Tomorrow’s Finance: How AI is learning the ropes

Written by Sam Moss , June 6, 2019

Welcome to the first edition of the Tomorrow’s Finance series. In this collection of blog posts I’ll be exploring some of the current tech trends in finance – what they are, why they’re needed, and which businesses are leveraging these technologies in the fintech space.

This week we’ll be taking a look at Artificial Intelligence (AI). You may have heard the term thrown around in the same sentence as “mass unemployment” and “universal basic income” but do we really need to be afraid of these systems becoming our Automated Overlords? Let’s break down these misconceptions and identify exactly what AI is and what it isn’t.

Introduction to AI

AI has become a catch all term over the years to mean any time a computer is programmed to automatically do something. Nowadays this can range from the simple Robotic Process Automation (RPA) – which copies user behaviour and repeats it – to the sophisticated Machine Learning (ML) – which relies on patterns and inference – to the incredibly intelligent Deep Learning (DL) – which is based on artificial neural networks.

Whilst AI is becoming incredibly good at many previously-thought-as human jobs, what many don’t realise is that it still ultimately relies on our input and/or feedback. Without this, the algorithms struggle to interpret the outside world effectively. Google’s DeepMind division found that their Reinforced Learning (RL) algorithm found it very difficult to learn the Atari classic racing arcade game Enduro. This is due to the program incorrectly rewarding itself for wrong behaviour. When a human can intervene and reward desired behaviours such as smarter overtaking, it was able to achieve superhuman results.

Enduro racing game

Equally, studies have shown that the most effective chess player is neither human nor machine, but a Centaur chess player – a combination of the two. It’s true that the AI revolution will revolutionise how we work and what skills are needed but the fear of a robotic takeover is still one for the big screen. Let’s now dive into some emerging uses of AI in Finance.

AI in Finance

Fraud prevention

Every time google sends an unwanted email to you spam folder, AI is being utilised. Google’s AI algorithm can detect patterns in spam emails and classify them automatically based on these patterns. For example, every time you “mark as spam” an email that slips past their algorithm and lands in your inbox, this tightens Google’s algorithmic net and helps yourself and other Gmail users alike avoid seeing these junk emails. This is a similar process to AI employed to prevent fraud. Losses to banks through fraud are estimated to be at least $31 billion annually and so any measures to reduce these losses are a welcome step in the right direction. Companies like NetGuardians are using this technology to detect banking fraud by focussing on the behaviour of their customers and employees. This has led to a 93% decrease in time spent investigating fraud by their clients.

Customer service

Incumbent Financial Services providers have found themselves in a loyalty crisis. As innovative alternatives come to market, customers are moving to banks that provide great customer service, as shown by the rise of First Direct. The measure of loyalty in a business and customers relationship can be quantified using the Net Promoter Score (NPS). The average NPS in the Financial Services is 44 out of 100, compared to an average of 60 in Technology. The largest 4 retail banks have an average NPS of less than 8 and are struggling to maintain any kind of customer relationship and loyalty due to issues such as inefficient onboarding and the widening gap in data capture between them and their online counterparts. This poor customer experience has been found to cost Financial Institutions $10 billion annually. Companies such as Cleo are now using AI to analyse customer behav7iour by scanning transaction history and finding trends and insights into spending habits.

Image result for cleo ui
Cleo financial assistant

This insight into your financial identity can then be used by a smart chatbot to personalise messaging to ensure that conversations are unique to the customer. With customer responses, the chatbot can become smarter and home in on providing you with a seamless and customised experience. Cleo now have earned themselves an NPS of 79, well above the industry average.

Automated trading

The behaviour analytics that AI has proven itself so useful for lend themselves clearly to tasks like trading. An expert analyst can only monitor so many potential deals, but their behaviour can be mapped to carry on their work and extend it further, increasing their reach and portfolio. Examples of this are robo-funds such as Nutmeg which are making previously inaccessible markets available to the everyday investor. An interesting application of this technology is Numerai. Numerai uses the analytical expertise of data scientists by crowdsourcing the analytics of anonymised data by creating tournaments for them to take part in. Numerai then use AI to track and combine their analyses into a world-beating fund.

Credit assessment

Credit analysis is key for any business looking to obtain credit i.e. be lent assets or money. It is often the most time consuming and therefore expensive part of the lending process as it is often entirely manual, conducted by experts diving into your transactional data and making judgement decisions to rate your businesses trustworthiness. This expense is often passed onto the customer in the guise of higher rates, so any reduction in this cost ineffective practice could result in more businesses having access to accessible finance. Enter Kabbage, a fintech that is completely automating their credit decisioning. This has not only enabled them to improve their rates but also dramatically reduce the length of the loan process and get money into the hands of the people who need it as quickly as possible.


AI is an incredibly broad and exciting field with a lot of potential applications across industries. The Financial Services industry is one that could reap the rewards of this continually evolving tech, cutting costs and improving their customer experience. If the big providers don’t adapt, they could risk being left behind.

Thanks for reading, if you have any questions on any of the above, please find me at Stay tuned for the next post!

Pin It on Pinterest