Most working-alone plans skip the only part that matters.

Why the plan is the deliverable, not the app

It is tempting to treat lone-worker safety as a purchasing decision, because a purchase is finite and a plan is not. But every tool in this category does the same three mechanical things — share a position, take a check-in, escalate a missed one — and none of them decides how long silence is acceptable, who responds, or what that person actually does.

Those decisions are the plan. Made in advance they take an hour; made during an incident they get made by whoever answers the phone, under pressure, without authority, and usually by waiting longer than they should because nobody wants to be the person who overreacted.

A plan is also what makes a tool usable. A check-in timer with no agreed interval is a guess. An escalation with no named recipient goes to nobody in particular. The app is the mechanism; the plan is what the mechanism is for.

1. Identify who works alone, and when

Start with the list, because it is always longer than expected and the missing entries are usually the intermittent ones — someone alone for forty minutes, not someone alone all week.

• Go through a normal week rather than a job description. Who arrives first, who leaves last, who splits off from a crew, who drives between sites, who does the one call at the end of the run?
• Include the periods, not just the people. Most lone working is a slice of a day, and that slice is what you are planning for.
• Include subcontractors and casual staff. They are frequently the ones nobody is expecting to hear from.
• Note who else is nearby and whether they would actually notice — being in the same building as other people is not the same as being accounted for by them.

2. Assess what could go wrong — and how long it would go unnoticed

This is an ordinary risk assessment with one extra column, and that column is the one doing the work. For each plausible event, write down how long it would take for anybody to realise.

The events worth planning for are usually mundane rather than dramatic: a fall, a vehicle that will not start, a medical episode, an allergic reaction, a slip into water, getting lost, a dog, an aggressive member of the public, a hand caught in something. What makes them serious is that nobody saw.

• For each event, ask: could this person still call for help themselves? Many of the worst cases are precisely the ones where they cannot.
• Then ask: if they cannot, how long until someone notices? That number, not the hazard rating, is what sets your check-in interval.
• Then ask: once noticed, how long until help physically arrives? Remote work stretches this out and should shorten the interval to compensate.
• Note anything that makes the person more exposed — working at height, near water, with machinery, with the public, or with a known medical condition they have chosen to disclose.
• Consider whether working alone should simply not happen for a specific task. That is a legitimate outcome of an assessment, and sometimes the correct one.

3. Decide the contact method and the interval

The CCOHS Working Alone fact sheet frames this well: decide whether a verbal or real-time check-in is adequate, or whether the worker must be accounted for by a visual check. Not every situation is satisfied by a message.

The interval follows from the assessment rather than from convenience. If the plausible worst case is survivable unattended for an hour, a four-hour interval is decoration.

• Write the interval down as a number, and write down what it was derived from.
• Vary it by task rather than setting one number for everything — a drive between sites and a climb are not the same exposure.
• Check that the method works where the work happens. A check-in that depends on cell coverage the site does not have is a plan that fails silently.
• Decide what the worker does if they cannot check in but are fine — because that will happen far more often than an actual emergency, and an unhandled false alarm is what teaches everyone to ignore the system.
• Agree a duress or code word, as CCOHS suggests, for the case where the worker can communicate but cannot speak freely.

4. Name a primary and a backup contact

Name people, not roles. "The office" is not a contact, and neither is "the supervisor" when the supervisor is on annual leave.

• A primary contact who is expected to notice, and who knows they are expected to notice.
• A backup contact for when the primary is unreachable, asleep, driving, or on the same job site with the same problem.
• Both should know the plan, not just be listed in it. Being an emergency contact is a role somebody has to accept.
• Both need working notifications on whatever channel the plan uses, and it is worth testing that once rather than assuming it.
• Record who is on duty when. A plan that assumes somebody is watching at 5.30am should say who.

5. Write the missed check-in procedure — the part most plans omit

This is the section that gets skipped, and it is the entire point of the exercise. When a check-in does not arrive, the person who notices needs a script, because in the moment the default human behaviour is to wait and to assume the best.

It should be short enough to follow from a phone in a car park:

• How long after the missed check-in before this counts as a problem — a number, agreed in advance, not a judgement call.
• Step one: try to contact them. Say by which methods, and how many attempts, and over how long.
• Step two: contact the backup, or anyone working near them, and share the last known position.
• Step three: physically go, if that is safe and sensible — and say who goes, and who does not go alone themselves.
• Step four: call emergency services. State the trigger for this explicitly, because the failure mode is always waiting too long. Nobody has ever been criticised for the call that turned out to be unnecessary.
• Have the information ready: last known position and the time it was recorded, the vehicle, what they were doing, what they were carrying, and any medical information they consented to share.
• Say who is authorised to make each decision, so nobody is waiting for permission.

6. Record, train, and rehearse

A plan nobody has practised is a document, not a control. The rehearsal does not need to be elaborate — running one deliberate missed check-in end to end will reveal more than another review meeting.

• Keep a log of check-ins and missed check-ins. CCOHS's example procedure includes a written contact log, and it is also how you find out that your interval is unrealistic.
• Brief every new person, including subcontractors, before the first time they are alone rather than at the end of induction week.
• Test it once: have somebody deliberately miss a check-in without warning the responder, and see what actually happens.
• Review after any incident, near miss, or false alarm — and treat a repeated false alarm as a defect in the plan rather than as an annoyance.
• Review when the work changes: a new site, a new season, a new shift pattern, or a new person changes the assessment.

What the authorities publish

Requirements differ by country, province or state, and industry — some jurisdictions have specific working-alone regulations and others rely on general duties. Nothing here is legal advice and no template substitutes for your own regulator's guidance.

These three are the ones worth reading in full before writing anything:

• UK Health and Safety Executive — Protecting lone workers: how to manage the risks of working alone, INDG73(rev4). Written for employers and for anyone engaging lone workers as contractors. hse.gov.uk/pubns/indg73.htm
• Canadian Centre for Occupational Health and Safety — Working Alone - General, OSH Answers fact sheet. Contains a worked example of a check-in procedure, including daily work plans, primary and backup contacts, a written contact log, code words, and a defined response to a missed check-in. ccohs.ca/oshanswers/hsprograms/workingalone.html
• US OSHA — 29 CFR 1915.84, titled Working alone. A sector-specific rule for shipyard employment requiring the employer to account for each employee working alone by sight or verbal communication at intervals appropriate to the assignment, and again at the end of the assignment or the shift. osha.gov/laws-regs/regulations/standardnumber/1915/1915.84

Where a tool like STALKR fits in the plan

STALKR can support elements of a working-alone safety programme. It is worth being precise about which elements, because the gap between what a plan needs and what an app supplies is where programmes quietly fail.

What it carries, what it does not, and what it does not have at all:

• The check-in itself — a one-time timer for a task, or a repeating dead-man switch for a shift.
• The escalation — server-side push and email to your Crew and the recipients you chose, carrying the last known location and the time. It never goes out as an SMS, and it never reaches emergency services.
• The position — live Crew location, labelled honestly as live, stale, or unknown, so a responder knows how old the information is.
• Zones around a yard, a site, or a hazard, alerting the Crew on entry and exit.
• An in-app history of Crew events, check-ins, Zone transitions, and SOS incidents, so an incident can be reconstructed afterwards.
• What it does not carry: your hazard assessment, your interval, your training, your decision about who responds, and your obligation to make that decision.
• What it does not have at all: a supervisor or web dashboard, a professional or 24-7 monitoring centre, man-down or fall detection, configurable escalation policies, compliance export or reporting, and any regulatory certification. It is an iOS TestFlight beta with no Android release.

Related NAVTRL Systems