Nearly every industry is monitored by some form of a regulatory body, each with their own seemingly never-ending list of regulations that must be accounted for and adhered to. Of course, these requirements are designed to protect the integrity of our economy and society but speak to a compliance officer at any major financial institution and you’ll quickly discover that compliance has never been an easy task, even considered a burden. Add to the equation a rapidly evolving playing field driven by technological evolution and you’ll find governments scrambling to keep up by passing new regulations and intensifying old ones. This makes an already complicated compliance environment all the more difficult for regulators and corporate entities alike.
Consider anti-money laundering compliance requirements in the European Union. When the European Union’s 4th anti-money laundering directive (AMLD4) came into force, regulated companies, such as financial institutions, insurers, and real-estate firms, have needed to record and maintain up-to-date ultimate ownership records on their clients with documentary evidence and provide these on request to regulators. EU member states were also required to establish UBO registers. Earlier this year, AMLD5 went into effect, extending AML-laws to digital exchanges and art dealers. Understandably, identifying the ultimate beneficial owner of a legal entity is often regarded as one of the most painful parts of the regulatory compliance landscape, especially since every EU member state has “localised” the interpretation of what is considered a UBO based on the national commercial laws and historic definition.
For a financial institution, a new corporate client’s true ownership may be buried beneath many layers of complex internal structures and in most cases, across borders and additional legal entities. Rarely are beneficial owners easy to identify and deducing the ownership structures is a timely process that can take weeks, if not months, of manual research, with different jurisdictions around the world offering the necessary data in a variety of formats and at different levels of accessibility. This presents a challenge for many institutions that can feel insurmountable to overcome with the capabilities of their current infrastructure and are often growing their backlog of client cases that need to be processed. But, the EU has made it clear: the verification and documentation must be done and the executives, not just the organizations, will be held liable if they fail to comply.
Of course, compliance is easier said than done. This is why many regulated sectors such as financial institutions, art dealers and digital (crypto) platforms, are turning to regulatory technology to transform their internal operations from traditional paper trails to a fully digital process. The ultimate goal is to turn the burden of compliance into a competitive advantage through the utilisation of regulatory technology.
What is regulatory technology?
Regulatory technology, or regtech for short, is the application of emerging technology to automate regulatory compliance
Over the last several years, regtechs have seen a huge growth in demand for their solutions thanks to their proactive approach to improving efficiency. In fact, the regtech sector is estimated to grow to more than $50 billion in the next five years. The key to innovating and automating compliance is the combination of robotic process automation (RPA), artificial intelligence, and cloud services, which some regtechs have already combined to specifically transform ultimate beneficial ownership discovery from a manual process to an automated process for some global financial firms.
Let’s break down the process.
Why does ultimate beneficial ownership discovery need regtech?
As explained earlier, in the world of AML compliance, achieving full insight into ultimate beneficial ownership (UBO) structures ranks as the most difficult challenge of all. Shareholder structures can be opaque, comprising vague networks across many jurisdictions.
To make it even more challenging, shareholder information tends to consist of unstructured data, often hidden in paper filings, with no consistency for how it’s collected, recorded or stored. Compliance teams are tasked with wading through PDFs , often compromised of inconsistently designed forms and even low-quality scans of hand-written documents! Even within the same country, different regional bodies can log their data in a variety of ways.
Therefore, regtechs quickly surmised that a significant technological innovation would be required to simplify this burdensome task for compliance teams.
Using AI to simplify UBO search and discovery
As a regtech, it’s essential to already have a strong foundation of clean data before adding AI to the equation. Austria-based kompany, for example, operates a proprietary real-time register network to access original company filings of over 110 million entities in more than 200 countries and jurisdictions. As they are delivered directly from the primary or government source, the data is considered a true copy. This makes the data legally reliable and certain, which is of critical importance for entities required to adhere to UBO related regulations.
The next step is to process the data through the artificial intelligence (AI) engine , which kompany is doing by applying the same rigor to unstructured shareholder documents as it did establishing its API connection to global registers . In their case, they began by tackling the jurisdiction with the most challenging shareholder data first: Germany.
Despite being home to Europe’s biggest economy, Germany does not have a centralized governance system for its company filings. Instead, over 100 regional authorities manage their own local registers. Considering the country’s commercial backbone, the SMEs, comprises around 1.6 million companies, this is a data nightmare for compliance departments, and as a result, UBO checks can take weeks to complete. The German UBO register, Tranparenzregister, has a very low filing rate of less than 10% and is not considered a viable source to conduct the required UBO checks. This is also the reason the German regulator has issued a circular recently allowing shareholder information to be primarily searched in the commercial registers.
In most cases, this challenge would have been dealt with via the use of manual review, but because of the huge complexity of this unstructured data, it has become essential to use OCR and machine learning to resolve the backlog and achieve a high level of confidence through automation.
Starting with 1.5 million data sets and over 70,000 different models that were tested to read and interpret the German shareholder filing, these were then whittled down to approximately 600 which were subsequently put into AI training. By narrowing the models down even further, it is now possible to analyze the unstructured data and make it machine readable by putting it into a structured format, achieving 95% confidence, comparable to human comprehension at a fraction of the time.
Based on a client’s request, unstructured data such as a German shareholder filing in pdf as an image layer, is pulled in real-time from the primary source or government register. After applying optical character recognition (OCR) software and turned into machine readable text, it is then run through an in-house tagging and machine learning based analysis engine. The final result is machine readable and structured data set with cross-border shareholder information that can be accessed through the API or fed into an advanced visualisation tool that displays connections between companies, directors and officers and shareholders, with consistency and integrity at its core.
An estimated 80% (though kompany estimates this number to be closer to 95%) of all company information in registers around the world is only available as unstructured data. This is where the kind of innovative solutions regtechs are developing become essential and where artificial intelligence proves itself to add a transformative difference to the way compliance teams have been forced to run their operations up to today.
