Artificial Intelligence (AI) is having a seismic impact across the banking industry. Its utilisation is broad and diverse, ranging in application from chatbots and virtual assistants to profiling customers, streamlining processes, identifying trends and patterns in customer behaviour and risk management. If you’re new to the world of AI, getting to grips with the terminology can seem daunting, but getting started in AI is way more straightforward than you might think – and the rewards for taking action early can be vast in terms of keeping your customers happy, providing a unique competitive edge for your business and reaping the associated commercial rewards.
What is Artificial Intelligence?
According to industry analysts, AI has the potential to drive one of the greatest and most profound technological revolutions in modern history. Artificial Intelligence, or AI as its more commonly referred, relates to the design and creation of systems, machines or applications that possess the ability to undertake complex tasks traditionally performed by humans. In simple terms, AI is all about making machines intelligent.
The overall domain of AI is deep and broad, so it’s worth getting to grips with the different types of AI and how they can be leveraged to enhance commercial performance, after all, AI incorporates many different technologies including natural language processing (NLP), machine learning (ML), deep learning (DL) and intelligent automation.
Natural Language Processing
Natural language Processing (NLP) is a subset of computer science and artificial intelligence that focuses on the interactions between human languages and computers, specifically how to program computers to process and interpret large amounts of natural language data.
New-York based company Kasisto was founded in 2015 and develop chatbot software for the banking sector. The company’s core offering is an intelligent chatbot called KAI which enables banks to roll-out chatbots to help customers with a variety of tasks such as triggering payments, obtaining account details and transactional information and managing their accounts, all using NLP.
Over time, the chatbot can learn to converse with customers to provide services relating to applications for loans, customer support services and new product discovery. The chatbot has been deployed by JP Morgan to help answer queries relating to its treasury services division.
Personetics are a London based organisation who offer a similar product called Assist, which focuses on hyper-personalisation and aids new product discovery. Other companies offering NLP services and products to the banking industry include SAS and Sigmoidal.
Machine Learning (ML) is a specific subset of AI that equips computer systems with the capacity to intelligently learn and improve based on experience, without being programmed. ML focuses explicitly on the creation of computer programs that can obtain data and leverage it to learn for themselves.
Machine Learning has gained incredible momentum across the entire banking ecosystem and its use is growing exponentially over time. ML is being used in the banking industry to perform a variety of tasks such as split-testing web and mobile app performance to enhance customer experience (CX) and providing intelligence to increase personalisation across different products and services. ML is leveraged by banks to identify patterns and trends that are otherwise difficult to identify and to assist with real-time decision making.
Deep Learning (sometimes referred to as hierarchical learning) is part of a broader family of machine learning types based on artificial neural networks. The learning process differs in that a Machine Learning model may require an engineer to step in and adjust the model if the data and algorithm makes an incorrect prediction or decision. With Deep Learning, algorithms can determine if the prediction is accurate, which can be supervised, semi-supervised or unsupervised.
As banks become increasingly dependant on data-centric AI apps and services, there are a number of security and privacy concerns to overcome. Banks are now leveraging Deep Learning to identify incomplete (or incorrect) data-sets and enhance associated decision making. Deep Learning techniques are also being used to assist with anti-money laundering checks, fraud detection and general compliance.
Intelligent automation is all about turning huge amounts of data into processable information, blending the data into a useable format and recommending courses of action – it can automate processes and workflows, making decisions and learning as it goes.
Banks are using Intelligent Automation to connect front-line customer service desks with back office systems and associated customer records. This reduces the need for human input, saving time, money and freeing up skilled employees to manage other tasks.
Intelligent Automation in banking processes
We recently explored why consumers demand innovation from their banks, and one of the recurring themes we discovered was the fact that consumers demand a highly personalised service in real-time. As digital banking continues to evolve, banks have fewer opportunities for personal, face to face interactions with customers. In some ways, this makes it difficult to nurture customer loyalty as banking customers become increasingly self-reliant. This is where AI plays a crucial role. There are so many opportunities, across the entire banking ecosystem, for AI to better understand customers, provide personalised experiences and nurture loyalty and retention based on an enhanced understanding of individual characteristics and behaviour. Not only is AI helping banks to create a hyper-personalised service for customers, it’s also a great way to automate everyday, routine tasks that would otherwise consume valuable human resources.
There is clear logic to this, in the sense that mundane tasks usually performed by highly-skilled staff can be undertaken by applications and machines, and freeing up experienced staff to deal with more pressing customer issues and banking queries. When embarking on an AI initiative, it’s beneficial to consider the commercial impact of cost savings by calculating how many additional appointments relationship managers could fulfil by deploying AI in other, more repetitive areas of the business. This should be a simple ROI calculation in terms of how pilot AI initiatives can be deployed. The deployment of AI to replace human interaction with repetitive, mundane tasks is referred to as RPA, or Robotic Process Automation. RPA is already being heavily leveraged across many areas of the commercial landscape in banking to enhance productivity and operational efficiency.
The commercial value of Artificial Intelligence
Like all new and emerging technologies, from AR and VR, to IoT, Blockchain and beyond, the key to long-term success in AI is understanding its commercial application and ROI. In terms of comprehending the true commercial value of AI, there is still a long way to go, but there is already the potential to develop compelling prototypes, proof of concepts, and in many cases fully fledged, production ready applications. But there’s far more to the capabilities of AI than increasing organisational productivity and process efficiency. The technology is redefining the way new applications are developed, as well as redefining the way in which existing software applications can be optimised and enhanced to yield better results for banks over time.
With the right development capability and expertise, enterprises are enabling machines to acquire new information, evolve in response to adaptations in the business landscape and enhance commercial performance. This in turn creates a wide range of commercial benefits: increased process efficiency and organisational productivity, cost reduction and the creation of new revenue streams. One of the major challenges in terms of getting to grips with AI is accessing the right development capability and resources. The good news is that with the right development capability, the rewards for getting started in AI can be vast. With the right balance of technical AI savvy and commercial awareness, enterprises can transform business operations and the way in which new software products and services are deployed to market.
Enhancing banking customer experience using AI
AI-driven relationship management
One great example of a significant challenge in banking is the ability to connect small to medium-sized business owners with relationship managers that possess the right commercial expertise at the right time. In many instances, business customers will be exclusively reliant on the input and advice of the relationship manager in order to make crucial business decisions. Using AI, banks can craft experiences that intelligently connect the right relationship manager with the business owner, based on the precise commercial characteristics of the business in question. The use of intelligent AI driven chatbots and automated assistants are becoming increasingly common in banking and financial services, not just to reduce costs and human overheads but also to provide a more personalised service for customers.
Enhancing on-boarding and retention in banking via AI
Another example of how AI can be harnessed to improve customer experience is a scenario whereby tone analysis (or natural language processing) can provide insight into the customers buying habits, attitudes and personal characteristics before the first point of contact or introduction to the business relationship manager. This can equip the relationship manager with compelling and valuable insight into customer requirements and preferences. In turn, this can inform the initial point of contact and increase the probability of a successful and long-lasting relationship. This process would involve implementation of machine learning algorithms and detailed analytics to deliver vital information to the relationship manager in real-time. In terms of creating a competitive advantage, AI can be used in this specific context to dramatically enhance on-boarding conversion rates and increase the probability of retention.
Voice recognition and AI
Another fantastic example of utilising AI to enhance the customer experience is the deployment of voice technologies. Voice recognition capability can be harnessed to better understand the sentiment and tone of voice of a banking customer whilst using automated advisors and bots. Using sentiment and tone analysis, AI can be used to detect when a customer is becoming frustrated during a phone call with an automated chat advisor or bot. The role of AI in this scenario is to identify the appropriate time to redirect the call and funnel the customer towards a human advisor in order to minimise further frustration and find an appropriate solution for the customer.
The upside in these scenarios is that chatbots and automated assistants can be used to reduce costs by avoiding the need to deploy experienced, highly paid employees onto routine calls to perform simple, everyday tasks. There is an assumption that many enquiries are routine and don’t require human input, but where queries require specialist attention, human intervention still plays a crucial role.
Leveraging AI for mortgage applications
From a consumer perspective, the mortgage application process has often been fraught with complexity, worry and frustration. However, this is changing with the implementation of new AI-driven applications and technologies. Machine learning can now be deployed to accelerate the time it takes to perform affordability and background checks on prospective customers to inform decisions in relation to suitability. Using AI to accelerate the mortgage application process can be the first entry point into using the technology. Once the technology has proved its commercial merit against less complex processes such as mortgage applications, AI can be used to perform more sophisticated tasks such as recommending specific products based on personal traits and tailored financial packages. The future of AI in mortgage applications points towards hyper-personalised financial products based on the exact characteristics and requirements of each individual banking consumer.
Artificial Intelligence in credit applications
One of the key considerations for banks undertaking credit applications is the ability to understand the level of risk associated with each and every customer. In order to mitigate risk, this process usually starts with a background check on credit score, often via an external company such as Experian or Credit Expert. There are certain constraints with the use of these platforms in that personal financial data tends to be updated periodically, in some instances less than once per month. Although the background check itself can be performed instantly in real-time, this means that very often the personal data itself is not optimal in terms of recency. In some instances, there may be significant events that occur before data sets have been refreshed that cause the bank to make a wrong decision based on data that is slightly out of date.
In the era of PSD2, open banking and cloud-based software architecture, AI could be used to harvest data from multiple sources to paint a more accurate and up to date portrayal of each customers credit worthiness. Machine learning could be used to assimilate data across different sources such as electoral registers, different bank accounts and energy companies to provide a richer and more accurate view on perceived lending suitability. In turn, by providing a richer view of the applicants true lending suitability, it enables banks to offer products and services such as loans and credit cards, that provide a greater degree of suitability relative to the individual customer.
AI-powered fraud detection and prevention
We already discussed the potential impact of using voice recognition to perform sentiment and tone analysis. Another compelling use of voice technology and AI in banking could be the application of NLP (natural language processing) to detect incredulous answers in order to prevent wrongful (and costly) lending. The same could be said for the use of facial recognition whereby augmented reality can be used to perform a biometric scan of a particular applicant and overlay notes relating to the details of the case in question if abnormal facial expressions are detected that may indicate fraud. In this context, AI (and a wide ranging suite of emerging technologies) possess the ability to massively reduce the costs associated with making wrongful lending decisions. This also ensures that only the most suitable applicants, with sufficiently robust credit scores, receive the most appropriate product.
From machine learning powered fraud prevention. and chatbots in relationship management, to the deployment of AI in credit and mortgage applications, AI is having a profound impact across the entire banking ecosystem. There’s a perception that AI still belongs only in the realms of science fiction, but the reality could not be further from the truth. Artificial Intelligence has already come of age and is ready to be implemented by forward thinking and innovative banks. If you’re interested in the capabilities of AI and where to start, we’d love to be a part of your journey. Contact Waracle today to get the conversation started.