Back to overview

November 2025

AI-Powered Receipt Analysis at DVS AG

From an unstructured document landscape to usable insights.

Upload documents

Extract data

Classify content

Return structured data

4M+
documents processed
GDPR
compliant architecture from day one
Scalable
stable even at high volumes

Deutsche Verrechnungsstelle AG

Deutsche Verrechnungsstelle AG (DVS AG) receives an enormous volume of invoices and receipts from its customers every month. These documents contain valuable information but, due to widely varying formats and layouts, could not be reliably extracted, collected, and aggregated. Together we developed an AI-powered solution that processes documents on upload, detects, categorizes, and structures the relevant data points. This creates the foundation for reliable analyses, new insights, and further use cases.

Challenge

Four requirements had to work together from the start. None could be bolted on later.

Very high data volume

Over 2 million documents in the first month made a robust, efficient, and scalable infrastructure absolutely essential.

Heterogeneous receipts & invoices

Different structures, layout variants, and contents were a challenge for reliable, consistent processing.

High quality requirements

Since the analysis results are used for real insights and decisions, consistently high extraction and classification quality of the AI models is essential.

Data protection & compliance

Processing highly sensitive financial and invoice data required a rigorously GDPR-compliant architecture, secure data processing, and strict access and security concepts from day one.

Solution

We built a system that automates and standardizes DVS AG's document flow:

Previously isolated receipts become standardized, actionable datasets.

  1. Simple upload workflow for incoming receipts and invoices

  2. AI-powered extraction of relevant data and line items

  3. Classification into uniform categories for better comparability

  4. Structured return of results in clearly defined data formats

  5. Scalable infrastructure that handles high volumes reliably and performantly

Outcome

A barely usable document stream became a reliable data foundation.

Standardized

Data from highly diverse receipts is now standardized, searchable, and aggregable

Actionable

DVS AG can derive reliable analyses and new insights on this basis

Automated

The previously barely usable document stream has been turned into a scalable, automated process

Working with Artesiana has been a very positive experience for us. What impressed us most was how quickly a productive system came together and how flexibly the team responded to feedback. The tool reliably handles our highly varied receipts and now delivers high-quality data that opens up new possibilities in our day-to-day work. We see a lot of potential in this collaboration and look forward to implementing further projects together.

Do you have a similar project?

We assess your document process and tell you openly whether and how automation can be implemented effectively.

Book a meeting