You probably have something in common with Garry Kasparov and Lee Se-dol – you are outstanding at what you do, but a machine can do it better. Today there are few tasks where a machine cannot outperform a human with regards to time, cost, and quality.
This is in itself nothing new. We have always strived for efficiency. With the Spinning Jenny and industrial robots, we have scaled up production for more than 250 years – and we all appreciate machines doing repetitive and strenuous work. What is interesting now is when machines enter the market of white-collar jobs, and it is no longer only about doing the same task better. Now we question if the work should be done at all. This is a consequence of automation.
To automate something, we must first understand and define it. When this is done, we often see that a lot of work in an organization is pure waste. Well-educated people manually move data in and out of Excel and related tools (CRM, Marketing Automation, Finance, ERP, BI, …). I call this the copy-paste cost of an organization. It may sound harsh, but how long will your business last against fast-paced competitors with a digital DNA who can deliver the same results to your customers, cheaper, more efficiently, and at higher quality?
Copy-paste costs are ingrained in most organizations, in part due to our reluctance to embrace change. The world is changing around us, and to keep up, we must train our staff and inject change into our culture. Doing the same work and expecting another outcome will not work.
Imagine being a constructor with ten carpenters, each having a hammer. If you invested in a nail gun, one guy could replace all the others after proper training, or if you get ten nail guns and train all carpenters, you could increase your market share. This is very similar to working efficiently with data. If you don’t do it, your competitors will.
Additionally, organizations are changing slowly because of a lack of knowledge, competence, and leadership. If leadership/management/those who can bring about change do not know what is possible, then defining/specifying the requirements is very difficult. If the requirements and needs are not specified, it transforms the analytics journey into a matter of tool selection simply because it is easier to understand.
I have had the privilege of looking under the hood of many companies, including a multinational insurance company. Actuaries are held in high regard at these companies. They have in-depth theoretical knowledge from many years of studies, and they often have long experience in a complex business domain. However, digging deeper into what they do, they do not differ much from any other field – little innovation and value added, much repetition and maintenance. Their models and scripts may be seen as semi-holy artifacts as they contain a lot of formulas only, they understand, but what I see are code snippets with no or little governance or too much maintenance needed.
The future is the analytics workflow, which means automated decision-making based on trusted data. There is a fancier term for this: Intelligent Automation (IA), a joint force of Business Performance Management, Robotic Process Automation and Artificial Intelligence (AI), and the foundation for this is Analytics.
Data processes need to become lean as industries have been for decades. This is unlikely to succeed from a big-bang implementation but would as a deliberate roll-out to achieve certain milestones, one at a time. Parts of the organization will, of course, be faster to implement optimization and AI, and their work should act as inspiration and guidelines for the rest.
All of this will have both positive and negative consequences. On the positive side, companies and societies will be able to create more with fewer resources; research and medical teams can (with the help of analytics) diagnose diseases with higher accuracy and simulate pandemic scenarios. Even in sports, this is being used, with the NBA already reaping the benefits of using Spatio-temporal analytics to analyze /predict gameplay scenarios.
The only limitation for the use of analytics now is your imagination. If you embrace the analytics journey personally and as a company – there is no end to the possibilities.
On a larger scale, however, the transition may not be as smooth. Analytics, automation, and AI are changing the world already, but so far, we have seen few negative consequences. We hear little of it, but researchers are looking at what comes from automation at the macro level.
For example, the ripple effect of self-driving cars is much greater than one may fathom. The loss of millions of jobs, both drivers and supporting staff. Related to this are the dependent industries such as roadside diners and their suppliers – and it goes on. Even the auto industry and its suppliers may suffer from this as the need for individual car ownership decreases. Masdar City in the United Arab Emirates showcased this with their Personal Rapid Transit system.
One significant difference between self-driving cars and Intelligent Automation within white-collar jobs is the skillset of the workers. Now people with university degrees are at risk of losing their jobs due to automation, making perfect sense.
Analytics and IA are about workflow and culture, not about tools and IT. No company would ask their IT department for advice on how to run their business – but that is the result when considering IT as a cost rather than an integral part of the process and therefore worthy of investment. What is possible becomes limited by an outdated view on data and IT. So, whenever there is discontent or misunderstanding shadow IT grows.
One prime example of this is an energy company I once worked for. When walking through their financial department, there was an endless number of screens of Excel and PowerPoint. Most of this work could be automated – and what we often forget is that if we work smarter on the business side, we simplify the work for IT and analytics teams. It is a snowball effect to strive for.
Another example is from the manufacturing industry, specifically a multi-billion company with production, suppliers, and customers worldwide. In discussion with their CFO, he claimed, “99% of my time is waste“. He explained how much time was spent chasing numbers to get it right in their decision-making process. I was introduced to his team of analysts, financial advisors, and controllers to aid them when gathering requirements. I soon realized what they were doing was defining reports which are short for how to get data to Excel. I questioned their method and asked why they were looking at so few variables at a time; the reply was, well, how do you visualize something with that many variables?
Well, you do not. You create a model on top of the governed data and start optimizing. People can’t keep up with exchange rates, risk, orders, staff planning, production planning, capital management (and everything else) using Excel. Elimination of waste is one of the significant benefits of IA. This particular company had an estimated copy-paste cost of ~ 200 million EUR per year. The total effect, however, was at least tenfold.
To understand and appreciate the value of IA, we must have a holistic view of the organization and its costs. The same company used Power BI, Qlik, SAS, and IBM Cognos to serve the different departments, making efficiency impossible. Once ingested to the data lake, all data was read by the different Business Intelligence tools and later consolidated manually using Excel. All this work performed by thousands of people – I would describe this as unnecessary waste.
This company is one of the world leaders in lean production. I visited their assembly line, and I was impressed. Robots everywhere, specialized teams at every task/site, and backup experts if the team fell behind—no room for guesswork or improvisation. An optimal process had already been developed by R&D, with every step being monitored and measured for further optimization.
This is what awaits us on the business side. Business and domain experts collaborating with IA-experts to map the processes and then automate them on robust yet flexible IT architecture. By breaking down every function into process and subprocess and being able to automate it, in the end, humans will only oversee and adjust the machinery with expert teams.
If I am modest and say 50% of all work related to information and decision making in an organization may be automated, would you believe me? Even if you adjust that number up or down to your liking, it still means something.
We can ignore reality, but we cannot ignore the consequences of ignoring reality – the words of Ayn Rand.
All of this is happening now, today – and it is happening faster. It is like standing on the rails and watching a train approaching in the distance. If we do not understand what the future brings, we will be run over. If we embrace it – we become that train.