AI the Next Automation: New and Not New

BrandPost By Paul Brook
Aug 19, 2020
AnalyticsBig DataHadoop

abstract technology
Credit: Dell Technologies

A few years ago, when AI started gaining popularity amongst business thinkers, the term automation was simply not applied into any AI conversation. Now you cannot escape from hearing about it. So, what changed and why does this matter so much for organizations in both the private and public sectors?

Automation, it is certainly nothing new. Forged most famously in the industrial revolution and refined across the years it is a core tenant of business, automate to promote efficiency. Indeed, unless you are in a highly artisan-based organisation, automation is a business essential. The Artificial Intelligence angle to Automation does have a few twists and turns because it is quite a new phenomenon.

First let’s start with a definition from Wikipedia: “Automation, or Labour-saving technology, is the technology by which a process or procedure is performed with minimal human assistance.” This definition points toward two critical areas of AI driven automation. I will highlight here two specific areas that have resonated with customers from across industry and the public sector.

Firstly, we have some levels of AI expertise available at scale. It is a little unfair to say that AI has been around for ages without clarifying that the concepts, research and capability has been around for a time yes; but the practical and wide use of AI, this is much more recent. What we are seeing is a wide acceptance with so many smart phones and associated interfaces (Google dot, Alexa) having an AI embedded in the basic functionality of the product. And business sees benefit in automation- that uses AI. The automation of the classic Amazon ‘people who bought this also bought that’ or ‘you may also like.’ This is all part of a multi-billion-dollar automation based upon classic retail upselling. Basically, a software robot does a better job of this than a person does. Who would have guessed? Well honestly, all of us. The over specification, routine, person to person interactions killed individual customer connection. Don’t believe me? Please remind yourself of any scripted greeting or closing. We all know the sound of the empty corporate upsell.

“Black Coffee please”

“Cream and sugar with that?”

“No thanks – I just asked you for a Black Coffee”

Ever ordered Fries and been asked if you want fries with that?

“You’re kidding me, right?”

“I gotta ask” comes the reply. Insane way to treat customers, even those who only have 99 cents to spend.

So, here is my first shot. Personalization at scale is where AI can automate your business and your customer interaction. If scale is your thing, AI can support your business. If Growth is an ambition, AI has a proven track record in creating sales growth. Do not underestimate how AI is changing this landscape. If you are operating at scale and intend to grow but have nothing up your sleeve related to AI, then you need to get something effective in place. Do it now, you need it now. If you are currently doing nothing, then you are being out competed at every turn by a business that is using AI and coupling this to an automation of process.

Secondly: AI is becoming a tool for effective process control and process improvement. Where can you see the benefit? Everywhere that you have complex process and clarity of process you have a potential AI based improvement to be made. Think, for example of the classic’ ‘swivel seat interface,’ a rather derogatory phrase implying that instead of building a proper automated process interface the organisation pays a person to sit in a  swivel chair, manually entering data into one system then ‘swiveling the chair’ so that the same person enters exactly the same data into a different system. AI will fit perfectly into these gaps. And this may range from the chatbots masquerading as call centre operatives providing perfect, legally appropriate replies to Frequently Asked Questions, to internal supply chain paperwork/process control where a fraction-of-a-cent saving adds up and quality is improved by removing human error. What automation was made for.

Now here is the killer tip: Apply AI to the process that you understand best, to a system where you have the greatest understanding of the routines that underpin the process. Because creating an AI to automate something you do not understand is clearly doomed to fail. Yet so many times I listen to stories about ‘failed AI’ when what I am really hearing is management describing a business problem, or process, that they simply do not understand well enough. What they wanted was a magic wand that made a problem go away, often the failed AI project simply spotlights a deeper business problem. So, before you turn to AI and the associated automation it may bring into your business process, first understand the business process as best you can. This typically leads to a better ‘AI’ fix. Think about where the AI systems that automate your precision and critical business process need to scale, because scale with precision suits AI so very well.

And finally, hiding amongst the weeds of AI are the ‘big three’ operational tips: Collaborate, use the best tools you can use and get a ‘corporate memory’ for your AI. The corporate memory is not a data hoard; it is a curated and valuable repository through which you realise your digital future. Mandate the use of tools that enable AI teams to collaborate and provide essential governance against what can often be a fast-paced project. These will pay for themselves in no time though repeatability and effective corporate oversight. Yes, you need flexibility and agility, but AI at scale cannot be the wild west. Even for a start-up. Like so many technology themes, the technology part is complex and resembles science fiction in terms of potential practical outputs. The business and general operational problems that Automation using AI are best at fixing are business problems that your great grandparents may well have recognised.