Technology is already more advanced than humans in some areas. But how can businesses benefit from this? One term you frequently hear in this context is artificial intelligence (AI).

Machine learning is a form of AI which is already being used by many businesses. In this article, we answer some of the most important questions on this topic.

Artificial intelligence is the broader term for applications that see machines perform human-like tasks, such as learning, reasoning, and problem solving. One sub-field of AI, machine learning, teaches computers to learn from data and experiences, and constantly improve task execution. Sophisticated algorithms can recognize patterns in unstructured data records such as images, texts, and speech, and use them to make decisions independently.

Artificial intelligence is the broader term for applications that see machines perform human-like tasks. This includes machine learning, natural language processing, and deep learning. The general idea is to converge machines and significant functions of the human brain: learning, reasoning, and problem solving.

This type of learning makes natural language processing (NLP) possible, among other things. NLP is the processing of texts and natural human speech, as used, for instance, in Amazon’s Alexa voice service. Deep learning uses deep neural networks with multiple layers and a high data volume. It is currently considered one of the most promising machine learning methods.

The algorithm for deep learning goes deeper than NLP: The machine recognizes and assess structures, and improves itself through several forward and backward passes. The algorithm uses numerous node levels (neurons) simultaneously to make well-founded decisions. Examples of this can be seen in medicine: Deep learning helps identify early stages of cancer and cardiovascular diseases, and can analyze a child’s DNA profile for genetic markers that indicate type 1 diabetes. In research, deep learning is used, for example, to evaluate thousands of cell profiles and their active genes, or particle showers created when proton beams collide in a particle accelerator. As this type of learning is able to solve complex, non-linear problems, it is also used in autonomous vehicles to correctly interpret complex traffic scenarios. Pedestrians, cyclists, weather, signage, and trees: the computer must correctly recognize and predict the behavior of all road users, taking all possible influencing factors into account.

When was artificial intelligence first created?

Once the potential of computers became clear in the 1950s, AI became a topic that captured many imaginations. In 1970, the “father of AI,” Marvin Minsky, stated that machines would be able to read Shakespeare in the near future, but this was not the case.

Computers were, however, celebrating success after success: first, for pure computing power. In 1997, a computer played chess with Kasparov, the World Chess Champion at the time, and won. In 2011, a computer won the game show “Jeopardy.” But it wasn’t until a machine played the Asian strategy game Go against professional player Lee Sedol, and unexpectedly won by 4 games to 1, that we first saw computers develop beyond pure computing power. Chess and Jeopardy only require a computer to understand questions and search for the appropriate answers in a database. But Go is much more complex. And it wasn’t just the speed of the computer that helped it defeat Sedol – the software was able to learn. A more advanced version of the software later taught itself the rules of the game from scratch and went on to win against its predecessor.

How do machines learn?

Machines learn either through training based on a dataset where the desired outputs are already known (supervised learning), or through algorithms that recognize patterns in data (unsupervised learning). They can also learn through rewards and punishments (reinforcement learning). This method sees the algorithm independently recognize whether learning components use the entire system (reward) or not (punishment). The data is either in a structured format, such as a table, or an unstructured format, such as a text, image, or language, as is the case in e-mails and social media posts. Machine learning can process all types of data, which is a huge advantage.

What areas can use AI?

AI is interesting for all industries that have large sets of data, for example, manufacturing companies where suppliers, sensors in machines, and the ERP system can deliver a high volume of data. In this case, self-learning algorithms support quality control and deliver forecasts for predictive machine maintenance. This is how companies can avoid production downtimes and minimize warehousing costs, to name just a few examples.

There are almost unlimited possibilities for using AI in the healthcare industry – from medical image analysis to robotic surgery. Every industry is currently generating ideas that often lead to significant increases in efficiency, as repetitive process steps are automated. This gives employees more time for creative and strategically important tasks. AI also leads to new business models. One example is if a company stops selling machines and starts selling their performance instead.

How do companies benefit from AI?

Artificial intelligence simplifies workflows, leads to more exact forecasts, and creates new business models based on data. It accelerates decisions thanks to better data, and increases the company’s ability to adjust to market changes thanks to real-time information and predictions well beyond human ability. As such, AI makes companies significantly more efficient, increasing their competitiveness.

How can companies use AI?

According to International Data Corporation (IDC), 94% of companies believe that machine learning opens up significant competitive advantages. This is not without reason. AI improves a company’s productivity and flexibility, and creates new business assets such as intelligent chatbots, used in customer service.

In this context, algorithms learn through direct contact with customers and react more precisely to their needs. SAP is also establishing intelligent processes, and offers services and applications to support its customers. They include:

  • SAP CoPilot, a digital chat assistant that asks questions and gives answers to help the user reach their goals.
  • SAP Service Ticket Intelligence. The application automatically categorizes service tickets, prioritizes tasks to be completed, and suggests answers to standard questions.
  • SAP Customer Retention predicts customer behavior.
  • SAP Resume Matching helps identify the most suitable applicants from a multitude of job applications.
  • SAP Brand Impact helps companies measure the influence of their brand and the effects of sponsorship and advertising investments. The solution records, for example, how often a company’s logo is used in a video, noting its position and size.

How does machine learning complement human ability?

Machine learning is particularly useful in activities where a person could make a mistake. Manufacturing quality controls are an example of this, as a self-learning algorithm can recognize even the smallest changes and predict their effect. Impressively, drones and satellites can be used to inspect long pipelines. Machine learning is also very important when it comes to data security, as it quickly identifies anomalies in transactions and processes, recognizes bribery attempts, and provides effective protection against hacking. It also simplifies daily tasks. If a train is canceled unexpectedly, an algorithm can readjust travel plans in a context-sensitive application. The customer can check online or in an app to see alternative routes which take them to their destination as quickly as possible even though the train is no longer running.

How can my company integrate machine learning?

Every company should know how machine learning can optimize existing solutions. You don’t have to be an expert in machine learning to use every intelligent application. Machine learning is already integrated into SAP products such as Concur and SAP S/4HANA and, using the SAP Leonardo Machine Learning Foundation, every SAP partner or SAP user can connect and implement readily prepared services, and use their own models to create intelligent applications.

SAP Cloud Platform is the foundation for the development and distribution of all types of intelligent application, and for high-performance operation. As a result, machine learning technology is available and many companies have found approaches to use it effectively.

How will AI and machine learning influence the future?

Machine learning is good at processing structured and unstructured data. A world where machines interact without commands, carry out actions that are highly context dependent, draw their own conclusions, and adjust their behavior is certainly imaginable, and is partly achievable already. It unlocks valuable time for employees to concentrate on what’s important: creativity and innovation. But AI still reaches its limits when it comes to creating ideas. In future, digital assistants and bots will support us in our everyday lives. New organizational roles will be formed to demonstrate how humans and machines can successfully work together.