Artificial intelligence (AI) in taxation is already a reality today. The prerequisite for this, however, is a sufficient digital maturity of the respective tax function, which must always be checked before using AI. The latest international AI study by the international tax practice, WTS Global and the German Research Center for Artificial Intelligence (DFKI) shows how the current status of digitisation is in taxation worldwide, and what a maturity model may look like in order to determine “AI readiness”.
As part of the study, WTS Global and the DFKI interviewed tax experts from a total of 34 countries on the digitisation status of tax functions by means of the international tax practice, WTS Global. In this regard, 85 per cent of the study participants stated that they were dealing with this topic in a professional context. However, the specific use of AI in taxation is currently only considered realistic by 5 per cent of those questioned.
“Tax functions have long recognised the need to digitise and automate tax processes, and are investing more and more in this area on an international scale. However, they are still reserved in terms of AI. Although intelligent tax solutions offer enormous potential, the organisational and IT infrastructures of most tax functions are not yet equipped for AI application”, explains Wim Wuyts, CEO of WTS Global.
Obstacles for artificial intelligence in taxation
For the application of AI in the field of taxation, certain conditions must be met. For example, the availability of data and the nature of data organisation play a decisive role. At this point, the discrepancy between the actual and ideal state becomes clear. For example, only 46 per cent of those questioned stated that their clients had tax-relevant data in digital form which could also be used for the purpose of process automation.
“The digitisation status of tax functions is very different. Especially when related to data as there is a great need for optimisation. For intelligent tax solutions to deliver their full potential, they need access to a unified data collection (“Tax Data Lake”). Often, however, separate data silos are used,” commented Prof. Dr Peter Fettke, Scientific Director of the study and the Competence Centers Tax Technology at the DFKI.
Other challenges identified included the lack of a clear digitisation strategy and the lack of a budget for its implementation. According to study participants, this applied to about two-thirds of all clients. But even the fact that business processes in tax departments are not always completely digitised by IT systems is a limiting factor – this is only the case for 18 per cent of companies.
Anyone wishing to use AI, not only needs new technologies and solutions, but also has to build up appropriate know-how on behalf of their employees. Many of the tax experts questioned had a basic understanding of AI, but no experience in greater detail in this area. As a result, AI training and a corresponding change management process are gaining in importance.
When intelligent tax solutions are used and what they do
With intelligent tax solutions, tax functions are able to set new standards in terms of quality, efficiency, compliance and cost savings. The application areas are currently focused on certain tax disciplines, in which large amounts of data are processed and the tax tasks are highly repetitive – such as wage tax, VAT, transfer pricing or customs. For example, process mining methods can detect errors and anomalies in mass transactions that are difficult for humans to detect.
“The study revealed a great need for approaches to uncovering unknown process patterns and anomalies. The process mining methods can be used precisely on this occasion, in order to identify real process flows and, for example, with regard to compliance issues. To assess the conditions for the use of AI processes, such as process mining, and to develop them specifically, a maturity model helps,” says Tim Niesen, who is writing his doctoral thesis at the DFKI on AI maturity models in the tax department.
Getting AI Ready with the maturity model by WTS Global and the DFKI
Conducting a study on the degree of maturity is an important basis for being able to use AI later on and to become AI Ready. With that in mind, WTS Global and the DFKI have developed a systematic assessment methodology for assessing the digital maturity of tax functions based on the study results. At its core, this is divided into four categories: Strategy, Data, Processes and Technology. This makes it possible to identify drivers and obstacles of digitisation in tax functions and to obtain a detailed overview of their digitisation status in the four areas.
“Before tax functions start purposefully with AI, they first have to become aware of their digital maturity. For this purpose, we have developed a model with which we can capture the digitisation status of tax functions, identify AI potentials and finally derive recommended actions for AI readiness,” says Vanessa Just, Project Manager, Artificial Intelligence at WTS Germany and Managing Director of wtsAI.
In the future, by means of the wtsAI joint venture which we founded in mid-2018 with the advanced analytics specialist, QUNIS, WTS Germany will conduct maturity research into corporate tax & finance and help companies to achieve AI readiness.