Many of us remember the habitual panic before going on a foreign holiday in the pre-digital age. Some of the last minute worries remain the same today — little niggles like “what type of clothes should I take?”, or “have I exceeded the baggage allowance?”. Others, thankfully, have been relegated to a bygone era — worries such as “where can I get travellers cheques in drachma?”.

Nowadays, innovative fintech companies are using mobile technologies to make financial transactions far faster and more efficient. For example, transferring funds from one currency to another can be achieved with a few swipes in a smartphone app. Customers can get a quick overview of their finances in seconds, while applications for loans, insurance quotes and many other services can be carried out faster than ever before.

The benefits of digital disruption do come with a price, however. One facet of this change is the decreasing number of banks and other financial institutions which have a physical presence. In the UK, about 60 bank branches close per month and a similar trend can be observed elsewhere in the developed world.

This has some fundamental implications for how financial institutions interact with their clients and acquire information. One of the fundamental roles of a conventional bank is to reduce “information asymmetries” between investors and borrowers. In this regard, banks face a challenge because they are often working with incomplete data.

The classic way to tackle this problem was by adopting a “relationship banking” approach. Firstly, bank staff would aim to get to know the client through face-to-face meetings. In addition, a pricing model would be developed that strongly encouraged the client to use the institution as a “one-stop shop” for all their financial needs. By building an individual relationship and analysing data across all of the client’s financial interactions, the bank could build up an accurate picture of the risk profile of each customer.

However, most fintech firms adopt an approach that looks nothing like relationship banking. Lean and agile, these firms typically aim to do one thing well. So if they have no physical presence and a narrow, focused product offering, how can they make good, data-driven decisions? This could be where artificial intelligence (AI) makes its most important contribution to the fintech industry.

Newspaper articles about AI are typically accompanied by a stock photo of a humanoid robot, but this sci-fi representation of the technology is quite misleading. While robots certainly have their uses in the industrial sphere, much of the technological advancement in AI in recent years has been in the realm of machine learning rather than robotics.

Machine learning algorithms examine vast quantities of data to come to conclusions about patterns or phenomena, without being given explicit instructions. For example, if you open the Google Photos app and search for a term like “cats”, it will analyse the images on your phone and return a custom selection which it thinks contains cats. In order to create this feature, a machine learning algorithm analysed a large quantity of photos that had been manually categorised as cats by humans. This is known as the training data. Afterwards, it tries to apply the lessons from examining the training data to the photos on your phone in order to categorise them.

So how does this apply to fintech? Well imagine that a machine learning algorithm examined the anonymised financial records of many clients who took out a mortgage. By analysing the data and looking for patterns, such algorithms could identify the “data fingerprint” of a customer likely to take out a mortgage in the near future. The customer could then be offered proactive advice on the mortgage market or tips on saving the funds for a deposit. Similarly, data could be examined to identify the warning signs of a customer who poses a default risk to help determine whether to approve a loan.

While this is an area involving important regulatory, privacy and ethical questions, AI holds the promise to greatly speed up decision-making in the financial sector and make digital financial assistants more useful. A similar approach could be adopted to approving small loans, setting insurance premia, or even offering proactive advice on pension or investment decisions.

In fact, there might even be hope for a form of relationship banking after all. A number of fintech firms have combined advancements in artificial intelligence (AI) and natural language processing to produce sophisticated chatbots which promise to usher in a new model of “conversational” digital banking. Such digital assistants could advise customers on investment decisions, provide market-related news, and field customer queries. Of course many people still find it very hard to imagine talking to a computer about their finances, but at the rate these technologies are developing, it would be unwise to bet against them.

Unfortunately, fintech innovators working on new AI services such as these often struggle to secure basic corporate banking services due to the outdated risk-modeling processes of legacy banks. Without a corporate bank account, many of these new financial services may never see the light of day. INITIUM aims to provide innovative firms in the new digital economy with a dependable banking partner. By marrying the best of relationship banking with careful, selective use of automation and machine learning, INITIUM aims to become the bank of choice for the new digital economy. Learn more and visit