AI, bots, and wizards

Process improvement has been part of the corporate world since the very beginning - an ideal of faster, simpler and consistent ways of working. But it only works when people moving through the process fit the assumptions on which the processes were built.

It started with Henry Ford’s Assembly Line, evolved into McDonald’s standardised kitchen routines, and now we have binary codes of automation and artificial intelligence.

Standard Operating Procedures (SOP) tell us, or the computers, what to do and when to do it.

A member can journey through the process, and when the answer doesn’t fit, what happens?

We can follow the symbols, arrows, and shapes of a process map until it stops working; the computer says no.

But what happens to the person standing on the other side of no.

This is where investigating complaints can tell us what a process map can’t.

And that’s also when organisations need to consider: is it the member not working within the process, or does the process not work for the member?

A complaint is often where this question becomes most apparent.

Cheques come to mind. They’re not very efficient. It takes time to send a cheque from the payer to the payee; the payee then needs to bank the cheque and wait for the funds to clear, which can take up to five days. In a cashless economy, a cheque doesn’t belong.

So what happens when the process evolves faster than the people expected to use it?

Are we creating an underclass of people who don’t fit 21st century frameworks?

If 21st century life has taught us anything, it’s that there is no one-size-fits-all and that we are all diverse in our own ways. So how do we reconcile a culture of standardisation and diversity, where the member experience says no but the system says yes?

Where do the exceptions go?

Complaints (in my experience). A complaint is not necessarily evidence that someone failed to follow a process. Sometimes it is evidence that the process has failed to account for someone’s circumstances.

The complaint then becomes the data point the process map didn’t know it needed.

Automation should remove repetition, not judgement.

Standardisation should create consistency, not take away curiosity.

And a process should tell us what normally happens, and not convince us that nothing else can happen.

The danger isn’t in the automation, it’s in the end of human thought.