Inject data science into KYC for improved client insights
Data science is changing the customer landscape in financial services. By utilising rich data sets, it is now possible to authenticate individuals faster than ever before, and in doing so firms have greatly improved insight into the risk each customer represents. Steve Elliot, Managing Director at LexisNexis Risk Solutions, below relates the opportunities that can come from improved data-driven working.
Who are your customers?
While name, address and date of birth may help provide a few markers as to the legal identity of your customer, they establish little more than that. Even expanding these data attributes to include nationality, gender or passport number doesn’t build a clear picture of who they really are.
A person’s identity is built up of a variety of aspects, including: devices they use; what they do, when and why they do it; who they engage with; their biology; their dreams, aspirations, wants and needs; their strengths and vulnerabilities; their finances and risk appetite.
Humans are continuously changing individuals, living as part of an array of interconnected networks, each of which are also undergoing continuous change. It is therefore surprising that some firms are still failing to look any further than the basic legal data attributes when building online identities to fulfil their ‘Know Your Customer’ (KYC) efforts. If humans aren’t static, how can their identities be?
Reducing risk and the cost of compliance
Data is critical to achieving improved KYC insight, and the more data firms can access and evaluate, the more businesses can know about their customer base. Still to this day, many firms continue to experience unnecessary costs and errors when seeking to screen external data against their internal customer files due to both records containing limited data. The use of these limited data attributes increases the likelihood of partial matches, also known as false positives, which require costly investigations and human involvement to determine whether to accept or reject the risk when the full picture has not been established.
Fortunately, the solution is simple and readily available; enrich the customer record using larger, external data sets in addition to matching it against consolidated internal records . This will ensure that a higher number of matching attributes exist in both sets of records, significantly improving the accuracy. By doing this, it will also allow firms to quickly identify duplicate accounts in their own records, trace lost customers and update records to ensure the current customer view is correct.
Still, the cost and complexity of KYC compliance remains high at some businesses because they have not yet embraced the benefits of data science – i.e data plus technology. This unwillingness can be traced back to many causes, including a misplaced nervousness about whether new solutions are compliant, confusion over the number of potential solutions that have emerged and fear that transferring to new solutions may take the organisation away from existing, ‘proven’ solutions.
Turning data into insights
Although larger, more complex data sets have the potential to provide greatly improved KYC and risk insights, they also require new tools to perform the complex analysis needed to achieve those outcomes.
A key tool involves artificial intelligence (AI), which in its broad sense, is very simple; it is the use of computers to perform tasks much quicker than humans could. It is often used to analyse large sets of data, finding patterns and connections that achieve an outcome without the need for human intervention. More sophisticated deep learning (DL) methods involving AI, use both structured and unstructured data to seek out ongoing patterns, which eventually lead to a target output. This complex, multi-layered analysis is solving significant challenges in medicine, aeronautical design and many other sectors.
“Knowing your customer is nowadays key for businesses survival, and data is critical to achieving improved customer and risk insight.”
– Steve Elliot, LexisNexis Risk Solutions
Despite our trust in this more complex form to find life-saving cures in medicine, it is still not used to drive better outcomes for financial service customers. This is largely due to AI being unable to explain exactly how it arrived at a particular outcome, which is key in financial services – being able to show you reached a decision, and the reasons behind it is a regulatory requirement. A challenge for regulators, therefore, will be in accepting that better customer outcomes can be achieved through the utilisation of ever larger and seemingly disconnected data sets.
Regardless, of the ‘explainability’ challenge for deep learning solutions, there are still other, simpler, forms of data driven statistical modelling and machine learning solutions that can be deployed now, and are fully ‘compliant’ with regulations. It is these that should be used, until more sophisticated deep learning methods can be married with regulation.
Reaping the rewards
It is clear that richer data sets, combined with effective technology solutions, improve customer experience, creating considerably less disruption for the customer and onboarding them far quicker and without the need for extensive questioning. Risk insights will also be improved through the effective evaluation of greater amounts of data and the cost of data errors, duplicates and inefficient decision making are significantly reduced when utilising these progressive but fully compliant data science solutions.
In an increasingly data driven era, it is vital for businesses to understand the value of data and all the potential possibilities from using it correctly. By implementing the right technology to facilitate the vast amounts of data needed to inform decisions, businesses can truly reach their full potential and reap the rewards of their initial investment. Knowing your customer is no longer a ‘suggestion’, but is actually key for businesses survival.
