When it comes to navigating the ever-growing universe of regulatory mandates for exchange traded funds, it may be helpful to think of "ETF" as standing for Expedited, Transparent and Frugal. But how to achieve all three? Pattern recognition and AI, with a sprinkle of data science, are the answer the industry is looking for.

Jasmeet Narang
Quality EDGAR Solutions
Data Scientist

Expedited. Asset managers need their reporting output fast. In a highly competitive market, speed to launch is imperative, and often reporting and SEC filing are crucial components to expedited launch.

Transparent. ETF operations involve a complex web of service providers, and no single party typically holds a complete reporting data set. A firm may have fund accountants and administrators, custodians, index providers, market makers and investment managers all providing their own siloed data to produce ETF reporting. This all gets aggregated together, but to have any confidence in your numbers and analytics there must be transparency in data lineage that flow to the final output.

Frugal. While ETFs are launching at a blistering pace, many never reach long term viability and can quickly turn into resource drains. Maximizing value across process and vendors ensures that potential profits outweigh risks of operational cost when considering a launch strategy.

Achieving this version of "ETF" is not a simple task, but emerging technologies can simplify the process for ETF providers of all sizes. ETFs, in fact, are ripe to benefit from pattern recognition-driven AI solutions when it comes to regulatory reporting and data management. Today's asset managers must analyze and interpret massive volumes of financial and operational data from sources such as fund accounting platforms, third- party administrators, and market data providers like ICE and Bloomberg. The challenge isn't just about processing data, it's about transforming raw, inconsistent information into structured outputs that is mandated by SEC. That can be done manually, as it still is in many corners of the industry, but it doesn't need to be this way, and new approaches can free up significant manpower in back- and middle-office teams that can help introduce a range of efficiencies in other parts of the business.

To reduce the manual process that still dominates much of the industry, it's key to look at what data is needed and where it needs to appear in reporting documents. Just because a firm may have handled things one way for years doesn't mean it needs to be that way forever. This process involves identifying key data from large datasets, performing necessary calculations, and converting information into the correct format. From there, new approaches allow for the automating of the entire workflow from retrieving values and understanding data flows to interpreting document structures and locating exact insertion points. Once mapped, scripts and algorithms can be developed to extract and populate these values automatically on a quarterly or annual basis. These outputs are then integrated into an internal system that assembles complete reports, such as TSRs.

To make the process even more efficient, there are now means through which asset managers can leverage machine learning, via a custom AI pattern recognition system that can identify patterns, trends, and structures within disparate data. The process requires training a model on a substantial set of labeled data with predefined rules and metadata that facilitate the identification of hidden patterns and the extraction of essential data points from large datasets. This capability allows for the efficient mapping of financial data, enabling easier generation of relevant information for various SEC-mandated reports.

Such a model can then be applied across different reporting frameworks and tailored to meet the needs of diverse fund series, strategies, and asset types. The key is knowing the right questions to ask before embarking on implementing these technological tools since they are, after all, tools and require a deft hand to make them work to their highest functionality.

There is a clear advantage for firms that are embracing these new approaches. They can respond faster to the new mandates, reduce time spent on the manual tasks, and deliver higher quality outputs. Forward-looking firms are already seeing results, in some cases cutting their reporting cycles in half, reducing human error, and improving collaboration across teams.

In contrast, those who cling to their legacy systems and manual workarounds risk falling behind. They may be meeting today's regulatory requirements but will struggle to keep up as data volumes grow and compliance become even more demanding. The longer the wait to take advantage of what AI is making possible, the higher the hill they will have to climb.

Interestingly, smaller asset management firms are demonstrating greater agility compared to large institutions that focus on building enterprise-scale AI infrastructure. Nimble asset managers are partnering with specialized vendors to implement targeted tools that automate critical tasks. These firms are not just catching up, they are leapfrogging outdated systems and setting new benchmarks for speed and accuracy in compliance reporting.

Having said this, it is important to recognize that AI is a powerful resource, but it cannot and should not be put to work in a vacuum. Rather, it should be used as a tool to simplify tasks and enhance user efficiency. AI, after all, comes with its own challenges, beginning with lack of transparency in so-called "black box" models. This opacity is a concern for regulators as it undermines market integrity, explainability, data quality, or model validation, ultimately introducing operational and systematic risk. Such limitations can make it difficult for fund managers to align model outcomes with compliance requirements. Moreover, the lack of explainability can contribute to "hallucination," where the system generates inaccurate or misleading information. This increases the risks of biased outputs or false positives. These issues can lead to erroneous financial calculations and negatively impact decision-making.

Nevertheless, while AI and other emerging technologies may not be the industry's panacea when it comes to the future of regulatory reporting, it is having, and will continue to have, a profound impact on the space. Firms that are embracing new tools now are poised to see competitive advantages for years to come.

Jasmeet Narang is a data scientist with Quality EDGAR Solutions (QES), where she leads the design and execution of data pipelines. Her work involves collecting raw financial data and transforming it into a structured format for regulatory reports mandated by the U.S. Securities and Exchange Commission (SEC).

Stay ahead of the news ... Sign up for our email alerts now
CLICK HERE