Operational systems for the work that actually delivers the service

Sometimes, and where one fits it is usually the better answer. Custom operational systems earn their cost when the way you schedule, prioritise or fulfil is genuinely distinctive — which for many organisations is exactly where their advantage sits.

Only if it is faster than what they do now. That is why these builds start by watching the actual work: a system designed around the report it produces rather than the task it supports gets quietly abandoned.

Yes, and for field operations it generally must. Work is captured locally and synchronised when a connection returns, with the conflict rules agreed rather than left to chance.

That is usually one of the first integrations. Operational events that never reach finance become a monthly rekeying exercise, which is often where the largest hidden cost sits.

Alongside the current method rather than instead of it, on a subset of work, until it is demonstrably better. Removing the fallback before that point is how these projects damage the operation they were meant to improve.

It comes free once the work is captured as it happens. That is the sequence though — capture first, reporting second. Building for the report first produces a system nobody on the frontline wants to use.