Role of AI in virtual banking


The world is moving toward a new normal with the realization of data鈥檚 power, adoption of mobile-first mindset, and user experience-driven consumers. There is a profound effect of this change on the banking ecosystem, where digital-only, virtual banking is exhibiting a multitude of opportunities for bankers, regulators, consumers, developers, and fintech players. The biggest building block for virtual banking (digital-only banking) is data, and the most effective use of data is through applying Artificial Intelligence (AI) for actionable insights.

In conversation with Michael Berns, Author and AI & Fintech Leader to understand the 鈥楻ole of AI in Virtual Banking鈥.

Top 3 drivers for AI emergence in banking

Traditionally, for the United States, AI was solving financial crime challenges including, FX manipulation or alignment of manipulation affecting trillion-dollar instruments globally. Rather than driving innovation, the objective was to understand the language at scale and its nuanced way of solving complex problems. Today, the region boasts strong AI platforms for collateral trading, automating a lot of these processes. Big players from JP Morgan, Stanley, Goldman Sachs, are all investing billions each year in a deal.

Unlike the Americas, the European market and German-speaking regions are not quite at the same maturity of use cases here. Initially, it was quite clear that people are looking to use AI for cost savings for efficiency for speeding up things and refueling to change the client relationship or building an entirely new business. So, when we think back some 15 years ago, it was very difficult as a vendor to enter a bank and kind of get its foot in the door. Now it is a desire, a realization that you cannot solve all the problems that you have yourself, you need to look outside.

This sums up that there are three key drivers which led to the emergence of AI in banking:

  1. Partnerships
  2. Innovative solutions
  3. Agility

2 key blockers to AI adoption

  1. Data availability 鈥 Banks are still relying on legacy ways to make data available in some regions. Whereas, other parts of the world rely on partnerships to strategically gain a sense of missing data and purchasing it.
  2. Data quality 鈥 Despite the availability of data, the banks fail to maintain the quality of data for building accurate AI algorithms.

Instilling customer data and trust in virtual banking

Post the financial crisis of 2008, disrupting the banking systems, customers lost the trust in their bank. I might go to an alternative service, but they need to offer something that in terms of having a nice interface and easier switch on switch off, maybe more insight.

The transaction needs to be of some benefit. If you deliver the benefit, the customer is happy to share some more personal data. But that benefit needs to be clear to the customer upfront, and then the transition will happen.

It is true for let鈥檚 say data and transparency as well as how do I gain customers in that segment? It is different for varied types of segments and customers, some of them might be a more, maybe slightly younger segment, who are more interested in using mobile-only bank having different requirements.

From the AI perspective, it is a trust that the institutions will not use customers鈥 data against them. It is a trust and an unbiased decision making in the background and trust in something like a responsibility on the framework, pushing for having these checks for in our data model around bias as well as robustness and building on those ethically.

What use cases do you want to go for that create a benefit for the client and whether you want to go with this or not?

Top 2 AI use-cases for virtual banking

  1. Automation of a typical KYC onboarding process 鈥 What you have is an additional system for let鈥檚 say, cases, which is called the enhanced due diligence, and for that one, you try to pull it together as much as information as you can from unscriptural sources.
  2. Customised offerings 鈥 how can I offer something personal for the client, at least in the German reaches, that is not done at scale yet because of very strict data privacy rules. Back in the UK, in English institutions, they are trying to look at transactions of the customer, and depending on the transactions, what products can I offer them.


twimbit Insights

Where the world discovers and shapes research