Tax consulting in Germany is undergoing a fundamental transformation. Firms traditionally provide primary services such as accounting, payroll, annual financial statements, and tax returns. These processes generate large amounts of structured data that have previously been largely ignored as valuable assets. However, in the digital context, they are evolving into a strategic resource.
DATEV, as a central platform, already stores virtually all of its clients' business-relevant data. The Nuremberg data center stores accounting records, wages, receipts, tax returns, and communication processes. Formally, this data continues to belong to the clients, while the law firm acts as the controller and DATEV acts as the processor. However, this legal construction obscures a crucial dimension: The actual power over data is not legal ownership, but "possession" – the ability to access, control over formats, and the power to exploit data.
“Property” vs. “possession”: The underestimated tension
The distinction between "ownership" and "possession" of data is crucial for the strategic positioning of tax firms. Ownership describes the legal control that undoubtedly lies with the clients. Possession, on the other hand, describes who actually has access, who controls interfaces, and who defines how data is used.
In practice, many law firms do not fully control the data. They only receive access provided by DATEV or other platforms. Thus, ownership—and thus the actual power—lies not with the clients or tax advisors, but with the platforms.
The risk is clear: law firms will become feature users whose services are integrated into the feature library of others. One could also say
“Unless you have a clear path to proprietary data or workflow ownership, you're building a feature that platforms will absorb.”
Vendor lock-in as a structural risk
This dependency is a classic case of vendor lock-in. Data and workflows are so tightly integrated into proprietary systems that switching is virtually impossible. This is particularly evident at DATEV: Data is stored in proprietary structures, exports are limited and complex, and value creation through new functions is increasingly shifting to the platform.
In the long term, this means that AI-supported booking suggestions, industry benchmarks, and reports are created centrally at DATEV or other platforms, while law firms act only as users. One could also say:
"These aren't tax firms with a database. They're service engines designed to collect proprietary, high-value datasets as a byproduct of client work, then monetize that data in adjacent markets."
The law firm runs the risk of providing data but giving up control over the resulting added value.
Three pillars of sovereignty: portability, semantics, AI
The solution lies not in avoiding platforms, but in consciously using them as infrastructure and systematically reducing dependencies. The strategic response can be summarized in three pillars:
1. Create portability
Data portability means that law firms must have a complete, structured export of their data at all times – regardless of the platform. This requires automated ETL/ELT processes that regularly extract data from DATEV (or other platforms), banking, e-commerce, payroll, and other sources and load it into a law firm's own data room.
Technologies like Microsoft Fabric with OneLake storage provide a suitable infrastructure for this. The key difference compared to DATEV lies in the format: OneLake stores data in Parquet/Delta by default. This open, column-oriented format is highly compressed, efficient, and readable by virtually all modern data and AI systems—from Spark to Databricks and Snowflake to open-source frameworks like Pandas or DuckDB. This ensures that the data is not locked away in a proprietary format, but is exportable and portable at any time.
This creates a fundamental difference to DATEV: While DATEV data remains in proprietary silos without exports, law firms can build a portable database in Fabric that remains usable in the long term, regardless of the platform.
An exit strategy is also part of the security: Regular backups in Parquet format should be stored outside of Microsoft, for example, in AWS S3 or on-premises. This way, the law firm remains resilient even in the event of changes to Microsoft's pricing model or technology policy.
2. Check semantics
Data alone has no value if its meaning is defined by others. Therefore, law firms must establish their own semantic layer.
This means that it is not DATEV or Microsoft that determines what constitutes "liquidity," a "risk profile," or an "industry comparison," but the firm itself. Proprietary KPIs, taxonomies, and benchmarks transform raw data into proprietary knowledge models.
For example, if a law firm defines and systematically maintains industry-specific key performance indicators for physiotherapy practices, construction companies, or e-commerce shops, a separate, irreplaceable semantic capital is created. DATEV thus becomes the data source, but the actual intelligence lies within the law firm.
3. Path to your own AI solution
The third and most important pillar is the development of custom AI solutions. Once data is available in Parquet/Delta's own data space and has been enriched with a semantic layer, it can be used for custom AI models.
This opens up three central applications:
- Anomaly detection: AI models detect anomalies in bookings or payment flows before they become a problem.
- Forecasts: Cash flow and liquidity forecasts, scenario analyses, or tax burden forecasts trained on cross-client data.
- Benchmarks: Anonymized, aggregated data pools enable comparisons between clients in the same industry and provide individual recommendations for action.
The crucial point: These AI solutions are developed within the firm. They are not a standard feature of DATEV or Microsoft, but rather proprietary products that clients only receive within that firm. This creates a new business model and a differentiation that frees them from interchangeability. Yes, platforms will certainly also offer such AI solutions – as far as legally permissible. But under their own terms and conditions.
Compliance as a strategic advantage
A frequently expressed concern is: Are tax advisors even allowed to use such data in this way? Here, a perceived obstacle turns out to be a strategic advantage.
Tax advisors are subject to particularly strict data protection and confidentiality requirements under the GDPR, Section 203 of the German Criminal Code, and their professional code of conduct. This builds trust. If a firm organizes data spaces in such a way that pseudonymization, aggregation, and purpose limitation are consistently implemented, a compliant and trustworthy framework is created. The contractual relationship exists between the client and the firm—not between the client and the platform, even if some platform operators would like it that way.
While (international) platforms are often viewed with skepticism, this is precisely where law firms can score points: In the future, they will become natural advisors – not only in tax matters, but also in data-driven matters.
The path to becoming a Service Engine Operator
The future of tax consulting will not be determined by the question of ownership, but by the possession and usability of data. Ownership formally lies with the clients. But whoever controls portability, semantics, and AI determines value creation.
Law firms that build on these three pillars use platforms as infrastructure without becoming locked into their systems. They secure access through open formats, create their own semantic models, and develop proprietary AI solutions. This transforms them from interchangeable feature users to service engine operators who not only manage data but actively translate it into new value creation.
Or to put it another way: Tax advisors who design their data rooms will meet platforms on equal footing. Tax advisors who fail to do so will become part of a feature defined by others.
