Data quality in Treasury

Standardization, automation, advanced analytics and the use of artificial intelligence promise to make Treasury processes more efficient, faster and more forward-looking. Yet discussions often focus primarily on new technologies and specific use cases, while their actual value is frequently determined by a less visible factor: the quality and availability of the underlying data.

Even the most powerful technology can work only with the information available to it. Inconsistent data structures, missing interfaces, different formats or unclear process ownership can prevent organizations from realizing the full potential of automation and cause AI applications to fall short of expectations. As a result, the existing data architecture is increasingly becoming a critical success factor: it can either enable or constrain technological transformation.¹

While Treasury-specific information was the primary focus in the past, modern management approaches require a much broader view of the corporate data landscape. Relevant insights increasingly emerge at the intersection of financial, corporate and process data, extending far beyond traditional Treasury systems.

A reliable data foundation therefore requires information from different areas to be consolidated and placed within a shared business context. In Treasury, this includes bank master data and account hierarchies, cash flow and liquidity positions, payment transaction data and market and counterparty data. A liquidity forecast, for example, can deliver its full value as a management tool only when planning data, operational cash flows and actual account movements are brought together consistently. A uniform data foundation is equally important in counterparty management: unambiguous master data and identifiers such as the Legal Entity Identifier (LEI) facilitate the consistent allocation and aggregation of counterparty positions.

What matters is not only the availability of data but, above all, its consistency, timeliness and traceability. Historically evolved data models, inconsistent definitions, such as differing interpretations of available liquidity, and a lack of standards often make it difficult to use information consistently. The ability to consolidate data systematically and safeguard its quality is therefore becoming a fundamental requirement for a data-driven treasury function.¹

Although many Treasury processes are already supported by digital tools, providing decision-relevant information still often requires considerable manual effort. Data must be collected from different source systems, validated and prepared before it can be used for reporting, analysis or management purposes.

Inconsistent definitions, differing or even contradictory data sets and limited transparency regarding the origin and quality of information can lead to inconsistent interpretations of analyses or decisions being made on an insufficient data basis. In Treasury, this is particularly evident in key management metrics:

• Missing or delayed account information can impair cash visibility, while inconsistently assigned cash flow positions can reduce the reliability of liquidity forecasts.
• Incorrect master data or payment data can disrupt automated payment processing.

A practical example illustrates this point: At an international group, incoming cash flows were categorized inconsistently across different country subsidiaries, resulting in identical transactions being allocated to different liquidity positions. Consequently, the rolling 13-week forecast repeatedly deviated from actual account movements. A reliable forecast was restored only after the categorization logic had been standardized and an authoritative source had been clearly defined for each data object. No new technology was required.

ISO 20022 also highlights the importance of consistent and structured data. The international messaging standard supports richer and more highly structured payment information. When applied consistently, it can improve interoperability and facilitate more efficient payment processes. The ongoing evolution of payment formats demonstrates that technical standardization can deliver its intended benefits only if the underlying master and transaction data is maintained completely and consistently.²

Reliable processes therefore require standardized data structures, clearly defined data flows and a high level of data quality. Without these prerequisites, existing weaknesses are often merely digitalized rather than resolved for good. This is why data quality is not a peripheral technical concern, but a key enabler of successful transformation.

In practice, data quality is not created within a single department. It results from the interaction of three functions: IT, Accounting and Treasury. Only when technical data provision, accounting principles and business management logic are aligned the organization can produce results that are substantively reliable and broadly accepted. Without this alignment, conflicting views of the same underlying facts can easily arise, resulting in management metrics that decision-makers do not trust.

Treasury should take the lead in this process. As the business owner of the liquidity and management perspective, Treasury defines the functional requirements, coordinates alignment across the relevant departments and is responsible for interpreting the management metrics. IT ensures the technical integration and processing of data, while Accounting ensures consistency with financial reporting requirements. However, Treasury retains ownership of the management perspective and is responsible for bringing the different viewpoints together.

This becomes particularly apparent in cash and liquidity management projects. The primary challenge lies less in the technology itself than in aligning business and technical requirements across several steps:

• Definition of data sources and technical integration: Treasury and IT must jointly determine which source systems are authoritative and how the relevant data will be accessed. This includes interfaces, update logic and access rights.
• Mapping of liquidity positions: positions originating from the source systems must be mapped to both Treasury’s functional definitions and the corporate and accounting perspectives. This ensures that the same position is understood consistently across functions.
• Definition and mapping of management KPIs: the management KPIs used by Treasury must first be defined from a functional perspective and then mapped consistently to the company’s balance sheet. This establishes an end-to-end link between operational management and the consolidated financial statements.
• Technology-enabled reporting: only aligned, technology-enabled reporting can ensure that validated management metrics are presented to decision-makers in a transparent, traceable and decision-useful format. The key component is the data layer, which harmonizes, transforms and stores historical master and transaction data for reporting purposes.

The value of this approach lies in the validated outcome. When Treasury takes the lead in coordinating data sources, position mapping and KPI definitions with IT and Accounting, the resulting management metrics are functionally accurate, compatible with financial reporting requirements and technically reproducible. Data quality is therefore not solely an IT matter. It is the result of aligned functional and technical responsibilities under Treasury’s leadership. Only this interaction makes management metrics verifiable and trustworthy for management.

The more heterogeneous the system architecture and the more widely data is distributed across different applications, the greater the requirements for integration, harmonization and governance. Companies with consolidated platforms, clearly defined interfaces and standardized data models can implement new technologies more quickly, scale analytical use cases more efficiently and respond to change with greater flexibility.

In addition to technical integration, functional data governance is becoming increasingly important. Key considerations include which source is authoritative for a particular data object, how its quality is monitored and who is responsible for its definition and maintenance. ERP and Treasury transformations provide an opportunity to harmonize historically accumulated data and establish standardized definitions at an early stage. The quality of the data architecture therefore increasingly determines how quickly innovations can be put to productive use and how much value new technologies ultimately deliver.

For Treasury functions, this means systematically embedding data quality in processes and responsibilities. A clear data ownership model for key data objects provides a useful starting point. While the functional data owner defines the relevant standards and quality requirements, data stewardship can support the operational maintenance and monitoring of data quality. An authoritative source, or “golden source,” should also be defined for material information to reduce conflicting data sets across systems.

Data quality KPIs covering areas such as completeness, timeliness, consistency and error rates provide transparency into the quality of the existing data foundation. Wherever possible, validation mechanisms should be applied at the point of data capture so that errors do not first become apparent in reporting or downstream processes. Transparent data lineage also provides visibility into the origin and processing of information relevant to management and makes it easier to trace. The importance of accurate, complete and timely data, as well as clearly defined governance structures, is also emphasized in the regulatory principles governing risk data aggregation and risk reporting.³

A simple maturity model can be used to assess where a company stands on this journey:

This transforms data quality from a reactive remediation task into a continuously managed component of the Treasury organization.

As analytics, automation and AI become increasingly important, data quality and governance are moving to the center of strategic decision-making. This requires a combination of a, clear standards, b, defined responsibilities and c, established quality assurance mechanisms. The objective is to provide transparency regarding the origin, use and quality of data and to ensure that it is interpreted consistently across functions.

This foundation becomes even more important when AI is used. AI-supported analyses based on incomplete, inconsistent or outdated data can produce results that appear plausible at first glance but have limited reliability. Data quality is therefore an essential prerequisite for transparent and trustworthy AI applications. The Financial Stability Institute identifies data quality, along with data privacy and data security, as one of the key challenges to the wider adoption of advanced AI systems in the financial sector.¹

Regulatory developments also underscore the importance of data governance. For certain high-risk AI systems, the European Union’s AI Act establishes specific requirements governing the quality and management of data sets used for training, validation and testing. Regardless of the specific regulatory classification of a Treasury use case, the ability to document data provenance, quality assurance mechanisms and responsibilities in a transparent and traceable manner is becoming increasingly important.⁴

At the same time, governance structures must be designed to enable innovation rather than impede it. Successful companies therefore view data quality not as a one-time remediation project but as an ongoing management responsibility.

The future of Treasury will increasingly be shaped by data-driven decision-making, intelligent automation and the use of AI. However, the long-term success of these developments will depend less on any single technology than on the quality of the underlying data foundation.

For companies, this means addressing data quality as a strategic priority from the outset. An integrated data architecture, consistent data models, clearly defined data sources and unambiguous responsibilities, under Treasury’s functional leadership and in close collaboration with IT and Accounting, create the conditions needed to implement innovation efficiently and scale it sustainably.

Data quality is thus evolving from an operational concern into a key value driver of Treasury transformation and an essential prerequisite for realizing the long-term potential of modern technologies.

_____________________________________________________________________________________________________________

¹ Crisanto, J. C.; Currat, A.; Ehrentraud, J.; Wu, W. (2026): In data we trust? Emerging policy and supervisory approaches to AI data use in financial services, FSI Insights No. 73, Financial Stability Institute / Bank for International Settlements, 26.03.2026. https://www.bis.org/fsi/publ/insights73.htm
² Committee on Payments and Market Infrastructures (CPMI) (2026): Harmonized ISO 20022 data requirements for enhancing cross-border payments – updated report, Bank for International Settlements, 26.02.2026. https://www.bis.org/cpmi/publ/d230.htm
³ Basel Committee on Banking Supervision (2026): Implementation of the Principles for effective risk data aggregation and risk reporting (BCBS 239 Principles), Bank for International Settlements, 06.01.2026. https://www.bis.org/publ/bcbs_nl36.htm
⁴ European Union (2024): Regulation (EU) 2024/1689 Laying Down Harmonized Rules on Artificial Intelligence (AI Act), in particular Article 10 “Data and data governance“. https://eur-lex.europa.eu/legal-content/DE/TXT/?uri=CELEX:32024R1689