The short answer, it stays where it already is
When an automation works on the programs the company already has installed, the data doesn't move anywhere, it stays in the business software, in the folders and on the server where it sits today, because the system opens the same screens an operator would open and writes into the same fields
That's the general answer, though. The real one depends on how we decide to run the system in your case, and it's decided before we start and put in black and white in the project contract, not discovered once the work is under way
The questions to ask, of Muffin Suite and of anybody else
Whoever offers you an automation has to be able to answer you on five points, and if they skate over them quickly, that in itself is the signal
- where the system runs, on an office computer or on a machine of its own, because that's where the data passes while it works
- which data it actually touches, because an automation sees what the process you handed it needs, not everything the company holds
- whether anything leaves the company, and in which cases, an external portal for instance, which the automation logs into because a person already logs into it today
- who sees what, which credentials the system logs in with, and whether the log of what it did stays there for you to read
- what happens when the relationship ends, to the access, to the written rules, to everything that was produced
They're plain questions to ask and awkward ones to dodge, and that's exactly why they work
Where something changes, and where it doesn't
An automation that opens your business software on your computer and copies into it what it reads from a PDF already sitting in your folder is one thing, a service that carries your documents away to process them elsewhere and then hands you back the result is quite another
In the first case the perimeter stays the one you have now, in the second it widens, and that isn't necessarily wrong, it's that you have to know beforehand and decide it yourself
Then there are the things that sit outside no matter what, the government portal where you upload documents, the mailbox if it's hosted somewhere online, but those pieces were already outside before any automation turned up, a person was logging into them with a browser
The choice of where to run the automation is the same choice that decides where the data passes, so the two questions get asked together, not one after the other
Muffin Suite doesn't ask you to take its word for it, it asks you to put those five questions to it, along with the one about the credentials the system uses to log into your programs, which is the other half of the same matter
The rest of the method, from the first look at the process through to the rollout, is on how it works