No ERP system is offering interfaces for artificial intelligence (AI). So several carmakers are cooking up ways to bring AI into their ERP platforms on their own. Automotive companies such as Volkswagen are experimenting with artificial intelligence in their labs. They have been setting up new AI-based processes on the operating engines of their ERP systems for a while and are working with neuronal networks and programming languages such as R or Python, AI libraries and AI frameworks from Microsoft, Amazon and Google. The question remains: Why are business software providers dealing with the issue so tentatively? Here’s a look at what three providers of ERP systems are planning.
SAP has intelligence but no true AI functionality in its product line. But the industry leader is confident this will change. “I see a great future for artificial intelligence in ERP systems,” said Carsten Polenz, who is responsible for Industry 4.0, IoT strategies and innovations at SAP. “The great strength of artificial intelligence lies in pattern recognition, which can help to auto- mate routine processes and tasks,” he said. That opens up the next logical step to the further development of ERP. Polenz said SAP is looking into deep learning opportunities. “Our research department is addressing the neuronal networks’ opportunities and possibilities within the framework of ERP software,” he said. Production planning is an important area of application. “In the future, AI algorithms and methods will augment many heuristic processes that today are used to fine-tune optimization algorithms,” he said. “Human empirical knowledge in this area will be supplemented with artificial intelligence.”
The German ERP specialist acknowledges that it is still in the early stages of AI incorporation. But it is confident that the technology will play a bigger role soon. “It is 100 percent certain that we will be moving toward AI,” said Karl M. Troeger, who heads Psipenta’s business development. At the moment, though, AI integration isn’t part of the systems on offer. Planning and forecasting areas are especially important starting points for AI integration. “Medium- and long-term planning in the old sense will no longer suffice with the implementation of Industry 4.0 concepts and a growing number of orders for low production runs,” he said. It’s an open question how quickly AI will start playing a bigger role. For example, experts are unsure whether enough big data are really available to induce neuronal networks to learn. It is highly conceivable that AI components could be incorporated in the classic ERP system with material needs planning, bill of material explosion and work plans. “For the approach to be successful, the AI must be intrinsic,” Troeger said. The Psipenta executive said customers’ lack of expertise and unfamiliarity with the new technologies also is holding back development of AI processes. “We are a reflection of our customers,” Troeger said.
The US business software group also is keen to include more AI into its ERP platforms. “Sometimes things take many years to mature,” Infor CEO Charles Phillips told a conference in July. “This technology is starting to creep into our lives and train people to get used to it.” Infor this year announced a business AI platform called Coleman, which uses machine-learning to help optimize and streamline operations. Still, today’s ERP systems do yet have dedicated AI functions, said Infor’s Matthias Sartor. Up-to-date applications are already highly automated and, for example, initiate cross-functional work flows featuring event- or release-driven operations. Infor is one of the first ERP providers that has turned to a cloud model. With its Infor ION platform, even systems from other companies can be integrated into an ERP infrastructure. “It makes sense to use an AI-driven front-end system – for customer service, for example,” Sartor said. Digital assistants of this kind can use data from the ERP system intelligently. AI in the ERP system can help employees obtain more resources for actual value-creation processes.