How SIRV AI can support Martyn’s Law readiness
Martyn’s Law makes preparedness a practical operational responsibility. The test is not whether a procedure exists, but whether people can use it when it matters.
The Terrorism (Protection of Premises) Act 2025, commonly known as Martyn’s Law, introduces requirements for qualifying premises and events. The precise duties depend on the applicable tier and circumstances. Organisations should rely on official guidance to determine their obligations.
For safety, security and resilience teams, the everyday work includes locating approved procedures, checking whether plans are workable, identifying gaps, preparing exercises and retaining a record of what changed.
From written procedures to operational readiness
A policy may be technically complete but difficult to execute. A lockdown procedure might exist, yet staff may not know who can initiate it, how the instruction reaches contractors, or what happens when the usual communications channel fails.
The operational question is therefore broader than document compliance: are arrangements appropriate to the site, understood by the people who use them, and revisited as circumstances change?
Readiness work is often fragmented. Procedures, site plans, training records, incident reports and exercise findings may be held by different teams. This makes it harder to connect a known weakness with the action intended to address it.
Where SIRV AI can help
SIRV AI is designed to support bounded operational workflows using approved sources, traceable outputs and defined review points. Applied to Martyn’s Law readiness, useful areas include:
Procedure access. Retrieve the relevant approved guidance for a particular site, task or situation, rather than relying on memory or searching disconnected folders.
Document evaluation. Compare plans, procedures and submissions against specified requirements, and flag omissions, unclear ownership or contradictions for human assessment.
Intake and routing. Help organise incoming reports, exercise observations and requests so the right people can assess them.
Operational analysis. Bring together previous incidents, exercise findings and review records to make recurring weaknesses more visible.
Follow-up and retained learning. Keep an auditable record of decisions, actions, owners and outcomes so later reviews can build on what has already been learned.
| Work | Without a connected workflow | With controlled AI support |
|---|---|---|
| Find guidance | Search folders and ask colleagues which version applies | Retrieve approved, relevant material with source references |
| Review plans | Manual checks vary by reviewer and available time | Consistent preliminary checks against defined criteria, with human verification |
| Identify gaps | Issues remain dispersed across emails and documents | Surface missing measures, assumptions and unclear ownership for review |
| Track actions | Follow-up depends on separate notes and reminders | Link findings, decisions, assigned actions and completion evidence |
| Reuse lessons | Past exercises and reviews are difficult to locate | Bring prior findings into the next assessment |
This is a qualitative operating-model comparison, not measured performance data or a guarantee of outcomes.
What this means in practice
A venue reviews its lockdown arrangements before an exercise. The current procedure specifies who can authorise a lockdown, but the supporting communication plan does not clearly cover temporary staff.
An AI-assisted document check flags the mismatch and points to the relevant source passages. The security lead confirms the gap, consults the appropriate stakeholders, approves a revised arrangement and records the change. The next exercise tests whether the instruction reaches the intended groups.
The important result is not the AI flag. It is the verified change, the accountable decision and the evidence that the revised arrangement was tested.
Why the control points matter
AI-generated output can be incomplete, incorrect or overly confident. For consequential work, a plausible recommendation is not sufficient evidence that a measure is appropriate.
Teams need clear boundaries around which sources can be used, how findings are checked, who can approve changes, and how unresolved issues are escalated. Human review must be a genuine decision point rather than a routine sign-off.
Nor does using SIRV AI establish legal compliance. Duties under Martyn’s Law remain with the relevant responsible persons and organisations. Technology can support preparation and record-keeping, but it cannot transfer statutory accountability.
Start with one defined workflow
A practical starting point is a single site or a single readiness activity, such as reviewing a procedure against agreed criteria or reconciling exercise findings with outstanding actions.
Define the approved sources, the required outputs, the person responsible for checking them, and the measures of success. Those measures might include review time, valid gaps identified, closure of actions and the quality of the evidence trail.
This provides a testable basis for deciding whether the workflow improves readiness, rather than treating AI adoption itself as evidence of progress.
Sources and further reading
- Terrorism (Protection of Premises) Act 2025, legislation.gov.uk.
- ProtectUK: Martyn’s Law, official guidance and updates.
- SIRV AI, workflow and control approach.
Frequently asked questions
Does Martyn’s Law require every venue to take the same measures?
No. Applicability and duties depend on the premises or event and the relevant statutory tier. Arrangements should reflect the specific site and official requirements.
Can AI determine whether a venue complies with Martyn’s Law?
No. AI can assist with information retrieval and structured review, but it cannot replace legal interpretation, site assessment or the accountable decisions required of responsible persons.
Can SIRV AI help with exercises and training?
It can support scenario preparation, retrieval of approved procedures, organisation of observations and follow-up records. People must design, conduct and assess exercises appropriately.
What documents could be assessed?
Examples include site procedures, response plans, assessments, training material and exercise records, provided the relevant approved sources and review criteria are defined.
What is the best first use case?
Choose one bounded, measurable activity with identifiable source documents and a named human reviewer. Procedure review and exercise-action tracking are potential starting points.
Discuss a readiness workflow
If your team is reviewing procedures, plans or exercise findings, SIRV can help identify a bounded workflow to test with clear evidence and human approval points.
Discuss a workflow
