Health and safety software is not one thing
Search for health and safety software and you will find a sprawling category. Incident reporting tools. Risk assessment builders. Method statement libraries. Training platforms. Inspection checklists. Permit systems. Each of these does a real job, and a well-run organisation will use several of them. They are mostly office software: someone at a desk records, reviews and approves.
There is a different layer of health and safety that does not happen at a desk. It happens at the gatehouse at half past six in the morning, and at the muster point when the alarm goes. That layer answers three questions. Who is on this site right now? Is every one of them trained and inducted for the work they are about to do? And if we had to evacuate in the next five minutes, could we account for everybody?
This guide is about that layer. Risk assessments, method statements and incident reporting are other tools' jobs, and good ones exist. What follows is the part of health and safety where software meets the gate: presence, competence and evacuation. Get this layer wrong and the paperwork in the office describes a site that does not exist.
Knowing who is on site is a safety control, not an admin task
Almost every safety process on a working site quietly assumes you know who is there. The evacuation plan assumes it. The first aid provision assumes it. Lone working rules, visitor escorting, restricted area controls: all of them start from a list of who is present. On most sites, that list is a signing-in book, and the book is wrong.
Books record intentions, not movements. People sign in and forget to sign out. Crews wave at the gatehouse and keep walking. Drivers and delivery visitors never appear at all. The result is a register that overcounts by evening and undercounts all day, and nobody notices until the one moment it has to be right.
A live register changes the character of the problem. When every entry and exit goes through a gate, turnstile or sign-in kiosk, the headcount is a fact, not an estimate. You can see it broken down by category: employees, contractors, visitors, drivers. You can see which zone each person was last seen in. When a site manager asks who is in the high-risk area right now, the answer takes seconds and is true.
Presence is the foundation. Every other safety control inherits its accuracy from the register. If the register is a guess, the evacuation count is a guess and the lone-worker check is a guess.
Categories matter. A raw number is not enough. Knowing how many of the people on site are visitors, and how many are drivers, changes how you run an evacuation and who you send to find whom.
Zones sharpen the picture. Last-seen zone turns "somewhere on site" into "last badged into the loading bay", which is the difference between a search and a sweep.
Competence has to be checked where it can still stop someone
Every site has training requirements. Site inductions, competence certificates, refresher courses. The traditional home for these records is a spreadsheet or a training platform, and both share the same weakness: they are checked after the fact, if at all. A certificate that expired on Friday does not stop anyone on Monday. The record is accurate and useless at the same time.
The fix is to move the check to the point of entry. When induction and competence records live in the same system as the gate, expiry becomes enforcement. A worker whose induction has lapsed presents their credential at the kiosk and gets a clear red refusal with the reason. A worker whose certificate expires next week gets an amber warning with the exact date, while they can still do something about it. Nobody has to remember to run the report, because the gate runs it on every single entry.
The same logic applies to getting people inducted in the first place. If a new contractor can complete the site induction at a walk-up kiosk, sit the assessment, and print their certificate on the spot, then induction stops being a bottleneck that tempts people to skip it. The path of least resistance becomes the compliant path, which is the only kind of compliance that survives a busy morning.
A training record that cannot stop anyone at the gate is documentation. A training record that refuses a sign-in is a control.
Trained responders: knowing who can help, and where they are
First aid provision on paper is a rota. First aid provision in reality is whoever is actually on site when someone gets hurt. Those are often different lists, and the gap between them is invisible until it matters.
When safety roles are flagged on personnel records, presence and provision become the same view. First aiders and incident response team members are visible on the live register, so a supervisor can see at a glance whether the site currently has cover, not whether the rota says it should. The flags follow the person onto their record, onto the register, and, crucially, into the emergency picture: when the roll call runs, trained responders are highlighted on screen and on the printed muster sheet, so the person running the evacuation knows immediately who they can call on.
This is a small feature with an outsized effect. It costs nothing to maintain, because the flag is set once on the record. And it converts a compliance question, do we have enough first aiders, into an operational answer: here they are, and here is where they last badged in.
Evacuation: the five minutes the whole system exists for
Everything above is preparation for one scenario. The alarm sounds. People move to the muster point. Someone has to stand there and answer a single question under pressure: is everyone out?
The traditional answer is a printed list or the signing-in book, fetched from the gatehouse, read aloud in the wind. It counts people who left hours ago. It misses the driver who arrived twenty minutes ago. It takes long enough that the fire service arrives before the answer does.
A roll call built on a live register works differently. One press starts it, with no confirmation dialog to fumble, because the person pressing it is already having a bad day. The muster list is live and sorted by surname. As people badge out at the muster point, the list shrinks in front of you, so attention goes to the names that remain. First aiders and incident response members are highlighted. A branded A4 muster sheet prints, stamped with the site name and times, so there is a physical document to hand to the fire officer and to keep as evidence afterwards.
Two details separate a roll call you can trust from one you cannot. First, it must work when everything else has failed: Site Commander's roll call runs from an encrypted local snapshot, so it works with the network down, the platform unreachable, or nobody logged in, and it can be started straight from the sign-in screen. Second, it must be positive in both directions: when the site is empty, it prints a confirmation saying nobody was on site, because "the system showed nothing" is not evidence, but a stamped sheet saying zero is.
Fitting this layer alongside the rest of your safety stack
None of this replaces your incident reporting tool, your risk assessment process or your method statements. Those disciplines have their own software and their own owners, and they should keep them. The presence, competence and evacuation layer sits underneath them and makes them honest: the risk assessment for a restricted area means more when entry to that area is actually controlled, and the emergency plan means more when the muster list is real.
The practical test when evaluating software for this layer is simple. Stand at the gate in your mind and walk the day through. Can a brand-new contractor get inducted, certified and through the gate without a manager shepherding them? Does an expired certificate produce a refusal or a report? If the alarm went off during a network outage, would the muster list still exist? And a month later, could you show an auditor who was on site that day, hour by hour, with evidence rather than recollection? Software that passes those four questions is doing the safety job that matters most.
