English | EN 
I’ve spoken before of situations where the adage for why things are done the way they are is because they’ve always done it that way. There is a lot of merit in having a well-established, robust and mature process or way of working but recent advances in artificial intelligence also provide an opportunity for truly transformational change.
Existing processes will have been developed and fine-tuned over time based on a combination of knowledge and experience gained from both successes and failures along the way. But equally these processes will have been shaped by the constraints of the time in perhaps the technology available throughout their lifetime. Some early ‘design’ decisions may have even impacted or constrained the process changing the process later.
The catch is that some parts of a process we wish AI to review and modernize may exist for reasons which nobody can remember anymore. “We’ve always just done it that way!” may be entirely valid and correct but steps which an AI may evaluate as inefficient or unnecessary may be removed. Only for the reason those steps existed to recur, and age-old problems resurface in the absence of the established processes which normally mitigate it.
There is indeed an apocryphal tale of chimpanzees being discouraged from eating a banana to such an extent that eventually the chimpanzees would themselves discourage newly introduced members to the troop not to eat the banana. So much so that in the longer term when all the chimpanzees from the original troop had been replaced, the troop maintained the behaviour of avoiding eating the banana.
The lesson here seems to be it is prudent when considering using AI to modernize your processes, to remember the adage “Fools rush in where angels fear to tread”, and to take the time to truly understand the process before allowing AI change it.
The uptake of AI has been phenomenal and there remains a persistent interest to increase its usage throughout business processes. There are also examples being reported where the transition has not delivered as hoped, and implementing AI has proven to be a detrimental step. That is not to say “Don’t use AI” but more to not rush and to remember to tread carefully. AI, specifically with Large Language Models, is still a very new technology and we are still working out how to use it most effectively.
My initial suggestion would be to ensure there remains a human in the loop at key points in any AI process. This would ideally be at a decision point but also to review what any AI is doing and to deal with those instances where the result is either not predicted or appears to be wrong.
But further than that I firmly believe AI gives us a really huge opportunity to make meaningful changes to existing processes and configurations. On the home front I have been able to fix several issues beyond my ken on the setup of my home servers and networks. They now operate much more as I intended them to behave, as per my original requirements by eliminating the work arounds I just found it easier to use.
In business we have the opportunity to look at how things work more deeply, and to use AI to work out if we could make things work more smoothly. For I challenge anyone visiting the World of Work for any amount of time not to find a single process which is not frustrating and time-consuming, and no one’s ever been able to fix it! It may not be seen as immediately adding value in the sense of a customer-facing hi-tech cool new thingmajig but if instead you can improve a few simple processes which consume a notable chunk of most of your employees or customers time for no discernible benefit, I see that as a win-win: Happier employees or customers; and less time lost to an unproductive process.
Whilst AI promises a glittering future, it also offers a chance to put down concrete foundations for that future by offering us an opportunity to apply new technology to our existing World of Work, to streamline that and to reshape it to make AI work for us all in the long term.