RAMS that hold up.

There are two kinds of risk assessment and method statement on a UK construction site. One is a document that exists so that somebody can say a document exists. The other is a document that tells a specific gang how to do a specific job on a specific day without hurting anybody. An inspector can tell them apart in about forty seconds, and so can a jury.

The generic RAMS problem

Almost every firm that has ever been asked for RAMS has a folder of them. They were bought as a pack, or inherited from a previous employer, or written once for a job in 2019 and renamed since. They are usually well formatted. They usually cover the right hazards in the abstract. And they usually fall over at the first question anybody asks.

The question is not "do you have a method statement". It is "which of these controls applies to the work happening behind me right now, and who told that operative about it". A generic document cannot answer it, because it was not written about that work. It was written about work of that type.

This matters more than it used to. Under regulation 3 of the Management of Health and Safety at Work Regulations 1999, an employer must make a suitable and sufficient assessment of the risks. Those two words are doing all the work. Suitable and sufficient is not a formatting standard. It is a question about whether the assessment engaged with the actual circumstances: this site, this sequence, this access, these people, this weather, this neighbouring occupier.

An inspector is not auditing your template. They are testing whether anybody thought about this job.

What "specific" actually means

Specificity is not achieved by adding the site address to a header. In practice, the parts of a RAMS that separate a real one from a laminated one are these:

• The sequence is the real sequence. A method statement is a description of the order in which work will happen. If the order in the document is not the order the gang will actually work in, the document is fiction, and every control hung off it is hung off nothing.
• Access and egress are described as they are on the day. Not "safe access will be provided". Which route, past what, shared with whom, and what changes when the scaffold moves on Thursday.
• Interfaces with other trades are named. Most site injuries are not caused by a trade doing its own work badly. They are caused by two trades meeting. The RAMS that holds up says who else is in that area and what the arrangement is.
• The controls are things somebody does. "Operatives to be vigilant" is not a control. "Exclusion zone barriered at 4m, banksman posted at the corner, keys held by the supervisor" is a control, because you can walk out and see whether it is true.
• Residual risk is stated honestly. An assessment that reduces everything to "low" after controls is not reassuring, it is unbelievable. Some work is dangerous after you have done everything reasonable, and saying so is what makes the rest of the document credible.

The briefing is the document

Here is the part most firms get wrong, and it is not a paperwork failure, it is a sequencing failure.

A method statement that has not been briefed to the people doing the work has not been implemented. It has been filed. If an incident happens, the question is not whether the document existed but whether the operative knew what was in it, and the evidence for that is a signature against a version, on a date, before the work started.

Three things go wrong here repeatedly:

1. The briefing happens after the work. Somebody realises on Friday that the sheet has no signatures and walks it round. Everyone signs. The record now says the gang was briefed, and it is worthless, because the timestamps are wrong and anybody checking will see it.
2. The version briefed is not the version filed. The RAMS was revised on Tuesday because the crane position changed. The gang signed Monday's. Nobody notices until it matters.
3. The people who joined on Wednesday never got it. Labour moves. The original briefing covered the original gang. The two agency operatives who arrived mid-week are working to a method statement they have never seen.

If you can produce the document but not the signatures against that version, you have evidence of intent and no evidence of control.

When it has to be reviewed

A RAMS is not a one-time deliverable. It has to be revisited when the thing it describes changes, and the triggers are more ordinary than people expect:

• The sequence changes, or the programme compresses and two operations that were consecutive become concurrent.
• The access changes: scaffold struck, hoist out of service, a new route through an occupied building.
• Plant is substituted. A different telehandler is not the same telehandler.
• Something happened. A near miss in that activity is the clearest possible signal that the assessment did not capture reality.
• The people change materially, in number or in competence.

The practical test is simple: if you would have written the document differently had you known, it needs reissuing and rebriefing. And a reissue means a new version with its own signatures, not a pen amendment on the copy in the cabin.

Dynamic assessment, and its limits

Site supervisors make risk decisions constantly that no document anticipated. That is legitimate and unavoidable, and formalising it as a dynamic or point-of-work assessment is good practice: a short, recorded judgement made at the face by the person in charge of it.

What it is not is a substitute. A dynamic assessment answers "the conditions in front of me differ from what was planned, here is what I am doing about it". It cannot answer "nobody planned this at all". Firms that lean on point-of-work assessment to cover an absence of method statements find that out at the worst possible time, because the record shows a supervisor improvising rather than a business managing.

What an audit actually asks for

When the request comes, whether from an HSE inspector, a principal contractor's compliance team, a SMAS or CHAS assessor, or the other side's solicitor, it is rarely "send us your RAMS". It is a chain:

1. The assessment for this activity, in the version current on this date.
2. Evidence that the people who did the work were briefed on that version, before it started.
3. Evidence that the controls it specifies were in place, which usually means an inspection record or a photograph.
4. Evidence of what happened when something changed: the revision, and the rebriefing.

Every one of those is a link. The chain is only as good as the weakest, and in most firms the weakest is number two, because signatures live on paper in a cabin that gets cleared out at the end of the job.

If you are starting from a folder of generic ones

Do not try to rewrite the library. It is a year of work and it will not get finished. Do this instead:

1. List the activities you actually carry out. Most contractors of this size have somewhere between twelve and thirty. That list is the real scope.
2. Take the three highest-consequence ones and write those properly, with the sequence and the interfaces in them. Not the three most common. The three that would end the business.
3. Fix the briefing record before you fix the remaining documents. A mediocre assessment that is demonstrably briefed to the right people on the right day is worth more at audit than an excellent one nobody signed.
4. Then work down the list, replacing generic ones as each activity comes up on a live job, so the specifics are in front of you while you write.

That order is deliberate. It gets you defensible fastest, which is not the same as getting you complete fastest.