As a manufacturer of more than two million sensors and systems annually, Endress+Hauser is a global market leader in measurement instrumentation and automation technology for process and laboratory applications. Devices, solutions, and services of the Swiss company are at home in many industries. Customers thus use them to gain valuable knowledge from their applications. This enables them to improve their products, work economically, and, at the same time, protect people and the environment.
Almost every instrument delivered is unique and adapted to the customer’s individual requirements. Analyzing and entering the customer’s configuration requirements for the instruments into the respective systems was a time-consuming manual task. However, this changed during a co-innovation project with the SAP AppHaus Network member sovanta. With the help of SAP Business Technology Platform, sovanta’s tooka.ai assistant now automates the time-consuming and error-prone task powered by artificial intelligence (AI).
Before: Error-prone manual paperwork
Endress+Hauser’s core competence is to provide customers with devices of the highest quality, taking into account their individual requirements. However, necessary preparatory work prevents employees from fully concentrating on this core competence. Before the instruments can be configured based on the specification data, the requirements must be processed manually. For one project, between 100 and 10.000 data sheets are provided by the customer. The format, layout and description of the attributes vary between customers and projects. In collaboration with sovanta, the challenge of the time-consuming and error-prone import of the specification data had to be solved. An ideal use case for sovanta’s new tooka.ai assistant powered by machine learning (ML).
tooka.ai in a nutshell
tooka.ai is a factory for digital assistants based on artificial intelligence. Endress+Hauser’s assistant automatically transfers requests and information to the right place within the SAP system with the help of AI. Data that come in via emails, PDFs or images are classified, contextually understood, transferred to the SAP system, and finally filed. Manual clicks become obsolete. Using a dashboard, the activities of tooka.ai can be monitored and corrected, so that it quickly learns and improves. The resulting process optimization can be achieved in order entry, cost classification, ticket assignment, invoice approval, master data maintenance and other repetitive and time-consuming business processes. Adaptable to any use case, tooka.ai is quickly and easily ready for the individual use case within a timeframe of six weeks. So how does tooka.ai support Endress+Hauser today?
After: Automation of requirement data processing
In light of Endress+Hauser’s initial business challenge outlined above, the answer is obvious. Communicating with SAP Business Technology Platform tooka.ai takes over the processing of customers’ configuration requirements for the instruments. For this purpose, tooka.ai is provided with the customer’s specification data as a trigger to proceed further. With the help of a machine learning algorithm, it decides how to process the data so that it can be transformed into the right output format. Customer attributes, such as the units used by customers (e.g. °C/Temp/Temperature) are automatically identified. Ultimately, the specification data is imported, in a fully automated fashion, into Endress+Hauser’s SAP Business Technology Platform. Employees can monitor and correct the processing in the project dashboard, at any time.
“Our customers send us specifications with measurement requirements every day, and we have to transfer these from PDFs, images or text files into our project format. This process is very time-consuming and error-prone. With tooka.ai we can already automate 60% completely! But we have set ourselves much higher goals, which we are getting closer to every week together with the team from sovanta.”– David Haener, Program Manager Engineering, Endress+Hauser Group Services
High optimization level with a steep learning curve
Endress+Hauser doesn’t supply mass products. The business segment of the Swiss company is particularly characterized by customer orientation and consideration of their requirements – a special challenge for tooka.ai, which it, however, solved with excellence. The recognition rate currently lies at 60% and benefits from continuous improvement, as the assistant learns with each new import. Additionally, as specification data sheets are standardized in the future in terms of format, layout and description of attributes, the automation rate will continue to climb up. Since its introduction, tooka.ai allows the sales and project teams to focus primarily and progressively on their value-adding core competencies, thus further ensuring the success of the company.