Just under half of respondents from Western Europe have already introduced AI initiatives, but so far mostly only for a limited area of application. But this is set to change.

Although for many companies AI is still a distant dream of the future, the maturity of intelligent technologies and the expectations placed on enterprise resource planning (ERP) systems have fundamentally changed business requirements. ERP systems must lose their clumsiness and be open to third-party applications so that companies can benefit from technological innovations more quickly. This is a requirement that - according to the experience gained from project practice with Epicor ERP systems- ERP providers and user companies alike must meet. This becomes clear in the following expectations for AI-enabled ERP.

1. platform concept for more development options

Cloud-based platforms allow users to combine different functional modules with one another - similar to app stores for smartphones. Companies that do not yet want to rely on a cloud model for ERP also have this requirement. ERP systems that have the same software architecture with a platform concept as an on-premises version and as a cloud version are therefore of interest. This means that companies do not face any technical hurdles when switching to the cloud in the future, but they can already benefit from the advantages of a platform structure today.

2. flexible process for the introduction of intelligent technologies

Intelligent ERP is based on the system's ability to access a variety of technologies for intelligent functions. In companies, this usually starts with automating repetitive routine processes in the background. Further functions are added step by step - such as machine learning through to deep learning for error detection in production, predictive analytics for demand planning and forecasting, natural language processing for voice-controlled systems in production or in the warehouse through to learning algorithms for autonomous systems. This approach ensures the cost-effectiveness of innovations and simplifies their acceptance in everyday working life.

Bitkom also assumes that "ERP systems, as a digital process and data hub in companies, will be increasingly enriched with AI technologies". ERP systems must therefore be prepared for the fact that innovations of this kind can be gradually integrated and further developed at any time as required.

3. scalable web services

Companies demand intelligent technologies. © Epicor

Intelligent systems require considerable amounts of data in order to recognize trends, patterns and correlations, including the ability to learn from them and draw conclusions. The IT systems required for this are only economically viable as cloud or web services. This has two consequences: Access to intelligent technologies is open to companies of all sizes and is no longer a question of IT budget. Technologically, however, this requires modernization to an ERP system whose software architecture is based on a corresponding industry platform, such as Microsoft .NET, in order to be able to seamlessly access Azure resources.

With an ERP platform that communicates with Microsoft Azure, for example, it is comparatively easy to connect company processes with intelligent cloud services and AI functions. Internet of Things (IoT) functions can be made available as "plug and play" to seamlessly connect IoT devices to the cloud. According to a recent Deloitte study, 65% of those responsible for AI in companies surveyed in Germany rely on "off-the-shelf AI". According to Deloitte, 61% of German companies also rely on enterprise software with integrated AI.

4. domain expertise for faster deployment of intelligent systems

Even if access to intelligent technologies via web services is hardly an obstacle in technical terms, they are still not an out-of-the-box solution for industrial use. Bitkom described what counts here in its report on "Artificial intelligence and ERP": "The actual service of the ERP provider is to deliver pre-developed models, which are then individually expanded for each customer. A critical factor is the semantics of the ERP data. This must be 'taught' to an AI by the provider." It is therefore about the domain expertise of the ERP provider for certain industry requirements or functional areas, which is crucial for ensuring that the systems achieve the required optimization reliably and quickly.

5. significantly higher operational efficiency

Managers are currently concentrating on optimizing their own company. Organic growth is the credo, new markets or strategic alliances play a subordinate role. As Western European companies have usually already achieved a high level of efficiency with established IT systems, the expectations regarding the results, prospects and innovative edge that an ERP modernization must deliver are considerable. If upgrades to the existing system are pending, a fundamental modernization is often considered in preparation for intelligent technologies.

6. faster system implementation

ERP implementations must reach "Go Live" status in just a few months, but at the same time keep all doors open for further development in the direction of AI. Industry-specific functions combined with industry standards in the software architecture are the prerequisite for ensuring that a lean ERP installation does not become a technical dead end. According to Bitkom, it is foreseeable that AI will change the ERP landscape and become a key competitive factor. Until now, the task of ERP has been to organize information, generate reports and visualize real-time data. But companies are demanding intelligent technologies to recognize trends and patterns, provide recommendations for action and automate complex processes. The degree of freedom offered by modern ERP platforms will help to meet these requirements. This is because they enable networking with the associated process improvement and thus create the prerequisites for the use of intelligent technologies.

Dirk Löhmann, Regional Vice President Continental Europe at Epicor Software / am