The engagement that recommended against buying software

Workplace health and safety records lived as three disconnected sets of forms: incident reports, hazard reports, and corrective actions. Each worked on its own. None of them knew about the others.

So a hazard could be reported, and a corrective action raised in response, and nothing in the system connected the two. Nothing chased an action that had gone past its due date. Whether anything actually got fixed depended on somebody remembering that it was outstanding, which is precisely the thing a safety system exists to stop depending on.

Silver Lake had previously proposed a custom-built command centre for this firm.

On working through what the requirement actually was, that proposal was withdrawn and the recommendation reversed: do not build custom software. The firm already licensed a business software suite capable of doing this properly, and its own staff could maintain a solution built there without depending on an external developer indefinitely.

Three linked lists for incidents, hazards and corrective actions, joined so every corrective action permanently carries the report that caused it.

Three automated flows. A submission saves the record and notifies. An assigned action notifies its owner. A daily job chases everything overdue, which is the one that changes behaviour.

Delivered as a step-by-step build guide, with flowcharts and exact interface labels, so the firm's own team could implement it without further involvement.

Because the most useful thing a software consultant can tell you is sometimes that you do not need software.

A custom build would have been a larger, more interesting and considerably more profitable engagement. It would also have left a small consultancy maintaining a bespoke application, and dependent on its author, for a problem that off-the-shelf tooling already solved. The recommendation ran against the commercial interest and it was the right call.

This was an advisory engagement. Silver Lake diagnosed the problem and specified the system; the client's own team built it.

That is the point of the study rather than a gap in it.