Digital transformation poses major challenges for medium-sized companies in particular because, in contrast to large companies, there are often no dedicated digitization teams or subject matter experts available to address issues holistically and over the long term. This often leads to silo-like digitization measures that are initiated in the short term and whose benefits and effects on the stakeholders involved are difficult to understand in retrospect.

Long process chains (end-to-end) such as Order-to-Cash or Purchase-to-Pay, in particular, conceal challenges at countless points because they are characterized by interfaces, media discontinuities, and manual activities. All these elements are potential sources of below-average customer and supplier servicehigh personnel costs, or inefficient cash management in the context of invoicing and recording.

In creating transparency, data-driven approaches such as process mining help during three key steps (Fig. 1):

  1. Analysis-To-Cause: reducing the time from the start of process analysis to identifying the cause of the problem.
  2. Cause-To-Change: Faster and more effective insight into where, how, and why it would be optimal to perform process optimization.
  3. Change-To-Impact: Gaining transparency and speed in evaluating changes already implemented through subsequent continuous monitoring.
3 central KPIs

In cooperation with Blumatix, we are now using an anonymized example case to demonstrate how quickly a return on investment can be achieved. In doing so, we show how useful it can be to take a holistic view of the problem by combining process analysis (with Noreja Process Mining) and the implementation of an optimized invoice receipt (with BLU DELTA AI):

Prozessdashboard Purchase-To-Pay-Prozess

1. Analysis-to-Cause: A look at the dashboard (Fig. 2) for the purchase-to-pay process provides initial transparency:

    • The analysis considers 15,000 process instances.
    • There is an average transition time of 13h 18m between the activities “create delivery note” and “post invoice” (with a total lead time of the purchase-to-pay process of 5 days and 9h).
    • A high rework rate of 2.1 can be identified for invoice posting.
    • There are 384 deviations from the specified process standard.

 For a more detailed analysis, it is worth taking a look at the timeline (Fig. 3). Here you can see the transition time between “create delivery note” and “post invoice”, which was previously only shown in key figure form, and its change over time. Between the years 2016 and the end of 2018, the runtime between these two process steps deteriorated continuously from 12h 30m to 14h 35m. In addition, seasonal patterns can be seen: In January, strong transit time peaks seem to occur; In May, short-term decreases in transit time can be observed. Targeted measures should be taken to counteract the negative developments.

Purchase-To-Pay-Prozess vor Änderung

2.  Cause-To-Change: The previously created transparency regarding the cause of the problem must now be translated into concrete optimization measures. High reprocessing rates and constantly increasing runtimes indicate resource bottlenecks in accounting (invoice receipt, archiving, invoice verification, invoice release). Particularly in the case of invoice receipt, media discontinuities between the invoice and the ERP system result in a high level of manual effort, which also harbors the potential for errors.

This can be remedied, for example, with an AI-supported document capture solution from Blumatix (Fig. 4). Their solution recognizes any features via OCR fully automatically and without customer-specific configuration with an accuracy of >95%. The models are intelligent and learn with each newly captured invoice.



3. Change-to-Impact: continuous monitoring utilizing process mining makes the implemented measures of the OCR/KI-supported solution from Blumatix measurable. The result shows that the runtimes and post-processing efforts in the business process could be reduced significantly by optimizing the invoice receipt. In our concrete example, the total runtime of the process step was reduced from 14h 35m to 13h 11m (Fig. 5), which corresponds to a reduction of 84 minutes (since these are transition times, waiting times are included here). Optimizing the lead time has further positive effects on the company:

    • Reduction of throughput time lowers personnel costs and frees up tied-up capital.
    • Employee relief leads to higher motivation and productivity.
    • Reduction of regulatory risks due to less intentional or unintentional human interaction.

BLU DELTA AI Rechnungseingangsoptimierung