Digital Safety Management: A Practical Pilot Brief

Measure the current process

Follow a record from initial capture to accepted closure. Record active work separately from waiting, repeated entry and correction. Note how records were selected and which roles participated. A baseline that excludes difficult shifts, poor connectivity or incomplete submissions will not describe the whole operating problem.

Agree acceptance before configuring the form

Define the task, expected evidence, responsible reviewer and circumstances that would stop the pilot. Include a rejected submission, a changed owner and a missing attachment. Choose locally justified targets and a period that allows the relevant work to occur; there is no universal pilot duration in this brief.

Include the people who capture and review records

Ask frontline users, supervisors, administrators and downstream task owners to complete realistic work with the proposed permissions and training. Log assistance, failed attempts and workarounds alongside successful submissions. A completed training session does not by itself demonstrate that the task can be performed.

Check performance and revise the process

HSE’s management guidance connects checking performance with reviewing and improving arrangements. Apply that principle by reviewing evidence quality, task completion and unresolved failures in the pilot. These proposed software measures are not HSE benchmarks. Increased reporting may reflect changed access or reporting practice rather than a change in underlying harm.

HSE: How to manage health and safety. Management principles in Great Britain; no software adoption, implementation-time or ROI benchmark. Checked 20 September 2026.

Plan migration and handover

Agree which records must move, which remain accessible in the previous system and how attachments and history will be reconciled. Assign support, routine administration and change approval. Expand after the process owner accepts the evidence and the operating team can support the next site or workflow.

Write down what you will do if the pilot fails

State before the pilot begins what happens if the acceptance criteria are not met: whether the vendor is asked to correct a defined issue within a set period, whether a different tool is trialled, or whether the existing process continues unchanged. Agreeing this in advance keeps a disappointing result from being reinterpreted afterwards, and it tells the pilot users that an honest account of problems is wanted and will not count against them.

Choose the pilot group and period on purpose

A pilot run only by enthusiastic volunteers on a day shift with good coverage will flatter almost any product. Include at least one group that is less comfortable with the change, one location with weaker connectivity and one shift that has less supervisor presence, then compare their results separately instead of averaging them away. Choose a period long enough for the less frequent tasks to occur, such as a monthly inspection or a repeated finding, and write down any event during the period that would distort the comparison.

Decide how the old and new process overlap

Running the spreadsheet and the new application side by side doubles the entry work, and people will tend to trust whichever one they filled in last. Agree a cutover date for each task, say which record is authoritative until then, and make the earlier records read-only after the date rather than deleting them. Note who answers questions during the overlap and how a record entered in the wrong place is moved, so that the pilot measures the new process and not the confusion of managing two.

Consult the people before the pilot, not after

HSE presents worker consultation as four stages: prepare, plan, consult and improve. Treat the pilot as one such consultation. Before configuration, ask frontline users which parts of the current process cost them time or lead to errors, and write their answers into the baseline. After the pilot, return the findings to the same group and record what was changed in response. A pilot that people feel was done to them tends to produce workarounds that the metrics never record.

HSE: worker involvement. HSE guidance on consulting workers in Great Britain. It is not about software pilots; applying it to them is our reading. Checked 3 October 2026.

Fit the pilot into the safety programme it supports

OSHA's recommended practices for safety and health programs are built around seven core elements and are written with small and medium-sized businesses in mind. Whatever the software does, it supports a programme that already exists: inspections, training, hazard reports and reviews. Before the pilot, name the element of your programme the workflow serves and who owns it, so that a result such as faster closure is read as a statement about that element and not about safety in general.

OSHA: recommended practices for safety and health programs. US guidance for employers. It does not evaluate software. Checked 3 October 2026.

Review the pilot as a management review

HSE's guide to managing for health and safety frames management as plan, do, check and act. After the pilot, hold a short review that follows those four words. What was planned, what was actually done, what the evidence shows and what changes as a result? Keep the written outcome with the decision record. If the review only repeats the vendor's summary, the pilot has not produced evidence of your own.

HSE: Managing for health and safety (HSG65). Guidance on managing health and safety in Great Britain. The pilot review is our adaptation. Checked 3 October 2026.

Bring five items to the closing meeting

The meeting that decides whether to expand, revise or stop needs five things on the table: the baseline figures and how they were collected, the pilot figures on the same basis, the list of workarounds that users invented, the unresolved issues with owners, and the decision options with the effort each would take. If one of the five cannot be produced, postpone the decision and name who will produce it. Deciding without it turns the pilot into an opinion.