Verify safety actions before closing the register
Compare approaches
TL;DR — what you’ll learn
- Treat completed work as a request for verification, not automatic closure.
- Agree what evidence would demonstrate that the original condition has been addressed.
- Return incomplete actions with a reason, owner and next step while preserving their history.
- Review recurrence separately from the first completion check so repeated problems remain visible.
A supervisor uploads a photograph of a new barrier and marks a safety action complete. The original finding concerned an access problem, and the new arrangement still leaves that problem unresolved. The action happened. The intended result did not.
Closure should depend on a qualified verification against the finding and its agreed acceptance criteria. The person reporting completion needs a clear way to submit evidence; the person authorised to verify needs a clear way to accept it or return it. The distinction matters more than the wording of the status button.
- Original finding
- Access problem
- The intended result defines verification
- Completion evidence
- Barrier photograph
- Shows an action, not necessarily its adequacy
- Verification
- Access still unresolved
- Return with a reason and retain the first attempt
The example distinguishes the activity completed from the access condition the verifier must assess.
Read the visual as text
Original finding: Access problem. The intended result defines verification. Completion evidence: Barrier photograph. Shows an action, not necessarily its adequacy. Verification: Access still unresolved. Return with a reason and retain the first attempt.
Original explanatory visual: NextEdge Labs. Illustrative records, not client results.
- STEP 01Agree the intended result
Owner: Safety lead and action owner
Define evidence and verification authority.
- STEP 02Submit completed action
Owner: Action owner
Attach evidence against the original finding.
- STEP 03Verify or return
Owner: Authorised verifier
Record the result and reason without erasing attempts.
- STEP 04Review effectiveness where required
Owner: Safety owner
Track recurrence and changed conditions.
Record verified closure for its stated scope.
Return the submission with a named owner and evidence gap.
Return path: The owner revises and resubmits; the verifier checks the new evidence against the same finding.
Read the workflow as text
1. Agree the intended result (Safety lead and action owner) → 2. Submit completed action (Action owner) → 3. Verify or return (Authorised verifier) → 4. Review effectiveness where required (Safety owner). After step 3, ask: Has the agreed condition been demonstrated? Yes: Record verified closure for its stated scope. No: Return the submission with a named owner and evidence gap. Return: The owner revises and resubmits; the verifier checks the new evidence against the same finding.
Workflow: NextEdge Labs. Apply the project’s actual authority and procedures.
Start with the condition to be resolved
Before assigning an action, describe the original condition, the location, relevant activity and the decision already made through the site's safety procedure. Record any immediate response separately. Administrative closure must never substitute for that response or delay escalation of an urgent situation.
The action owner and safety lead should agree the intended result and the evidence needed to assess it. This does not mean prescribing a standard technical remedy from a generic register. The competent project team determines the appropriate measures for the actual conditions.
An action such as install barrier describes an activity. A verification criterion must address the reason for that activity. If the original issue is safe access under a particular work arrangement, the verifier needs to assess that arrangement, within their authority, rather than simply count installed barriers.
Follow a returned action without losing the work
In this fictional example, the action owner submits the barrier photograph and identifies the affected location. The verifier compares it with the original finding and inspects the relevant conditions. The access issue remains, so the verifier records not accepted, explains the gap and routes the action back to its owner.
The system retains the completion submission, photograph, verification date and return reason. It does not delete the first attempt or restart the finding as a new record. The revised action can then address the gap, with the responsible safety authority reviewing any changed approach.
On resubmission, the verifier sees both attempts. If the intended result is now demonstrated, the record can move to verified closure under the approved procedure. The closure entry identifies the verifier, scope and evidence. It does not certify every surrounding activity as safe.
Separate three kinds of review
Completion review asks whether the assigned action was carried out. Verification asks whether the evidence demonstrates the intended result for the original finding. An effectiveness or recurrence review asks whether the result continues to hold in the relevant operating context. These checks may occur at different times and need not have the same owner.
For a straightforward action, the approved procedure may combine some checks. For more significant or recurring issues, the safety owner may require further review. Record the chosen approach and the reason instead of forcing every action through an identical sequence regardless of risk.
Procore's observation workflow distinguishes ready for review from closed and not accepted. That product pattern helps illustrate the submission-and-verification boundary. It does not determine who is competent to verify your action or which evidence is technically sufficient.
Give deadlines and escalation a purpose
A due date should come from the site's approved risk and action process. The register needs an owner, a delegate where appropriate, an escalation path and a reason for any changed date. Keep the original due date visible when an extension is approved so management can understand the actual history.
If a verifier is unavailable, route the task to an authorised substitute. Do not let the action owner gain closure authority simply because the primary reviewer is absent. Where the approved procedure permits a combined role for a particular case, record that permission explicitly rather than allowing a blanket self-approval setting.
Reminders help with administration, but they cannot resolve an unsafe condition. The system should distinguish an overdue upload from an unresolved operational concern and direct each to the appropriate process. Avoid presenting notification delivery as evidence that the responsible person has acted.
Handle recurrence honestly
Suppose the same access problem returns a week after verified closure because the work arrangement changes. Create or reopen the record according to the site's procedure, link the previous finding and preserve the earlier evidence. The safety owner should decide whether this is a recurrence, a new condition or evidence that the earlier action was inadequate.
Record what changed: location, activity, contractor, equipment, sequence or other relevant conditions. A recurrence report without context can encourage superficial blame. The aim is to help the team understand why the previous response did not remain effective.
Do not calculate a success rate from closed actions alone. Show returned submissions, time awaiting verification and recurring findings as well. A low return rate can mean good first submissions, but it can also mean weak verification. Sample the records before drawing a performance conclusion.
Configure the evidence handoff
The completion screen should show the original finding and agreed criteria before asking for evidence. The verification screen should show the submitted material, relevant history and the verifier's permitted actions. Require a reason when returning an action, and make that reason visible to the person expected to respond.
Test role changes and offline submissions. If an old mobile record says open while the server has a revised action, the system should present the conflict for resolution. It should not overwrite the newer decision because the older submission arrived later.
Retain an export containing the observation, assignments, completion attempts, reviews and final disposition. The site's record-retention policy should determine how long it is kept. Access to incident details or personal information should be narrower than access to an ordinary action summary.
Close with a reproducible reason
An effective closure record lets another qualified reviewer understand what was found, what changed and why the evidence was accepted. That is a stronger operating result than a green dashboard tile. It also gives the next team useful history if the condition returns.
- Take one completed action and compare its evidence with the original finding.
- Have the safety owner define the appropriate verifier and acceptance criteria.
- Test a returned submission, reassignment and later recurrence without losing history.
- Review verification backlog and recurring conditions before celebrating closure totals.
Compare this workflow with your current setup
Share the step that takes the most manual work, the systems involved and the exceptions your team handles. That gives us a concrete starting point for comparing approaches.
References and further reading
Sources checked 15 October 2026. The worked examples and diagrams are original editorial explanations, not client results.
- Procore — Respond to an ObservationReady for review, closed and not accepted states
Product review-state example, not a prescribed NCR procedure.
Visual credits
Original explanatory visuals and workflow diagrams by NextEdge Labs. Labels and text equivalents are included in the page. Numerical scenarios are illustrative; diagrams preserve the project owner’s approval and exception decisions.