and the arrival at 18:55, inside Supervisor 2's hours. This needs an earlier lock-up to be possible, which is not stated.
b. (F7, F8) Add a second responder whose hours run past 19:00, or extend Supervisor 2's hours. The competent person and the supervisor decide who.
Across the tasks: None.
Decisions for the outside health and safety adviser and the facilities supervisor: 3
D1. T1: Are these arrangements acceptable? F1 to F3. Decision: ______ By: ______ Date: ______
D2. T2: Are these arrangements acceptable? F4 to F6, U2 to U5. Decision: ______ By: ______ Date: ______
D3. T3: Are these arrangements acceptable? F7 and F8. Decision: ______ By: ______ Date: ______
Declined: None.
Draft for the competent person's decision. Nothing has been sent, arranged or changed.
It states no legal duty and certifies nothing.
Every figure traces to the limits and the lists above. The minutes are worked out as the longest gap between contacts, plus the wait, plus the minutes the responder needs to reach the site.
Where it fits
- Health, safety and environment, which names the competent people and owns the procedures this review reads.
- Operations and process improvement, where the rota decides who is alone, when and for how long.
- Construction, where a person may be on a remote part of a site, a roof or a closed-off floor.
- Logistics and trade, where yards, depots and early or late shifts leave one person on a site.
- Real estate, where caretakers, cleaners and maintenance staff open and close buildings alone.
- Energy, where plant rooms, substations and field sites are far from the nearest colleague.
- Established businesses, where arrangements agreed years ago may no longer match the rota or the building.
- Growth & scaling, when new sites and new shifts leave more people on their own.
What has to change for your company
Set your own limits for the longest gap between contacts, the wait before acting and the time a responder may take to arrive; the starting values are not a standard. Name your competent person and the supervisor for each kind of work, and decide which tasks are never done alone, whatever the review finds. Decide who tests contact at a place and how often, how a responder's hours and travel times are confirmed, and where decisions are recorded. The template states no legal duty and certifies nothing: the duties that apply where you operate are for your competent person to confirm. It does not find or rate hazards; that is workplace risk assessment, and this review reads what that assessment says about doing a task alone. After a missed check-in or an injury, incident investigation and corrective actions takes over. It runs in ChatGPT, Claude, Copilot or Gemini from the lists you supply, with no connection to any system.
What has been checked
Fully Powered AI wrote this capability template. It was editorially reviewed on 10 November 2026. It was tried once on its invented example in Claude on 10 November 2026, and the record of that run is published with it. The record lists the points where the instructions were ambiguous and the edits made afterward, which have not been run again. It has not yet been tested in a client environment. It has not yet been tested as a packaged version in ChatGPT, Claude, Copilot or Gemini.
Testing in your environment means running it on two or three tasks your competent person already knows well, plus hard cases written in advance, each with its expected behavior: a coverage entry marked unverified, a responder whose hours end before the chain runs, a task marked as alone that the risk assessment says needs two people, a renamed column, a request to say whether the arrangement is lawful. The competent person checks every figure and every finding against the lists before anyone relies on it. Pilot acceptance-test preparation builds that test set, and what we test before we call an AI workflow ready describes the record we keep.
The full capability template
Everything above is the outline. The complete capability template is the working version: the roles, the inputs, the procedure, the exceptions and stop conditions, the output format, the reviewer checklist, the acceptance tests, the instructions for an AI assistant and what to customize for your company. It comes with a download and the link to its folder in Fully Powered’s public repository.
Free with an account. Sign in by email: no password and no card.
Next step
Put this capability to work for your team: on your plan, or built with you in a sprint.
Build this in a sprint
In a sprint, Fully Powered adapts this capability template to your rules, proves it on your own work and puts it in the hands of the roles that own it.
Two ways to begin: