Angermann, H., Haenisch, T.
16th Annual International Conference on Information Technology & Computer Science, Athens, Greece
Publication year: 2020

Robotic Process Automation (RPA) is the emerging topic across research areas. This has different reasons. Firstly, RPA has proven to be an efficient technique for automatizing less and medium complex repetitive work without changing the business logic of the underlying processes. Secondly, RPA provides a broad application area, as most processes in firms are less complex. And thirdly, since applying RPA allows generating data in a standardized manner, it is considered as the entry point for later applying Artificial Intelligence. However, due to the wide application area, identifying RPA use-cases is still considered as main challenge requiring further research. The main concern is to recognize the use-cases in a structured way in order to implement those with a holistic solution. This is particularly challenging in terms of heterogeneous IT-landscapes. Nowadays, different business applications are used in an enterprise, as every system supports only a part of the relevant business processes. For example, a typical enterprise makes use of an ERP system to integrate financial activities, utilizes a MES for core manufacturing processes, applies a CRM system to manage clients, and an e-commerce application is required to allow distributing the products online. Of course, the heterogeneity becomes even more complex, when the systems have to interact with external systems. For example, when interfacing product data with marketplaces. For efficiently identifying RPA use-cases in a structured manner, the research at hand is presented. Using the proposed methodology KIM-RPA-Map, it will be possible to identify RPA use-cases in business applications context – regardless of the underlying RPA complexity and heterogeneity of the IT-landscape. To do so, KIM-RPA-Map performs in three consecutive steps. The first step focuses on the IT-landscape to find out existing interactions between business applications. Hereby, all required internal business applications are examined with regard to the individual process components (e.g. CRM social listening component). Afterwards, all single components of the different business applications are compared with other existing components to state if an interaction exists or not (e.g. ERP component A to CRM component B). Same for the external business applications, respectively the utilized services (e.g. ERP component A to the product integration platform of a marketplace). Hereby, the main concern is on the underlying taxonomies (e.g. product taxonomy), which not only allows to standardize possible interactions, but also to allow context-based reasoning for more complex (Smart) RPA use-cases and to make use of background knowledge. The second step focuses on the components of the business applications. More precisely, the identification of the required human work. To do so, all components are investigated according its main reason for use. This allows finding the use-cases having the highest potential of automation and ROI. Next, for each component, the high-level work tasks are identified (e.g. ERP: generate invoices, order new material). The third step finally focuses on the use-case and its potential. Now, all in step two identified work tasks are investigated in detail. Each task is subdivided according to the individual single work steps that have to be performed by a resource. Afterwards, each work step is formulated as user-story. Based on the single user-story, the problem is categorized according its complexity, and a description has to be given. Based on knowing the concrete problem(s), the potential for automation can be estimated, before finally stating a possible solution.

© 2021 | Dr. Heiko Angermann | | Privacy Policy