I recently gave a TEDx speech in Portugal on the topics of FinTech and how AI, ML, and society are changing the financial world. I only had a short 10 minutes to speak, so here I will flesh out the topics in more detail since I have fewer restrictions on my time.

 

The Problem with Traditional Finance and its Institutions

The banks and financial institutions are often viewed negatively by the public. In fact, the lowest trust in banks was recorded in the EU, but globally banks tend to be viewed with suspicion and dislike. Banks can provide many useful services, and in the past, they maintained networks that guaranteed assets could be accessed at various places, eliminating the need for travellers to protect themselves from armed bandits on the highway.

Banks also did (and still do) connect those with excess capital and those in need of capital through loans. They allowed the un-landed to still found businesses because the founders could borrow money and trust banks as intermediaries to assist in leasing property. Insurance companies gave people peace of mind in return for periodic payments.

As for international finance, the banks and other financial institutions assist in cross-border trading, insurance, and financing. For businesses that need to regularly pay foreign suppliers, it is far more useful to employ a bank for that service than to hold cash and adjust reserves internally. The financing efficiency of today’s global corporations is enabled by overnight loans, often disbursed for sums of millions of US dollars, to cover one day of operations until more cash is received – this means capital is fully employed at all times.

Image Credit: Telegraph.co.uk

In the 20th century and especially in the 21st century, it seems the traditional roles of finance had been shoved aside in favour of simply making money. Banks still fulfil the traditional roles, but they also house prop trading desks that participate in markets precisely for making the bank money – not for the customers, but for the bank itself (and consequently the shareholders).

Banks and their reckless self-interested behaviour have been blamed for the financial crisis. Accounting firms that were meant to protect the investing public have failed in their audits (see Arthur Andersen). Insurance companies still provide peace of mind, but in some systems, particularly the United States health care system, the insurance companies have taken on a villainous public image, denying coverage of expensive drugs or due to notions of “pre-existing conditions”. The public may no longer maintain peace of mind when they use health insurance companies in the United States.

With a more mobile global citizenry, many people dislike the fees associated with transferring money across borders, especially when it seems so much of the financial world is now automated. If banks are not paying salaries for workers to transfer the money, why should customers be paying so much for those transfers? Clearly the banks are not putting those fees back into the pockets of employees.

Banks are in the business of money, and they are particularly adept at making more of it. The sole purpose of traditional corporations is to maximize profit, ideally by filling a market need, but often the ends justify the means – even when those means are immoral or detrimental to the host society (these are the economic externalities such as pollution or social consequences of wage policies).

This is where FinTech takes over. The businesses in Silicon Valley, like Google, Amazon, and Facebook, were founded partially on the drive to change the world. Of course, traditional corporations can also be based on positive values and morals, but tech companies often seem to be admired for their adherence to social desires, while traditional firms, but especially financial ones, tend to be reviled for their seeming dismissiveness of the public’s wishes.

FinTech wishes to follow in the footsteps of such tech giants by changing the financial world.

 

How FinTech Measures Success

The conventional measure of corporate success was the margin and the profit line. Greater margins meant greater success, as did higher profits. Other indicators of success are largely financial, and the stock market tends to react accordingly: higher earns translate to a higher stock price, perhaps due to the efficient market hypothesis, but also possibly due to the influence of herd psychology in finance.

FinTech, taking its playbook from Silicon Valley instead of Wall Street, wants to align itself with the people. The measures of success in FinTech are expedience and satisfaction to the customer, increased use of automation, and system-to-system communication.

 

The First Measure: Expedience and Satisfaction

To many people, convenience is king. And to be more convenient, expedience is required. With the successful implementation of convenience, satisfaction rises. FinTech companies may provide immediate access to funds anywhere in the world at any time and make transferring between two people as simple as touching two NFC-enabled smartphones together. A wire transfer may take days to complete – not to mention the exorbitant fees, especially when transferring between countries and currencies. The instantaneousness of electronic money transfers are an essential ingredient in FinTech’s attempt at surpassing the banks in the financial industry.

Some aspects of traditional finance have been automated and thus made into easily-accessible electronic processes, such as EFT and direct debit for bills. FinTech companies promise to streamline these processes and make them much more commonplace than they currently are.

 

The Second Measure: Automation

Automation is integral to FinTech – its name is financial technology, after all. There is a push to include automation in every mundane and repetitive task that requires little thought. This will lower the cost of processing for customers and relieve employees of boring tasks. No longer will so many humans be required to process, confirm, and audit transactions, dramatically reducing salary overhead.

Another benefit for customers is expediency and accuracy. In pre-automation and pre-connectedness days, a written check had to be physically delivered to the bank and processed by a human, the two banks had to contact each other to confirm the funds were available for remittance, and someone had to enter the new transaction amount into the bank system.

With connected systems, banks can communicate more effectively. More importantly, consumers now have a digital interface into which they can enter an amount to remit, that information is processed by machines, and the amounts are immediately deducted or credited. The central bank must still shift the money between banks for the finalization of the transaction, but these inter-bank transfers are done in large lump sums. Consumers no longer need to wait so long for money to be transferred between banks. For the misanthropes and distrustful, there is the added security benefit of impartial machines processing everything. This also increases accuracy, as computers are very good at copying data from one place to another while humans tend to make mistakes when copying data (especially if the task is tedious).

 

The Third Measure – System to System Communication

The third measure of success for FinTech is system-to-system communication independent of human interaction. This includes the above scenario of bank systems interacting with other bank systems. It also covers internal systems, designed to process different information, talking to each other to share data and create synergistic value.

For public-facing systems, we often use APIs to accomplish this communication. Open APIs provide the means for systems and even members of the public to create their own systems that automatically receive non-original data, free of charge. CityFALCON offers one such API.

 

How this evolution is different from the past

The current revolution in finance and FinTech is different from before. The pace of innovation is faster today, and globalisation allows the best people on the planet for any single job to collaborate, regardless of their citizenship or residence. The world is also becoming more and more educated, meaning more minds are working on these problems. These two factors (more education and more communication) mean we can solve bigger problems faster than before.

Some call the current integration of sensors and collection of information the Fourth Industrial Revolution. We can produce cheap sensors to capture myriad datapoints then store all of that on commodity hardware (i.e., cheap hard drives, either our own or those operated by someone else, such as through AWS or Azure). We can also process all of that information, often in parallel, leading to real insights in real time.

To accomplish this kind of insight, we cannot manually inspect the data and make manual analyses. We need to automate our analysis, which is exactly what machine learning (ML) has aimed to do. By feeding huge amounts of information into a well-designed system, we can use pattern recognition to discern patterns and their implications in new data. This kind of automation underpins another important benefit to consumers of FinTech and automation: we can find patterns of potentially fraudulent activity, preventing damage to customers. Such technology is already employed by credit card issuers to block credit card thieves from using the cards undetected.

 

Some Uses of AI

ML is a subset of AI, which is a broad field dealing with imbuing machines with intelligence, or the ability to reason about new concepts and scenarios.

One example of AI to improve customer satisfaction are chatbots. These can parse our questions, query a database, and return some meaningful set of information. CityFALCON has released an Amazon Echo-based system for this very purpose: the human asks the Echo to read off the most recent headlines, and the chatbot responds by reading today’s more important financial headlines. There is also system-to-system communication at work here, since Amazon processes the voice data, makes requests of CityFALCON, which returns the requested data to be processed into voice output by Amazon.

Another use of AI is pattern recognition as it is applied to biometrics and security. The modern bank will ask personal questions to authenticate a user. But these personal questions have discoverable answers (especially ones like “which street did you grow up on?” or “what was your first pet’s name?”). Conversely, our biometrics, like fingerprints or iris patterns, are much harder to discover and replicate. Humans are unlikely to be able to differentiate fingerprints or irises from each other quickly, if at all, but AI can.

Example of discoverability (and reproducibility): one scam that appeared on Facebook was designed to discover some secret banking question information. The scam relied on individuals’ interest in their “superhero name”, which was concocted by concatenating the name of the street of the victim’s childhood home and their first pet’s name. Crenshaw Fluffy or Oxford Rex. Sure, these names may sound cool or entertaining. But now the scammer has answers to two banking questions meant to secure financial accounts.

While it is true that fingerprints can be lifted from trash or public places, replicating a text string is far easier than replicating a fingerprint. Additionally, scanning irises is likely to be invasive and suspicious, and reproducing them is even harder than reproducing fingerprints.

 

One Potential Future: Blockchain and Cryptocurrencies

We’ve written extensively on blockchain and cryptos at CityFALCON. The technology would be fascinating, even if it were never widely adopted. The use cases are broad and largely coincide with use cases of distributed, redundant systems. Cryptocurrency as a currency has its obvious uses as a complement to fiats and a potential way to cut out banks and their fees.

Blockchain itself has more uses, especially when it comes to smart contracts. By formalizing transactions (legal or financial) into computer code and uploading them to a blockchain for automatic execution, we can ensure agreements are fulfilled and we don’t even need to interact with the system.

In my talk, I used the example of an insurance policy on a stolen phone. By uploading a legitimate police report, I could trigger the contract’s payout. The blockchain receives the police report, parses its content, determines whether the claim is eligible, then pays out the contracted amount. The last human-system interaction point is the completion of the police report; after that, everything is handled by machines on an automatic (and impartial) basis.

In Summary

FinTech is different from traditional finance in our founding principles, which are more aligned with the society-changing goals of Silicon Valley firms. We, in FinTech, also measure our success differently – profit is still important, but we focus on other areas as well if we wish to be considered successful. And FinTechs are just as diverse as traditional finance, being involved in insurance, banking, and transactions across many platforms.