(The original and more detailed German version of this article is published in Controlling)
Artificial intelligence is currently regarded as one of the key technologies for the future competitiveness of companies and economic areas. It changes the life of each individual through numerous applications, but also has considerable effects on companies and their controlling. It enables the use of digital assistants, which makes communication much easier. Artificial intelligence improves forecasts and business planning. It also helps to monitor business more efficiently. When used correctly, artificial intelligence has the potential to significantly increase a company’s competitiveness.
What is artificial intelligence?
Artificial intelligence (AI) comprises a large number of methods with which human intelligence is to be simulated. This enables computers to perceive and understand facts such as the recognition of objects in images. On the other hand, they can learn from feedback and thus improve their ability to perceive and understand contexts.
In recent years there has been considerable progress in the field of artificial intelligence. In areas where cognitive performance can be measured and mapped by fixed rules, computers are already superior to humans. These include, for example, board games such as chess or go. IBM’s Watson AI system caused a worldwide sensation, defeating two extremely successful human participants in a TV quiz show in 2011.
Artificial intelligence draws on an extremely broad range of methods, including so-called machine learning. This is regarded as one of the central technologies of artificial intelligence. In machine learning, knowledge is generated from experience by developing models based on existing data that can be continuously improved by new data. Such data can be, for example, texts, images, language information or sensor data, but also company key figures.
There are two important prerequisites for the application of artificial intelligence methods: On the one hand, large amounts of data must be available, because this is the only way to train algorithms and achieve improvements that come close to intelligent behavior. On the other hand, the methods usually require high computing power. This is the only way to achieve an adequate processing speed that is necessary for many applications.
Digital Assistants in Controlling
Digital assistants are an important and rapidly growing field of application for artificial intelligence. Apple’s Siri and Amazon’s Alexa have now reached private users. Digital assistants can answer oral questions and contribute considerably to the simplification and acceleration of processes, because many routine tasks no longer require the written input of the corresponding information. For example, instead of tediously dialing a telephone number, only the name of the person to be called has to be mentioned and the connection is established automatically.
In controlling, there are numerous fields of application for digital assistants. This is because controlling already often plays the role of a business partner who has to answer questions from the operative units and advise them. The acceptance of an order or the price quotation for a customer are examples of such inquiries. This is because such decisions usually have to be made quickly, and those responsible are dependent on rapid and reliable information as well as methodological competence provided by Controlling.
Due to their ability to recognize language, translate it and access an extremely large database, digital assistants can answer such queries in real time and with a broad database. They can be used by both decision-makers and controllers. On the one hand, they partly take over tasks of the controller, which are currently still associated with a considerable effort and thus give him more time for the consultation of decision makers. They can send numerous inquiries directly to the digital assistant. On the other hand, digital assistants can also make the life of controllers easier by taking over partial analyses for ad hoc management queries.
An IDG study comes to the conclusion that digital assistants are already used or planned to be used by 50% of the companies surveyed. These include decision support as well as internal and external customer enquiries.
Artificial intelligence improves forecasts
One of the core tasks of Controlling is the determination of forecast values for central success factors of a company such as sales and turnover figures or profits. This so-called forecast takes up a considerable part of the daily work of the controller. The reason for this is that the forecast forms the basis for many planning processes, such as planning production capacity or marketing expenses. In addition, it shows deviations from the planning at an early stage. This means that countermeasures can be taken in good time if any disadvantageous deviations are detected.
Machine learning methods are already very mature in terms of their ability to deliver good forecasts. Studies have shown for numerous fields of application that forecasts based on artificial intelligence are often superior to man-made forecasts. Both statistical methods and linear regression, in which the forecast values are determined from a large number of influencing variables, are used. In many situations, however, more sophisticated methods such as artificial neural networks, which can be trained with large data sets and often deliver better results in complex forecasting processes, are also useful.
All methods have in common that the forecast values are always based on previous knowledge. Unpredictable changes, for which there is no indication in the previous data, cannot be predicted even with methods of machine learning. Therefore, it makes sense to use a mixture of artificial and human intelligence in important forecast processes. While artificial intelligence provides an automated forecast, humans can intervene in the process at any time and change the forecast value through a personal assessment.
In the meantime, there are several providers of business software that use artificial intelligence methods in their solutions for controlling questions. For example, the enterprise software Unit4 Prevero uses artificial neuronal nets for the determination of Forecast values. These can be determined automatically, can be replaced however at any time by another Forecast method like for example a linear trend or be corrected by the personal estimate of the decision maker.
More efficient business planning with the help of artificial intelligence
The coordination of corporate planning is one of the most important tasks of controlling. This requires the coordination of a large number of sub-plans, such as sales, production and procurement planning, as well as financial and liquidity planning. Corporate planning ultimately involves making a large number of decisions based on confusing amounts of information. Artificial intelligence is already used today in individual areas to support planning and has the potential to play a central role in corporate planning in the future. According to the IDG study, 41% of the companies surveyed already use artificial intelligence systems in their planning systems, and a further 21% are planning to do so.
One area in which some pioneers have already implemented artificial intelligence is price planning. AI makes it possible to consider a multitude of factors that can have an influence on the level of optimal prices and to adjust them dynamically and constantly. These include, for example, competitor prices, own stock levels and purchase prices, regional customer preferences, time factors, seasonal trends or even the weather. Petrol stations, airlines, hotels and numerous webshops are already using such dynamic price adjustments. Artificial intelligence methods provide for automation and an increase in turnover and profit.
A study by Deloitte comes to the conclusion that German companies often use artificial intelligence as a standardized service. 65 % of the surveyed companies use artificial intelligence “as a service”, while 15 % of the study participants rely on a company-internal implementation. A majority of 61% of the companies surveyed already use business software with integrated artificial intelligence. This opens up a simple possibility for small and medium-sized enterprises to use artificial intelligence without having to build up corresponding development capacities.
Artificial intelligence allows more precise monitoring
In addition to corporate planning, an important task of controlling is the ongoing monitoring of business and the development of adjustment measures in the event that corporate goals are at risk. This applies on the one hand to financial data such as sales and costs, and on the other hand to non-financial key figures such as cancellation rates or machine running times.
Artificial intelligence methods can contribute in many ways to a more efficient and precise monitoring of the business. This already starts with the fact that such a system can check manually entered values for plausibility and thus ensures that the data quality of a company’s information systems increases. The possibilities of pattern recognition in large amounts of data mean that companies can use artificial intelligence to identify customer migration at an early stage and react with countermeasures. Another important monitoring task is to ensure machine and computer availability. Here, artificial intelligence based on sensor technology and the resulting data can prevent imminent failures and automatically initiate maintenance measures in the right time windows.
Artificial intelligence requires responsible action
Artificial intelligence is used in the first applications in controlling and is about to fundamentally change controlling and a multitude of business models. In retail, for example, customer needs can be predicted much better and an individual customer approach with a comprehensive range of solutions can be realized. Such new business models require a new management philosophy that focuses on the customer solution rather than the product. Controlling has to deal with the changes caused by artificial intelligence and analyse the consequences for strategy, ongoing business and everyday work in controlling.
At the same time, the methods of artificial intelligence must be critically scrutinized. Even if the use of large amounts of data leads to better corporate decisions, it must be ensured that their use is ethically justifiable and that the corresponding legal regulations on data storage and use are complied with. This applies in particular to data obtained from social networks. In this area, companies must act with particular sensitivity in order to take appropriate account of society’s and individuals’ desire for privacy.
Companies can gain a considerable competitive advantage with such a responsible approach and the courageous use of the possibilities of artificial intelligence.
This is a repost from the original article on LinkedIn.