IoT
Everyone is talking about artificial intelligence or AI. As I find the topic very exciting, I wanted to investigate the extent to which this also makes sense within an ERP solution and how far we have already come in this area. And since my passion is SAP, I did this using SAP S/4HANA.
There are already numerous supporting processes relating to ERP that use AI and ultimately work with ERP data and interfaces. Together with ETH/EPFL, Swisscom has developed various topics such as automatic sorting/delivery of incoming e-mails in customer service, intelligent classification of scanned documents and "schwiizerdütsche" speech recognition. From the Industrial Internet of Things (IIOT), we are familiar with the predictive maintenance solution, which uses a large amount of (sensor) data to predict when maintenance is due. To do this, customer and equipment data is pulled from the ERP and new service orders are automatically created in the ERP depending on the result of the calculation.
SAP has now integrated various self-learning algorithms into the ERP in its latest version of S/4HANA 1709.
The system automatically allocates incoming payments to the invoices. If the customer provides the customer and invoice number when making the payment, e.g. via ISR, this already works without artificial intelligence. However, if this information is missing and, for example, only the customer name is available in the text or fragments of the reference, manual intervention was previously necessary.
The system is now trained on the basis of past assignments. To do this, about 75% of the data is used to learn. The accuracy of the assignment is then checked with the model created by simulating the remaining 25% (of which the assignment is known) through the model and comparing it with the expected result. This allows you to calculate the statistical accuracy.
When new payments are received, the model is used. If the accuracy is below the threshold value (e.g. 90%), manual approval of the allocation proposals by the clerk is required; all other allocations run automatically and can be displayed and cancelled in a log at any time.
The SAP HANA database has machine learning capabilities. However, the algorithms are usually connected to the ERP via APIs and the cloud so that you can programme in an agile manner alongside the ERP (no need to upgrade the core ERP) and use the latest libraries on an ongoing basis.
A second example is the predictive purchasing contract analysis. Here, the system predicts the expiry and utilisation of existing contracts (e.g. volume contracts, etc.) so that you can benefit from better conditions and renew contracts in good time.
Many more intelligent functions are already planned for the future, for example:
| Consumer Retention | Integrated Digital Content Processing for Content Management |
| Brand Impact (Marketing) | Quotation Conversion Probability Rate |
| Reduce Free Text Item Purchases | Finance Predictive Accounting (Richtiges Konto aufgrund der Erfahrungen) |
| Procure: Invoice Payment Block | Customer Behaviour Segmentation |
| Forecasting Key KPIs in Accounts Receivables Management | Delivery Performance/in Time |
| Account Payables Mgmt and Cash Discount Analytics | Email2Record |
| Cash and Liquidity Management | Product Compliance |
| Produce: Defect Recording (inclusive Text Recognition, Image Recognition) | Customer Retention |
| Predictive Physical Inventory Counting | HR Resumee Matching (Bewerberprofile) |
| Automation in Source of Supply Assignment (Bezugsquellenfindung) | HR Learning Recommender (Ausbildungplanung) |
| Smart Variant Configuration: Merkmalsvorschläge aufgrund der bewährten Konfigurationen | Project Forecast based on historical data |
| Cognitive Operational Procurement | Solution Recommendation |
| SAP CoPilot (Mobile Assistant, quasis ein Business SIRI) | Service Ticketing |
This is going to be exciting! With its expertise in artificial intelligence combined with SAP know-how, Swisscom is naturally ideally positioned as a consulting partner for our customers. Not only are SAP standard algorithms used, but the combination with other enablers is possible at any time, so we can also score points against the international software group, especially with Swiss specialities, or develop new automations with artificial intelligence together with our customers where there is no solution yet.
Chief Solution Architect
Find the job or career to suit you. A career where you can make a difference and continue your personal development.
What you do is who we are.
