Automation Platforms

There’s an app for that

Not so long ago, as the smartphone revolution gathered pace and everyone with a serious consumer business developed the mobile habit, the phrase ‘There’s an app for that!’ became commonplace, to capture the point that apparently there was no problem that a decent smartphone app could not resolve.  We’ve become equally familiar with the Gartner notion of the ‘hype cycle’, where expectations for a technology rise to fever pitch and then drop off even more quickly than they appeared, as efforts to engage with the technology grapple with reality.  In many ways, the smartphone is now coming through the other side of its ‘trough of despondency’.  Where does Artificial Intelligence sit on the hype cycle?

Our fear of Artificial Intelligence is ramping up the hype

AI has been around in some form for decades, but seems to have gathered pace recently, and there is nothing that it will not do, to the extent that if we allow unfettered development, it will take over our jobs and threaten the very fabric of civilisation as we have known it.  In many ways, it is a knee-jerk reaction driven by fear, and the feeling that we are giving the machines so much more capability that it seems we are developing their ‘brain’ power at the expense of our own.   True, we should always be cognisant of the ‘Law of Unintended Consequences’ but that seems a bit extreme.

Experience of technology in business shows that when it becomes invested into the fabric of the way a business runs, we cut through the hype to find applications of technology where practicality and business impact enable us to get down to what really makes a difference.  Unless there is a clear business benefit to doing something, company executives are hard-nosed enough not to allow ‘hammers looking for nails’ to overtake their thinking, and after a while, meaningful applications emerge.  In that sense, AI is now heading quickly into the trough.  Whether or not it emerges will depend on finding the use cases that really do make a difference in helping business.

Let’s be clear on what we mean by AI

In looking for use cases in business, most are not looking for the cure for cancer, or ways to fight drought and poverty in sub-Saharan Africa.  Whilst these may be valid uses for AI, they simply don’t have meaning for most companies.

Getting to the business impact means initially breaking down the technological challenge in such a way that the role of the technology can be understood, and in this context, it’s unlikely that most businesses will deploy truly Artificial Intelligence in the sense that the machine develops its own capability to learn and take actions as a result.

To do this, the system must be programmed to act in a certain context, and what is really meant may be no more than an element of process automation based on workflow driven by a more complex decision tree powered by more data than previously.  Throw in the speed at which it can work, and it becomes understandable how this might be considered a level of intelligence that is non-human, hence ‘artificial’.

That is somewhat different from cognitive applications where the system can assess historical data and feedback from observations and past actions, and go on to draw its own conclusions about what to do next, or choose not to i.e where it is learning, making a suggestion as a result or going on to take action unsupervised.  Full AI, where the computer draws conclusions from inadequate data that has gaps in it, or where context requires judgement that can vary in its understanding and its application, or where a human might do something that is thoroughly illogical when driven by emotion rather than common sense, remains a long way off.  Answering questions is easy, understanding questions and context is much more difficult.

We need to reset our expectations.  Where businesses are starting to see that benefit emerge is in three areas:

  • Intelligent Automation, where the workflow of back office tasks can bear further automation if the machine can be fed with a better rule base to achieve it
  • Cognitive Insight, where increasing volumes of data can be analysed and understood to improve the interpretation in business and its impact in critical processes
  • Cognitive Engagement, where software can be applied to understand human intention through language processing, eliminating a lot of basic tasks that might otherwise require human intervention

None are ‘Artificial Intelligence’ in the headline grabbing sense, they are really the continued development of applying machines to business functions where (a) they are able to help and (b) humans need help.  So let’s get real.

Artificial Intelligence is a big challenge

It may be a challenge, but that should not deter us from trying to find the use cases that can be of benefit.  This is what we have always done to try to improve business, to make things easier to run and deliver better outcomes.

The intention should be to find applications where the machine can do the ‘heavy lifting’ of work that means that humans can be relieved of having to spend time in non-value adding activity (aka ‘boring jobs’).  If that means ‘replacing’ humans, then it is only humans doing non-value adding activity that would quite rightly be at risk of ‘replacement’.  In practice, to date, it is the non-value adding activity that is being replaced rather than the humans, who are freed up to be re-skilled into value-adding roles.

In general, companies are not prepared for AI and there is a long way to go.  If you are to allow the computer system to take autonomous action on the basis of data that you hold, and recognise what to do in a given set of circumstances, then your data and processes had better be in decent shape first, otherwise all that will happen is that the computer will take actions that will be no more educated and at worst produce completely the wrong result.  Data quality and clarity of process are fundamental with most businesses failing on both points, and it could take a period of some years of improvement to this foundational business infrastructure before AI has the right base from which to operate.

Similarly, companies looking at AI must consider the culture and skills of the organisation to understand whether or not their people are ready to embrace the continuous change programme that will arise.  Fear of computers and the impact on jobs must be addressed by better understanding all round of what is really involved, and what the benefits will be – much the same as any other change programme.  Just forget the ‘Artificial Intelligence’ label.

In Summary

Artificial Intelligence may be the headline grabber, companies that are now beginning to grapple with what it means are finding the reality a challenge that still holds promise, when understood for what it really is.


If you wish to share your thoughts on this article, you are welcome to comment below, or to contact the author through the Contact Form or by phone on +44 113 242 3795.

Leave a Comment