It should be fairly obvious, but an automated system won’t perform a task the same way a person can. The automated system will be fast and consistent, but the person will be able to use their judgement and provide flexibility where needed. So automation means compromise and change, and it’s best when everyone can accept that early on.
Here’s an automated task (garbage truck with bin-lifter arm). These trucks are great – they’re quicker, quieter, and we don’t have split garbage bags and strewn rubbish on the street after they’ve gone past. But we had to change our behaviour: we all have to have the exact same bin, and when we put them out they have to be correctly positioned. Automation needs predictability!
Automation works best when the tasks it performs are limited and repeated. Imagine if they’d tried to accommodate all the different sized existing bins with those garbage trucks, just so people didn’t have to change what they were already doing. The lifter arm would have to manage all the different heights, widths and lids. We’d end up with an expensive, over-engineered “solution”, with multiple break points making it unreliable, that doesn’t even work for every bin. All that hard work and we ended up with something everyone says is over-priced and useless.
Yet I am regularly told by customers that minimal change to existing practices is a “top priority requirement”!
There are many tasks that shouldn’t be automated at all – either because it’s not cost effective or because a level of human judgement is required (and we’re not designing artificial intelligence here). To continue with the bin metaphor: they may have automated the task of emptying the bin, but we still have to fill it and put it out on the street, and someone still has to drive the truck.
So when designing your IAM solution, try and focus on tasks that can be completed cleanly, with minimal decision branches and “moving parts” that weaken the system. And make it clear that existing data and existing processes will have to change.