Are Analysts paving the way for The AI Revolution?
Over the past few years, it has been hard to ignore the stream of stories prophesising the impact of the coming artificial intelligence (AI) or robotics age.
Warnings from experts about the potential harm it could cause are coupled with companies reeling at the prospect of making even greater efficiencies and savings. Unlike the previous revolutions that impacted manual or skilled labour, this one has highly paid professionals such as doctors, lawyers, and accountants worried about their futures. A realisation and concern around the risk of their roles being replaced by a super brain or some animated hologram previously only dreamt up in science fiction shows. Though some versions of the above scenario are likely, you have to put AI in the same context as the paperless office. Over 20 years ago, everyone was talking about a world in which paper tickets and a whole host of other documents would be digital. It is fair to say that a lot less paper is used compared to 20 years ago, but it is more paperlite than paperless.
The concept of employees who you only have to pay for their upkeep, which will work for free twenty-four hours a day, seven days a week and take little to no breaks will naturally excite organisations. This concept is driving leaders to either look at or commence their version of a robotics programme. However, the significant savings can often blindside them into investing in expensive software and hardware without fully quantifying how this new technology can be fully adopted within their organisation. In over 20 years of working, I have seen countless businesses put all their eggs in one basket with one or two big strategic programmes and then only two years in seeing them either reducing the scope or having a blank cheque to cover the ever-increasing costs associated with the delivery.
The simple rule when it comes to Robotics is to keep it simple. Robotics should be business, not technology-led, and your analysts should play a fundamental and strategic role in the programme, which will be carried out through a phased implementation. Like any revolution, it starts small and builds momentum. Let us now look at the different types of robotics implementations and potential strategies for adoption.
The first stage is a basic level of automation. Think of this type of automation as a keystroke or simple, repeatable process whereby the information is readily available and in the right format. There are a lot of applications on the market in which the user can record a series of actions or steps that they undertake for a particular process. Once recorded, you can assign a trigger or prompt that, when it occurs, the computer program will run the sequence of steps. A simple example of this is shortcuts on your keyboard; in most applications, pressing the keys Ctrl and the letter ‘P’ triggers the printing of a document. You could use basic automation for workflows driven from your website or, for example, prompted processes, such as whether the report has been received, yes or no. Though it may not be as advanced as the other phases, you will realise savings and prepare your organisation for what is to come next.
The second stage is Advanced or Enhanced Automation, taking what we did in the first step but doing more and going further. In advanced or enhanced automation, we might have workflows that trigger actions based on data contained within emails, paper documents or user-prompted actions. Typically, most robotics applications do not have Optical or Intelligent Character Recognition (O/ICR) capabilities as standard, which means that your robots will not be able to read. By adding this functionality, your robots will be able to read emails, attachments, or even scanned paper documents, whether through printed text (OCR) or handwritten text (ICR). This phase of the AI evolution will open up more processes and savings, whether by performing tasks triggered by customer communications or by freeing up resource-intensive processes. By now, you will likely begin to see significant savings across your organisation.
The last and final stage is the full artificial intelligence cycle of adoption or evolution. Developments within this area are frequently changing. Most organisations would start by building decision tree algorithms and even monitoring production or live transactions, enabling them to begin a reactive, cognitive model within their organisation. Let’s use an example of a complaint process: a customer sends in a complaint via email. The email states that the client’s flight was delayed for 8 hours, and no one offered them any form of subsistence or compensation. The airline has a policy in which they will provide compensation of up to 10 dollars per hour where there is a delay of 5 hours or more. In the pre-Robotics process, the complaint would likely be passed internally within the organisation, eventually to an airline employee who will assess the claim and finally settle it by paying the customer 30 dollars. In an AI world, this could be handled much quicker and simpler. The customer would be able to fill out all of the necessary information via an online form. Once submitted, this would trigger a series of workflows and, based on the information provided, would derive a decision. The decisions would not only be based on the information and evidence submitted but also based on similar past cases. The robot would be able to calculate the most probable outcome and, based on the results, would automatically calculate the compensation and even pay the 30 dollars directly into the customer’s account without a single human being involved at any stage of the process.
Though this was a straightforward example, by feeding your AI engine with past decisions and actions, you begin to help it create a probability model. Each time a new case comes in, the AI would be able to assess the nature of the communication and would then follow a set process, which would, in turn, enable the robot to provide an outcome based on the most probable and best solution. AI is expanding, and already, there are sophisticated cognitive developments in areas such as speech and visual recognition, making the application and opportunities exciting for the future.
It is hard to ignore the warnings, and I do believe there is some merit in the concern around using AI in the wrong context or setting; however, from an evolutionary perspective, AI is exciting. Imagine a world in which doctors are virtual, able to examine you via an advanced smartphone, portable, and available wherever you are in the world. It is not something to fear, as humans would still likely play a significant role in building and maintaining robotics and in processes that require specialist knowledge or do not follow the norm. A world of virtual doctors would inevitably see health care costs fall, but so would unnecessary deaths, especially when health care is not readily available due to the expense and availability, being in remote parts of the world or just because help comes too late.
Analysts should be the ones building the workflows, defining the parameters and tweaking the robots to ensure optimum results. Ultimately, though, for it to be truly effective, it cannot be seen as a fad or a must-have; instead, it should be a strategic platform deeply ingrained within any organisation.
