DATEV processes 2,8B digital documents a year from round about 1M German SMEs and public organizations. One of the most common but also time-consuming jobs for the tax advisors and their employees is accounting and taxation of different types of accounting documents. Each document needs to be classified regarding the underlying tax issue to produce valid accounting records that correctly represent each business case while also fulfilling legal obligations.
Therefore, DATEVs goal is to improve efficiency for its customers within their daily business by reducing the amount of time spent working on recurring tasks. To achieve this goal, DATEV develops solutions supported by machine learning and artificial intelligence, which aim to automate these customer processes. Quality and explainability play a central role in these automation solutions to minimize errors (and hence costs) as well as to fulfill the (legal) obligation of tax consultations and other business supporting roles to check the validity of (automated) processing steps.