Why AI isn’t taking over the business world just yet

Many companies will also have invested significant sums of money testing various forms of AI – typically targeting operational efficiencies through robotic process automation or improved digital self-service through chatbots.

Yet, few will have managed to move beyond the pilot phase into full-scale production.

In most cases, this is not due to the technology but the existing business reality. Data continues to be stored in different places and in different formats. Many operating systems are not fully integrated. 

And, many decisions and actions continue to be taken by staff, outside of systems, and it is not that easy to automate these when key pieces of the puzzle are not connected.

Getting it connected usually requires significant changes to existing systems, processes and people. That is why most, if not all, disruptive businesses are started from scratch rather than emerging from existing business models, unlocked by the application of AI.

For existing businesses built and perfected over decades, the disruptive implications of AI remain daunting. This slows down adoption.

Breaking down what has worked and replacing it with what should work takes courage, leadership, time and money. It also requires shareholders with a long-term investment horizon that tolerates a dip in performance before the lift can be realised – not something most executives have the luxury to work with.

In many cases, it is forced on companies as they scramble to respond to a disruptive market entrant that threatens their very survival.

So, how do executives prepare their organisation for the digital era while driving improved performance via their existing business models?

How does one migrate a legacy business model to a digital reality without being forced to run parallel business models; or engage in a transformational overhaul from day one?

Take the contact centre. Sure it's tempting to get rid of all your agents and ask your customers to self-serve via chatbots. The financials make sense.

But, that requires you to first build a chatbot capable of answering all customer queries in a way that satisfies the customer. Doing this within an innovation lab, and hoping the digital team will get it production ready, is wishful thinking.

There are simply too many possibilities to capture using traditional decision tree coding techniques, and the amount of rich unstructured data is seldom enough to achieve acceptable predictable outcomes via machine learning.

As a result, the project struggles to get out of pilot phase simply because the risk of a poor customer experience is too high.

A more pragmatic approach is to stick with your existing human interfaces and augment them with digital intelligence. This gives you room to learn and make mistakes because your staff can step in when your digital logic is found wanting.

It means approaching AI from a both/and position, not an either/or one. In other words, not excluding staff in your digital mix. Only once you have perfected your digital logic do you then look to adopt purer forms of digital autonomy.

This approach not only applies to contact centres. Sales is another example. Instead of trying to perfect a chatbot to effectively sell your products online, start by offering all sales reps access to a digital sales advisor that they can use when dealing with customers.

Let the digital advisor help them sell better. Perfect the logic with them, co-piloting the customer engagement. Allow them to learn the differentiating behaviours that will make them better than a chatbot.

And, all the time, keep learning and improving your sales logic until its time to give it a digital interface.

This can also be used in technical support and any number of other operational areas. By initially building digital professionals that augment, not replace staff, it allows you to realise significant business benefits without being forced to transform the business model to make it work.

Not only will it enable staff to do more with less training, but you start building critical data that helps you shape and optimise your digital logic.

As the logic gets more robust and accurate, you can look to adopt more digital interfaces where relevant. By that time, you will have empowered your staff to think and operate within a digital world; one where prescribed decision logic will increasingly be tackled by technology, and where the human role shifts to perfecting the behaviours that help differentiate the customer experience.

Intelligence augmentation (IA) offers existing businesses a very pragmatic first step into the digital era. It focuses on getting the back-end logic or intelligence perfected before you try perfect the front end experience.

It also allows staff to increasingly develop their EQ given that their IQ can increasingly be supplemented digitally.

2018 is going to be the year where AI moves out of the labs and into the mainstream. And intelligence augmentation solutions will be one of the key ways existing businesses achieve this outcome.  

For more information, visit www.clevva.com. You can also follow CLEVVA on Facebook or on Twitter.  

Contrary to what many people think, humans play an integral role in applying machine learning-driven solutions to real-world problems. Read more in our article, The vital role of humans in machine learning.