Every element of IT appears to be ripe for automation. As we seek to apply Artificial Intelligence (AI) and Machine Learning (ML) to every tier of the modern IT stack from databases to applications and all the interconnected services that form the synaptic connections of modern systems, automation really does appear to be everywhere.
Given this automated, autonomous, artificially engineered truism, at what point do we draw a line around these new controls and efficiencies? Where do we rein in the Robotic Process Automation (RPA) robots? In other words, what should be not automate?
Rule #1, there are no rules
Our first truth is to remember that there are no rules, no industry standards and no limits. Automation doesn’t even have some form of de facto industry litmus test to recommend where the line should be drawn.
Every use case for automation in every industry vertical will of course be different. Despite the popularisation of ‘accelerator’ templating technologies designed to help shortcut software deployments at infrastructure, platform and application level, a business will need to examine each workflow inside its organisational structure to determine how much automation to apply to it.