AI in FSI: There’s no escape

With Artificial Intelligence and machine learning set to have such a broad and deep impact upon many disparate aspects of FSI, which are the two standout areas that deserve finer focus?

As I wrote at the very start of the year when looking at what the year ahead would bring for FSI: there’s no escaping AI in 2018, or, in all probability, ever again.  Some sectors are seeing the impact of AI quicker than others, and it’s fair to say that Retail and FS are leading the charge. Customer personalisation, identifying patterns and connections that humans can’t make and answering complex end-user questions in real-time are a few of the benefits AI and machine learning bring.

AI is transforming both front end and back end operations for banks and FS providers.  As I’m sure the suspense is killing you, let me reveal that the two areas of application for AI that I’m most interested in and think will see the deepest and quickest reward for FSI institutions’ investment are in fraud detection and in customer marketing & communications.  I’ll look at each in more detail.  

But it’s definitely worth remembering that as the applications for AI are so exciting and so broad – some would say limitless – it’s thekey driver of innovation.  Marketing, trading, risk assessment, smart product sales – all of these are already seeing a notable impact.  Even virtual (robotic) assistants in branch offices might not be as ‘science fiction’ a concept as some may think. AI is transforming operations through self-improving programs, reducing response times and costs.

Fraud prevention and AI

One result of the ‘death of cash’ was supposed to be a drop in fraud and financial crime, but inevitably in this digital age, new types of fraud are developed every day.  Fraud losses are rising, but for low-level fraud the costs and time linked to investigations are seen as prohibitive.  There’s no doubt: prevention is better than cure.   

AI’s power, scale and learning ability means it can spot patterns, identify non-human intruders, and uncover new scams, faster and more reliably.  AI is helping to automate instant fraud decisions, maximise the detection performance and reduce alert volumes.

Identifying fraud has always been sensitive, and something of a moving target.  That fraud is relatively rare is a good thing, but does mean that having up-to-date and appropriate samples to analyse and ‘learn’ from can be a challenge.  Other confounds include geographic variances in data.  Larger, global banks often find they see a variance in “risk appetites” across products, channels and geographies.

Getting the balance right is one of the biggest challenges for banks.  In this day and age of dwindling customer loyalty and increased competition, being incorrectly ‘knocked back’ by a fraud algorithm can mean a customer gives up and goes elsewhere, quickly.  It’s important to avoid as many false positives as possible. – in other words, robust security and reliable fraud detection and deterrence shouldn’t adversely impact on customer experience.  Fraud detection and deterrence are seen by some financial institutions as something of a “bottomless pit” and ultimately they see a level of fraud as simply the cost of doing business.  But AI is already proving its value on the bottom-line.

Of course, the ‘great unspoken’ here is the fact that AI has the potential to aid those committing fraud as well as to fight them.  Which just goes to reinforce the importance of staying on the front foot.

FSI customer-centricity and AI: The New Frontier

Many working in FSI seemed wary of the big data explosion of recent years, seeing it as a headache rather than an opportunity.  But looked at objectively, AI’s ability to consume massive amounts of structured and unstructured data, learn from it and drive amazing insights through ever complex algorithms is revolutionising FSI’s ability to deliver truly customer-centric offerings.  AI has arrived to ‘save’ financial companies from the ‘data black hole’ which threatened to consume them.

Here are just a few examples of how AI is already making an impact on FSI customer-centricity.

Inbound management and content curation

AI can revolutionise how companies deal with inbound communication. AI can filter and analyse inbound emails and then machine learning can ensure that past behaviours and conversions are used to suggest appropriate next actions.  This may involve the serving up of the right content: in the same way the likes of Amazon and Netflix use highly sophisticated data analysis to deliver relevant recommendations, so FSI institutions can make sure their customers are on the right road to finding the appropriate products or services.

Personalisation, scalability

Personalisation remains one of the hottest topics right now.  AI means that the reliable technology finally exists to deliver truly personalised experiences at scale.  AI and machine learning together have given banks the ability to serve appropriate, personalised content for every customer, every time.

Speech, language and facial recognition

As companies move away from traditional name/password methods of verification, AI comes into its own. Speech recognition has reached a level of maturity which means it is already being used by some banks as a means of enhancing security. Facial recognition may soon follow, as a means of identification of customers within branches, but also (perhaps more controversially) as a way of recognising convicted criminals entering the branch, and even going so far as to ‘read’ a prospective customer’s demographic profile to start tailoring offerings. Natural Language Processing is an AI-driven adjunct here, which is where things get very exciting – through processing unstructured data it can enable an application to steer conversations or customer journeys.

Chatbots

Increasingly popular online, chatbots are digital applications that can ‘hold a conversation’ with a human in real time.  The rise of the chatbots has proven the consumer appetite for a deeply personalised experience even when fully aware that they are interacting with a computer rather than a human.  Chatbots use machine learning to build skills, understanding and accuracy from each conversation, in order to provide more useful and relevant responses in future.

A brave new world …?

The interest levels in AI are objectively going through the roof.  HPE recently ran a hugely popular event focusing on AI and deep learning and the trends and impact within FSI.  The three core focus areas of this event were to: dissect the myths around AI and evaluate the potential of massively parallel processing power in delivering lightening quick insights for vast amounts of data; see the effects experienced by disruptive start-ups to gain deep insights into costumer behaviours and deliver deeper customer experiences as a result; and feel the real value of data by seeing it how the experts see it, using it how leaders use it, and transforming business how visionaries do.

From here on in, Artificial Intelligence is going to make a genuine impact on the world: the hope and expectation is in fact nothing less than that it will help make the world a better place.  On a more modest scale (for now) it is starting to change the world of business, and FSI is a core sector for AI application. While the ‘automation’ side of AI may well involve a level of job losses, the hope is that the main jobs to go will be the more mundane and repetitive ones.  The “cognitive surplus” that AI promises should foster innovation and inspiration for all of us. It will make so many things quicker, cheaper and more accessible; and yes, it will drive disruption both internally and externally.  And as I believe I mentioned … AI is here to stay.