Financial institutions still process payment instructions containing incomplete, inconsistent, or unstructured address information. While much of the industry conversation around structured address has been driven by ISO 20022 requirements, the operational impact extends well beyond compliance.
Poor-quality address data can contribute to payment repairs, manual investigations, screening complexity, and inconsistent information across systems. As payment operations become increasingly automated, the quality and structure of the underlying data become even more important.
This was a central theme of IntellectEU’s recent webinar, Beyond Structured Address Compliance: Turning a Regulatory Requirement into an Opportunity, with Xavier Heyvaert, Deputy Head of Payment Modernization & Swift, and Mark Swift, Product Owner for Catalyst Data Intelligence, moderated by Hryhorii Vasylenko, Product Manager for Catalyst Integration Manager and Catalyst Core.
“This was never really about compliance in the first place. Better structured address data means better payment quality.”— Xavier Heyvaert, Deputy Head of Payment Modernization & Swift, IntellectEU
Watch the full webinar recording
The timeline changed. The data challenge remains.
On August 27, 2026, Swift announced an extension to the structured address migration timeline for ISO 20022 payment messages. The original plan was to remove fully unstructured postal addresses as part of Standards Release 2026. Following feedback from banks, central banks, payment market infrastructures, and other industry participants, Swift decided to defer the payments changes and consult the community on the timing and approach for structured addresses.
Importantly, Swift continues to encourage financial institutions and payment market infrastructures to move forward with their structured address programs rather than wait for the revised deadline. That makes sense because the underlying data-quality problem remains unchanged.
Addresses are inherently difficult to standardize. The order of address components varies by country. Abbreviations, transliteration, spelling differences, missing information, and regional naming conventions all create inconsistencies in data that may still be understandable to a person but difficult for automated systems to process reliably. The same information may then flow through payment engines, sanctions-screening platforms, fraud controls, investigation workflows, and other downstream systems.
The objective should therefore not simply be to move existing information from an unstructured field into a different message format. It should be to improve the quality and structure of that information so it becomes more useful throughout the payment lifecycle.
Better data can reduce operational friction
The impact becomes more visible at scale. During the webinar, Xavier and Mark discussed how payment investigations can require several minutes of manual work. Even when only a relatively small proportion of transactions requires intervention, that can translate into a considerable operational workload across high payment volumes.
Screening presents a similar challenge. Address quality is not the only factor behind sanctions or fraud alerts, but clearer and more consistently structured party information gives screening systems better data to work with. The same applies to payment repair and exception-management processes. When important information is embedded in free text, downstream systems have to interpret it. When elements such as country, town, postal code, street name, and building number are clearly identified, that data can be processed, validated, and reused more consistently.
Structured address data can therefore support better straight-through processing, more efficient investigation and exception-management workflows, and stronger data inputs for screening and other automated processes. This is a broader business case than compliance alone.
Automation makes data quality more important
Financial institutions continue to invest in automation and AI, but the effectiveness of these technologies depends on the quality of the data they receive. Automation does not compensate for inconsistent underlying information. In many cases, it makes data quality more important.
ISO 20022 provides a richer and more structured data model for payments. Realizing its value, however, requires more than adopting the message standard. The information entering those fields also needs to be sufficiently accurate and consistent for the systems consuming it.
Address data illustrates this challenge particularly well. An incoming record may contain a misspelled city, an incomplete street address, a market-specific abbreviation, or several address components combined in one field. Simply separating the string into new fields does not necessarily create reliable structured data. The process needs to identify what each component represents, correct or enrich information where appropriate, and determine how confident the institution can be in the result.
From parsing addresses to improving payment data
During the webinar, Mark demonstrated how Catalyst Data Intelligence approaches this problem. Catalyst Data Intelligence uses AI to transform unstructured address information into structured or hybrid formats aligned with ISO 20022. It can identify address components, enrich incomplete information, and support more consistent data across payment workflows.
The demonstration included individual addresses as well as XML containing multiple address records, showing how address structuring can be incorporated into broader payment-data processes rather than treated as an isolated remediation exercise.
The solution uses country-specific models and can be deployed on-premise, allowing institutions to retain control over their data and models. It does not depend on a general-purpose large language model for address processing. This is particularly important in financial infrastructure, where the objective is not simply to generate an address that appears plausible. Institutions need repeatable processing, clear controls, and mechanisms for identifying records that require further review.
Explore Catalyst Data Intelligence
Knowing when not to automate
Not every address contains enough information to produce an equally reliable result. Some can be structured with a high level of confidence. Others contain missing, ambiguous, or contradictory information.
Confidence scoring provides a practical way to distinguish between the two. Institutions can define thresholds based on their own operating model. High-confidence records can move through automated workflows, while lower-confidence results can be routed for review or repair.
The aim is not necessarily to remove people from the process entirely. It is to reduce unnecessary manual intervention and focus operational teams on cases where human judgment is actually needed. This reflects a broader principle for the application of AI in financial infrastructure: automation should be paired with appropriate control.
What should institutions do with the additional time?
For institutions already working on structured address programs, the extended timeline creates an opportunity to reassess the business case rather than simply pause implementation. Compliance remains one driver, but it does not need to be the only one.
Institutions can look at where poor-quality address information is already creating friction. How frequently are payments being repaired or investigated? How effectively can screening platforms use existing party data? How much customer and counterparty information still sits in free-text fields? And where could cleaner data improve automation across the payment lifecycle?
Those questions can help determine whether structured address should remain a standalone standards project or become part of a broader payment modernization and data-quality strategy. The second approach may create significantly more long-term value.
Better address data is ultimately a payments question
The transition to ISO 20022 is giving financial institutions access to richer and more structured payment information. But richer message formats do not automatically create better data. The value comes from making that information accurate, consistent, and usable across the systems and processes that depend on it. Structured address is one practical example.
The immediate requirement may be driven by messaging standards and regulation. The longer-term opportunity is to give payment systems, screening platforms, operational teams, and increasingly AI-driven processes better information to work with. In that sense, structured address is not only about how an address is formatted. It is about building better data foundations for better payments.
Watch the webinar
Watch Beyond Structured Address Compliance: Turning a Regulatory Requirement into an Opportunity for the full discussion with Xavier Heyvaert, Mark Swift, and Hryhorii Vasylenko, including a live demonstration of Catalyst Data Intelligence.
Watch the recording
Explore Catalyst Data Intelligence
See how Catalyst Data Intelligence helps financial institutions structure and enrich address data for ISO 20022 payment environments.
Explore Catalyst Data Intelligence
Explore the Catalyst Product Suite
Catalyst brings together payment and system integration, financial messaging, managed connectivity, ready-to-deploy solutions, and AI-powered data intelligence in one enterprise suite.
Explore the Catalyst Product Suite
.png)
