Most Swiss companies are actively automating their operations using digital tools. This is shown by an extensive study carried out by ICTjournal with around sixty Swiss CIOs. From customer service to marketing, including human resources and IT itself, automation is experiencing a new boom. Several reasons explain this. First, the almost systematic use of digital tools in management, the service sector and more generally white-collar work, so that many decisions are documented and operations executed in digital form, making them manageable by the computer tool. Then the arrival of two complementary technologies: software robots (RPA) and artificial intelligence – whether machine learning, automated language processing (NLP) or, more prospectively, content generating systems. According to our survey, 29% of Swiss organizations have RPA projects and 18% are already using it. The figures are higher for AI with 25% adoption and 37% pilot projects or in deployment.
>> Exclusive survey: the adoption of RPA in Swiss companies
>> Exclusive survey: the adoption of AI in Swiss companies
Together, these two technological fields make it sometimes daring to speak of “intelligent” automation. Gartner, which prefers to speak of “agnostic automation software”, estimates that this market should register double-digit growth in 2022.
If the technologies evolve, the objectives remain the same as in the other waves of automation: gains in productivity and quality, greater stability, ability to quickly scale up or down. With the difference that today we speak more readily of freeing employees from laborious tasks, rather than replacing them with the machine.
RPA: software to automate routine tasks
Both deployed in automation projects and sometimes combined, RPA and AI have little in common. More than a technology, RPA is a category of software that makes it possible to reproduce the work done by employees on screen on a computer medium, and in particular to automate a series of administrative tasks requiring several applications, by saving part of their integration. Its use is often tactical and aims above all for productivity – a benefit targeted by 89% of the CIOs surveyed.
If a task has few exceptions, is time-consuming, and is performed entirely on digital tools, it is a candidate for automation by RPA. It is no coincidence that the deployment of these solutions often comes up against processes that are too fragmented or that it leads organizations to have to start by redesigning their processes, as evidenced by the speakers at a round table a few weeks ago. at the Swiss IT Forum(s).
After the banks which have massively used it to automate their back office, the RPA is now spreading to most companies and branches, starting with public administrations. The use of RPA is particularly popular in accounting departments and 43% of Swiss organizations using RPA have ongoing projects in the area of customer service.
AI: a versatile, complex and developing technology
Things are very different with artificial intelligence. The technology is first often used implicitly via other software (ERP, CRM) in which it is embedded. When developed by the company, projects using artificial intelligence are also more complicated. One in three organizations thus indicates a lack of skills in this area, just after the lack of usable data to train the models.
AI is then much more versatile, the automation of simple or repetitive tasks only comes in third place among deployment objectives, after improving the offer and assistance for complex tasks. Thus, 30% of AI projects aim to reduce costs compared to 43% for RPA. This versatility probably explains why the main challenge for organizations in AI is to identify a solid business case and why many of them struggle to understand the benefits of the technology.
Besides the variety of benefits, AI is also likely to invest more functions. Or to put it another way, rare are the professions of which a part cannot be supported by artificial intelligence. If, according to our survey, it is today in production that AI is most used (27%), more than one in three organizations is experimenting or has started to deploy this technology in marketing, (36% ), innovation (38%) and customer service (38%) – an area where its language processing capabilities are particularly exploited (sentiment analysis, chatbot), sometimes to serve as input for RPA.
Computers are no exception. Given the amount of data generated by systems and the growing complexity of stacks, IT departments often have no choice but to use AI (often embedded in their tools). We are thinking in particular of infrastructure and operations management (AIOps), but also of security (log analysis), application development (Github Copilot), or even artificial intelligence itself (41% of organizations surveyed using the AI use AutoML type tools).
What threat to employment?
Impossible to approach the subject of automation, without mentioning the question of employment. Even though CIOs see automation as a way to save money, only one in five say downsizing is a targeted goal. For both RPA and AI, job loss fears are only rarely seen as a major impediment to projects.
What about the facts? Looking in the rear view mirror, it appears that the digitization of recent decades has had an impact on the Swiss labor market. According to an analysis published a few weeks ago by the KOF (the Economic Research Center of the ETHZ), the cognitive routine professions (office workers, bank workers, counter workers, cashiers, secretaries, etc.) are the more likely to be automated as a result of digitization. However, these professions have experienced a decline, going from 20% of workers in 1992 to 14% in 2021. At the same time, non-routine cognitive professions (teachers, traders, doctors, etc.) have increased sharply, from from 31% to 43% of assets.
Read also >> Which jobs are threatened by “intelligent” automation?
According to the researchers, the decline in cognitive routine jobs does not translate into lower wages or increased unemployment on the Swiss labor market, unlike the phenomenon observed in the United States and Great Britain. They believe that this resilience of Switzerland can be explained as much by an extended welfare state that slows down the growth of new low-paid jobs, as by a dual education system that is more responsive to the changing needs of employers. However, they note that the disappearance of certain routine occupations has an impact on movements in the labor market, since these occupations were previously a means for physical workers to access better paid employment.
This retrospective analysis, however, does not make it possible to make a prognosis on the future impact of “intelligent” automation. With AI, it is not only laborious and routine activities that are threatened with substitution. Recruitment, translation, X-ray analysis, writing texts, creating content (like the image on the cover of this magazine), cannot indeed be described as routine tasks. And if for the time being, AI assists these activities more than it replaces them, it is legitimate to wonder both about the number of professionals who will exercise them in the future, and about the interest of their new work of “corrector of automatic systems” (see editorial).