What to do with the map when the audit is over
Three routes out of a process map, including the two where you do not hire anybody.
An audit ends with a set of priced processes and an order to attack them in. That is a decision document, not a purchase order, and there are three reasonable things to do with it.
Route one: build it yourself
For a good number of the processes that surface, this is the right answer, and the map is most of the work. A process with a named trigger, an ordered set of steps, the tools written down and a frequency attached is a specification. Plenty of companies have someone who can build against it in an afternoon using tools they already pay for.
The processes where this works share a shape: one person, one or two systems, no judgement in the middle. If the map contains several, take them. We would rather the cheap wins go early, because a project that starts by removing an obvious annoyance buys patience for the harder work later.
Route two: delete instead of automate
Some mapped processes should not be automated at all, and the map is what makes that visible. A step that exists because of a problem that stopped happening two years ago does not need a workflow. It needs somebody with the authority to say it can stop.
These are easy to spot because the reason behind them is thin. When the answer to why a step exists is that it has always been there, that is a candidate for deletion, and deletion beats automation on every axis. It is cheaper, it is immediate, and it does not need maintaining.
Route three: hire the build
What is left after the first two routes is where automation is genuinely worth paying for: multiple people, a handoff with waiting time in it, more than one system that does not talk to the next, and enough volume that the hours are real.
This is the engagement the audit is designed to scope, on the terms in [what paid on acceptance means](/guides/paid-on-acceptance). The point worth making here is that the map sets the order, not the sales conversation. The first project is chosen by the criteria in [the wrong first process](/guides/wrong-first-process), and those criteria will sometimes rule out the biggest number on the page.
The map has a shelf life
Process maps go stale. People leave, tools get replaced, and a workaround quietly becomes the official way. Six months on, a map is a reasonable historical document and a poor plan.
That argues for acting on the top two or three items while it is fresh, rather than building a roadmap out to next year. A roadmap written against a map this age spends most of its life describing a company that no longer works that way.