The Promise of Global Digital Identities: Facilitating Ecommerce In A World Where Everyone Is A Merchant
A few weeks ago, I participated in a panel session at the K(NO)W Identity Conference in Washington D.C. to discuss the significant and growing Know Your Customer (“KYC”) challenges in the payments space, as well as ways in which technology can help with effective and efficient risk assessment and compliance. During this conversation, I highlighted how the increased automation of fraud presents challenges for WePay’s current KYC program, and how digital identity systems, as proposed by the World Economic Forum and other innovation and technology leaders in the financial services industry, could address this pain point for WePay and many others in the industry.
What Are Our Challenges?
Making our product as seamless and invisible as possible while meeting the regulatory expectations of KYC
Everyone in payments today adheres to prescribed standards of KYC and identifying their customers, as required by the Bank Secrecy Act (“BSA”). WePay’s challenge is to adhere to this prescribed KYC model in a manner that meets regulatory and partner expectations while also remaining a creative and disruptive FinTech company.
Our approach to this challenge has been to adopt a multi-step on-boarding and underwriting process. This means we allow our merchants to accept payments immediately with only a name and an email address and then, later in the payment lifecycle and before settlement, we require completion of KYC and underwrite the merchant. This model allows for flexibility that both our platform partners and our merchants find valuable. You can imagine how much higher the conversion rates for our partners are since we make it so easy to accept payments quickly. Similarly, you can imagine how powerful being able to accept payments immediately as a small business can be in growing the customer base.
While we have successfully implemented a model to provide seamless payments and meet our KYC obligations, digital identity systems can serve to eliminate the bottleneck that failed KYC verification creates. Here at WePay, approximately 15% of users fail automatic identity verification and may be subject to manual review and authentication, which is quite time consuming. As a further challenge, current standards for identity verification are based on physical records, i.e. the ability to prove identity depends on access to and authentication of physical documents, like passports and government-issued ID cards. This creates tons of friction in the digital economy in which we play, because 1) physical identity documents can be falsified and altered as well as lost or stolen, and 2) the transfer and review of physical identification documents is slow and creates the potential for human error.
On the other hand, a digital identity system would exist as a set of digital records that the user can control and use to complete transactions. In this way, proof of identity would be communicated between an individual and financial institutions in a standardized, digital format. Proposed identity systems would also adapt to the nature of the transaction, and continuously adapt to requirements by integrating additional information to create a rich view of the user.
This type of system would allow WePay as the payments partner to interpose itself in the merchant/platform relationship as little as possible, thereby enhancing the user experience, and have a better understanding of the user, thereby more efficiently monitoring for fraud and other suspicious activity.
Fighting the increased automation of fraud
Machine learning and artificial intelligence are all the rage and, to a certain extent, how we techies have differentiated ourselves from traditional financial institutions. WePay employs machine learning models trained to distinguish between fraudulent users and good ones. The model uses a random forest algorithm and the model score is available as a variable for writing rules. The rules refer any payment or withdrawal with a high enough risk score to human analysts for review. Read more information about WePay’s machine learning approach here.
Despite the great promise of machine learning, it is still considered a black box. Financial services has been at the forefront of machine learning advancement, but not everyone is deploying machine learning because the industry has not been able to substantially overcame the hurdle of financial regulation. That’s because it might prove quite challenging and inadequate to tell a regulator that a computer made a risk decision and show them a piece of math as proof.
Since we’re not regulated on the federal or state level, WePay is currently fortunate to not have to be too concerned about scrutiny associated with machine learning models. That might not always be the case for FinTechs generally, and payment facilitators specifically, as the conversation regarding regulation expands beyond traditional financial institutions.
In addition to machine learning technology, traditional information sharing (e.g. with other financial services companies and law enforcement) also remains an important factor in effectively monitoring transactions, and therefore minimizing both fraud and compliance risk. That’s because transaction monitoring software is only as good as the data we feed it and sharing information allows for greater insight into a merchant’s activity across the ecosystem.
To succeed in this payments environment, payment facilitators like WePay will need to successfully address the challenges around risk, speed, and competition. I feel confident that WePay is building a product and risk solution that adds maximum and unparalleled value for small businesses and our platform partners. Nevertheless, challenges around KYC are a critical pain point that inevitably slows down our operations.
This is just one problem that the adoption of digital identity systems would address. Such systems enable proof of identity for transactions that take place in the digital economy, unlike physical identity systems, which were designed to support face‐to‐face transactions. Digital identity systems will be crucial to continuing innovation and delivering efficient, secure, digital‐based Fintech products. These are systems that will make the question of competition more about the technology that most efficiently improves the user experience and minimizes fraud and compliance risk.