Bill McDermott has reprogrammed SAP. Under the leadership of its American chief executive, the German software group has acquired companies for €20 billion ($21.3 billion) since 2010 so as not to miss out on cloud computing.
It’s been a successful move as the company’s quarterly figures show. Business is growing steadily, the share price and value of the company have almost reached an all-time high, and the development team, once considered lethargic, has picked up the pace.
Now, though the company’s realignment isn’t quite complete, Mr. McDermott is getting SAP ready for the next big tech trend: artificial intelligence and its sub-discipline, machine learning.
The software giant will “invest massively” and, in doing so, will become “the ultimate market leader in machine learning in the enterprise. Hard stop and bar none,” Mr. McDermott said in an interview with Handelsblatt. “We are going to drive machine learning with everything we have.”
Market research firm IDC estimates that spending on machine learning will grow to $47 billion by 2020.
The term Artificial Intelligence, or AI, conjures images of the malignant computer HAL 9000 from Stanley Kubrick’s “2001: A Space Odyssey,” or perhaps of the seductive virtual assistant Samantha in Spike Jonze’s 2015 film “Her.” But such complex systems are still fiction.
Still, scientists have made great progress in enabling computers to solve defined tasks, including voice recognition, translation, optimization of data centers and assisting in diagnosing cancer. In the process, the systems acquire knowledge from large quantities of data.
Experts have been discussing these concepts for decades, but only in the last few years have they managed to put them into practice. On the one hand, the technology is constantly getting better and cheaper. On the other hand, the mountain of data in which companies can dig for knowledge is growing, thanks to smartphones and sensors.
The “datafication of the world,” as Oxford University researcher Viktor Mayer-Schönberger calls it, is the prerequisite for machine learning.
Technology companies and market researchers alike agree that the potential is enormous. The AI age is now beginning, hot on the heels of the mobile age, said Google Chief Executive Officer Sundar Pichai, who is having virtual assistants installed into many devices.
Timotheus Höttges, the head of Deutsche Telekom, called it a megatrend, saying: “German companies have no excuse not be using AI.” The company uses voice recognition software in its call centers and uses algorithms to optimize data centers. And the possibilities go much further: leading market research firm IDC to estimate that spending on machine learning will grow to $47 billion by 2020.
Using artificial intelligence is a logical step for SAP. With its software, the company helps its customers guide production, keep control of their finances and manage personnel. SAP boasts that 70 percent of all worldwide business transactions pass through its systems, at least at some point. This creates large amounts of data, from which the algorithms can extract knowledge.
One potential application, according to Mr. McDermott, is to use an intelligent assistant to fill jobs. “Machines and computers don’t have a bias, they just have an algorithm that matches the profile of a job to the applicant.” SAP recently announced that it was developing software capable of analyzing brands in videos and images, a valuable tool for the advertising industry. In addition, customers can use SAP software to access IBM’s Watson system, a feature made possible by a joint venture between the two giants.
“We don’t need to buy a small start-up to exert our will on machine learning.”
Machine learning could also play an important role in the networking of machines, also known as Industrie 4.0 or the Internet of Things. This is an investment priority for SAP. With the help of technology, for example, it is possible to predict when a device will fail. Anticipatory maintenance could save immense costs.
SAP’s efforts are still modest compared to what technology giants like Google and IBM are doing. Some 100 of SAP’s 22,000 developers are currently addressing the issue, although that number is expected to be much higher in the future.
Nevertheless, Mr. McDermott is not eyeing any takeovers, stressing that SAP merely has to apply its own experts to the right topics. “We don’t need to buy a small start-up to exert our will on machine learning, we just needed to partition our best talent and put them on this initiative.”
According to analyst Holger Mueller, SAP is relatively well-positioned in the market for commercial software. “Everyone starts small. The competition isn’t much further along. But now it’s time to get going.”
Mr. Müller, an expert for the corporate client business at Constellation Research said, “The German company needs to make it possible to process large volumes of data with the Hadoop software, the de facto standard for big data applications. The takeover of startup Altiscale could help. On the other hand, voice recognition will now play a key role. What works on smartphones could also make sense in the office.”
Sven Afhüppe is the editor in chief of Handelsblatt. Christof Kerkmann is an editor for Handelsblatt Online and writes about the technology sector. To contact the authors: firstname.lastname@example.org, email@example.com.