Automation of service processes

Humans and robots working hand in hand, completing coordinated processes until the finished product leaves the factory – this scenario has been commonplace in production facilities all over the world for many years. Robots and other computer-controlled systems are integral components of automated production processes, and are generally accompanied by a substantial increase in efficiency. The tasks which fall to human workers on modern production lines are quite different from what they used to be: they now control production processes, fulfill any special customer requests and take action to solve problems in the event of a failure. Instead of traditional conveyor belt work, employees now tackle more demanding activities and act as experts in automation technology, robot control or production design. Accordingly, employees’ skills more significantly involve the interplay between human and machine.

Unlike the production environment with its highly refined technology, the processes and tasks involved in business administration and company services have hardly changed in recent decades – despite digitization and ever-more-powerful information technology. Manual processes continue to dominate. Administrative staff busily input information from Excel lists by hand, and media failure means that data must be recorded redundantly. In other words, there is still considerable potential for optimizing processes and working more efficiently.

There have been repeated attempts within the service industry to increase productivity. In the past, decision makers have introduced new platforms and have been willing to invest in automation solutions. However, these efforts met with little success for various reasons, including the increasing complexity of process and product landscapes and the fact that the automation technologies implemented did not achieve the stability and economic efficiency required. Finally, the necessary investments exceeded companies’ financial means. In the past, the alternative to process automation was often found in outsourcing particular work steps to low-wage locations, and completing the work there with little or no change except for lower labor costs.

In the age of digitization, however, this strategy is reaching its limits. The digital transformation has fundamentally changed the conditions under which companies act and the requirements to which they are subject. The advantage of higher efficiency is no longer the only argument in favor of automation. Rather, customer expectations can today only be met by employing a high degree of automation. In addition, technologies which are essential for the consistent automation of service processes have reached a new level of maturity. Robotic process automation (RPA), machine learning and natural language processing (NLP) – which includes speech recognition and the translation of texts – can now reasonably be used by wide circles of users, and are close to making a general breakthrough as realistic alternatives for the handling of service processes.

The two Horváth & Partners studies, “The Use of Bots in the Financial Services Industry” and “Production Management in Service Centers”, demonstrate that the backlog in terms of process automation in the business administration and services sectors is not due to a lack of willingness on the part of managers. The reports examine the effects of using robots and artificial intelligence on the future design of processes – and bring to light a considerable discrepancy between intentions and reality. “In most cases, decision makers are aware that their companies are lagging behind technological developments, and recognize the need to automate processes and create intelligent working environments,” states Sebastian Ostrowicz, Principal at Horváth & Partners.

This self-assessment is just as clear as the expectations with regard to process automation. In customer administration and financial processes in particular, decision makers are counting on average increases in efficiency of between 20 % and 30 %, with peak values as high as 50 %. From a cross-sectoral perspective, approximately two in three companies expect that the implementation of bots and intelligent machines will be of high to very high relevance in the future. The studies show that the automated handling of mass processes is especially important for shared services center organizations, such as those in the areas of HR, IT, finance and operations.

The study paints an equally clear picture of the barriers which obstruct the implementation of automation solutions: the main problem identified by more than 70 % of respondents is a lack of process standardization, followed by insufficient expertise and employee resistance, each at 50 %. The differences between the various industries and sectors – sometimes huge – concerning the level of dissemination of bot technologies are striking. While three-quarters of companies in the financial services sector have already gained experience through their initial pilot projects, or are currently implementing such projects, only every second company in the manufacturing industry is planning to use bots in service processes.

RPA functions precisely and reliably, but reaches its limits as soon as it is faced with unstructured data such as emails or letters, or processes that deviate from defined rules. If, on the other hand, bot-assisted process automation is supplemented by learning software, even unstructured data can be transferred to the system. Self-learning software validates parameters, constantly learns as it goes along, and indicates when it has reached its limits. Other significant development stages are already within reach, with the integration of artificial intelligence (AI) into processes, and improved speech integration. Now, it is simply a matter of time until they reach their full potential.

It is not only the experts who agree that artificial intelligence will make its way into an ever-increasing number of sectors. Analytics company IDC estimates the current volume of the sector at $8 billion and is predicting that this figure will grow to $47 billion by 2020. “AI systems work on binary code, like conventional computers,” explains Sebastian Ostrowicz. “However, they are set up in such a way that they can perform millions of calculations simultaneously, constantly evaluate their work and are involved in the pending resolution of new challenges. The computers have a built-in optimization program, so to speak.” The artificial brain consists of thousands of computer chips. The engineers looked to nature in designing and structuring the system: similarly to neurons in a brain, the chips react to one another, send each other signals and autonomously become active or passive. In this way, huge data sets are collected in the control and nerve centers of the machines, and these data sets help the systems to learn. Taking a wide range of parameters into account, they make autonomous decisions and imitate human reasoning, as well as the resultant behavior. AI software is trained using training sets as examples. Such technology results in robots that think, shorter response times and higher customer satisfaction levels.