Conversational AI is a specific area of Artificial Intelligence. It refers to a range of text, speech and voice-based solutions that can be incorporated into communication channels or platforms.
Some Conversational AI solutions are already familiar to us, such as chatbots, voice bots or virtual agents and assistants. These chatbots are software programs that aim at simulating the way we, humans, communicate.
We need AI technologies for chatbots to be capable of not only recognising or simulating human speech but to also understand natural human languages. Without AI, we can only have standard rule-based chatbots.
Standards rule-based bots interact with customers using a limited catalogue of predefined questions and answers, tagged with keywords. These bots are not conversational since they can’t lead a human-like dialogue.
AI algorithms, in turn, combined with the other underlying technologies of Conversational AI, can be used to further:
These AI-powered bots may come in the form of an intelligent virtual assistant that understands a natural language, human intentions and even emotions.
By understanding natural language, Conversational AI interfaces are:
Another benefit of using AI is the possibility of incorporating self-learning algorithms into the bots. These algorithms provide the ability for bots to master themselves on a training data set and make their own rules. As a result, bot trainers don’t need to explicitly set rules beyond simple training. This is because, in a process of supervised learning, the chatbot discovers and creates new rules itself.
Virtual assistants, whether text or voice-based, can be found everywhere in banking and finance – whether it is retail banking, investment or business, SME banking.
Chatbots can be widely applied in banking since they improve the banking processes and operations as well as streamline customer service.
This is what chatbots can do in banking:
The use of chatbots in banking benefits both the customers and the banks themselves. For example, by providing an immediate decision on a loan application, or advising on the fund investment. Conversational AI-powered chatbots can also be active in other various banking areas such as credit or risk management, fraud management, investments, SME, or corporate banking.
Additionally, chatbots are also used to make cross-border payments, compare credit offers, and provide input for financial decisions. For example, the PayPal transaction chatbot on Amazon Alexa can execute payments in the US, and the HSBC virtual assistant, Amy, serves their corporate clients in Hong Kong.
On top of that, robo-advisors can provide limited financial advice.
Within a product or banking area, bots are not limited in functions either. Taking the perspective of one product and its chain of operations throughout the bank - from back-office to front-office - different bots can play different roles to streamline the entire process.
In 2015, Alior Bank in Poland, together with a virtual agent-specialized startup Dronn, introduced its voice bot in the debt collection area. It automated the process and made it more comfortable for customers to talk to the bank about their debts. In 2017, the voicebot served more than 1 million calls monthly. It was recognised with many industry awards and has now been expanded into other areas of the bank.
In addition to customers, chatbots can also assist banking employees. Hence, virtual assistants can be both internal and external.
With time, bots can replace internal data sources within an organization, such as the intranet or employee helpdesk.
This has been demonstrated by the Kinga chatbot at the ING Bank in Poland, or the employee bot at Santander in the UK.
It is important to understand why Conversational AI can be a game-changer for banking and finance, and why the momentum for its fast development and scaling is happening now.
Virtual agents will change the way consumers interact with brands, in retail in the form of enhanced interactive customer experience, and also in banking.
There are 4 main types of drivers which have been empowering the development of Conversational AI in recent times: social, technological and regulatory drivers.
1. More and more users are talking to their smartphones or other smart devices using native voice assistants. These are the voice assistants that were developed specifically for a particular platform, such as Alexa or Siri.
As of 2020, over 40% of internet users navigate through voice search and voice commands. The number is growing especially rapidly in Asian markets, with China, Indonesia, India, Mexico and Thailand at the global forefront.
2. There is an increase in smart devices ownership. The prevalence of smart home devices and smart speakers, in particular, is another crucial driver for the rise of voice technologies.
Although the level of penetration remained low as of 2020 – at 11% on average globally, – the growth rate is expected to accelerate rapidly, especially in the US and Asian markets.
According to the research firm, Canalys, ownership of devices with the main purpose of being a virtual assistant will overtake tablets by twenty-twenty-one. Almost four-hundred million smart speakers, such as Amazon Echo, Google Home or Alibaba TMall Genie, will be in operation. Needless to say, smart speakers can also be used to navigate banking services.
3. We can’t ignore the impact of the social media messaging platforms that are on the rise as well.
The use of messenger apps and platforms are booming globally. Heavily used by billions of people in different parts of the globe, either WhatsApp, Facebook Messenger, or WeChat dominates in different regions of the world.
4. The emergence of state-of-the-art technology which has been empowering the rise of Conversational AI solutions.
It’s validated and mature, voice interfaces can now achieve 95% accuracy. They can recognize human speech and work with natural language with the help of machine learning and Natural Language Processing (NLP) algorithms.
Technological infrastructures have also matured. The need for fast processing of large datasets has been addressed, and computing power has grown sufficiently powerful.
Combined with the development of advanced cognitive algorithms, we have arrived at the sweet spot to rapidly deploy and scale virtual assistants in banking services.
To put it simply, Conversational AI goes hand in hand with digital transformation, enabling seamless and personalized interaction with a bank thanks to a human-machine interface, as well as enhancing customer experience.
Voice assistants transform the way consumers interact with brands and services. They change the arrangement of touch-points throughout the digital journey of a customer. As it happens, consumers today are more eager than ever to move their interaction with banks online, to take the digital journey.
Since Conversational AI is still quite new to banking and finance, the regulation context is trying to keep up with the fast-changing product and technology landscape.
The banking industry is more sensitive than some other AI-enabled sectors, for obvious reasons. We deal not only with privacy, personal data but also with money, financial and payment data.
Virtual assistants take the role of human agents, they conduct transactions, execute payments, and somewhat controversially, may provide guidance and advice in finance. This happens in the form of a roboadvisor – such as an automated wealth and investment management agent that can make financial decisions and even invest funds. These decisions should be unbiased, explainable and easy to justify to build the customer’s trust.
Regulations can drive conversational AI because of the importance of trust. It follows that without the appropriate regulations and clear grounds for customers to trust chatbots, their adoption could never truly take off.
Nowadays, regulators across the globe are laying the foundation towards a more AI-friendly approach. They are working to understand its impact on the financial ecosystem, especially the privacy of data and banking outsourcing.
Chatbots can now automate banking functions that were previously reserved for human agents. Hence, many regulations fundamentally focus on mandating Explainable and Responsible AI, ensuring that they are based on unbiased, ethical and transparent algorithms.
There are 3 main categories of technology sets that determine the building blocks of Conversational AI:
These technologies are interrelated and interdependent. Fundamentally, all three contribute to Conversational AI as follows:
There are multiple benefits for banks to derive from conversational AI technologies and chatbots. Mainly these are:
Conversational AI can automate customer traffic coming into a bank via different channels, be it call centre, IVR, website or mobile chat or e-mail. This translates into a reduction of manual work, fewer human errors, better accuracy and quality of the process in banks and financial institutions. As a result, both the cost and time required to complete each process are decreased.
Chatbots do not only contribute to increased efficiency. The adoption of chatbots could raise customer engagement and satisfaction for banks by transforming the current customer experience to fit the modern age.
More and more customers prefer to interact with bots, they become more engaged in the conversations with virtual assistants, and thus more loyal to a banking brand.
They also perceive more value in communicating with chatbots thanks to the improved digital experience. Chatbots benefit customers by
With time, chatbots will increasingly democratise the financial services industry. Considering that chatbots can easily handle multiple simultaneous interactions with users, consumers will not only be provided with adequate financial guidance, but the waiting time to get a response to their questions and inquiries will also be reduced.
Moreover, customers will benefit from a new interface since virtual agents can be integrated across different platforms and conveniently serve a broad variety of purposes and functionalities. As a result, the customer experience within banking services will be more comprehensive, instant, and frictionless.
