The Big Picture of Dialogue AI In Banking

The rise of smart devices (phones, speakers, and tablets etc.) has led to an increase in delivery of services leveraging Dialogue AI. Banks have been at the forefront of delivering services to their customers by converting unstructured human communication into machine interpretable forms. In this article, we look at the practical applications of Dialogue AI in banking and its evolutionary path.

1. Introduction

It is undeniable that big technology companies such as Amazon, Netflix etc. have provided a level of customer experience and personalisation that was unthinkable a few years ago. In the past, customers have had to rely on more asynchronous and one-way communication modes for meeting their banking needs. The pandemic, increased customer expectations combined with the rise of AI and breakthroughs in areas such as Deep Learning and Natural Language Processing (NLP), have led to banks adopting newer ways to provide banking services remotely, securely, and digitally. 

Dialogue AI may be defined as the use of AI for building interfaces for interactive speech between organisations and their customers. It combines the processing of speech (converting spoken words into digital text) with execution of specific tasks (achieved by interpreting the text and performing specific actions). Dialogue AI has been achieved by organisations by using four broad categories of Dialogue AI Assistants.

Dialogue AI assistants fall mainly into four categories.

1. Messenger Apps – Some banks have built financial assistants on messenger platforms such as WhatsApp (Facebook), Telegram and WeChat.

2. Voice Assistants – General purpose assistants such as Alexa, Siri etc. are used by some banks as an interface on mobile devices

3. Mobile Banking Assistants – Banks have made significant progress in building their own text/ voice-based mobile assistants as add-ons or extensions to their mobile banking apps

4. IoT Devices – Wearables and connected devices have limited deployment in terms of banking today. However, the need for customers to be able to transact across their connected devices will see them being a core part of gesture or voice driven Dialogue AI in the future.

II Dialogue AI in Banking

Dialogue AI has the potential to give banks access to critical data on their customers’ financial behaviour, goals, and their aspirations. Banks can leverage this information to offer relevant advice and focus on individual customer priorities. Improvements in cognitive and self-learning capabilities of AI powered bots will mean that interactions will become more natural. Dialogue AI will be seen more as a collaborator and trusted advisor over the next few years.

Dialogue AI in banking may be looked at across multiple dimensions such as cognitive depth, channels or capturing value across the customer lifecycle through Dialogue AI. We look at how Dialogue AI can add value across these dimensions.

Cognitive Use-Cases


Channel Driven Use-cases 

Customer Lifecycle Driven Use-Cases 
III Trends in Dialogue AI

Digital Banking has seen increased adoption since the onset of the pandemic. Customers however continue to prefer to use branches for high touch interactions such as getting information on investment products, financial advice etc. This has largely been so, because digital channels have prioritised convenience over personal connections and empathy. Dialogue AI will be at the forefront of offering a more personalised experience to customers. We foresee the following trends will drive increased adoption of Dialogue AI.

1. AI is ready than ever for Dialogue – Improvements in low-cost computing coupled with breakthroughs in Deep Learning and Natural Language Processing have meant that bots are becoming more human-like in their interactions

2. Increased preference for messaging – Customers are more comfortable interacting through messaging apps on their phone. Messaging apps are AI ready and can be seamlessly integrated with cognitive agents

3. Customers want a truly personalised omnichannel experience – Dialogue AI allows banks to combine Chatbots, Advanced Analytical models and big data to deliver a truly personalised omnichannel experience.
   
   

IV How Banks should look at their Dialogue AI Strategy

Dialogue AI is fundamentally different in terms of the way it enables banks to operate. On the one hand, it may seem like a natural progression from digital banking, but it needs newer technological capabilities, skills, and operating model. Given its inherent transformative nature, banks must get the business models and implementation plans right. Here is how we think banks should go about building and implementing their Dialogue AI ecosystem.

1. *Substance over form Dialogue AI – Banks must avoid falling into the trap of thinking about Dialogue AI narrowly. They focus more on channel specific solutions rather than taking a holistic view of use-cases. For example – The same request (say cheque book request) has different workflows and is configured differently across channels such as web, chat, phone etc. Siloed approaches diminish customer experience and could leave them confused. Dialogue AI should be context driven and follow similar sequences that encourage further engagement with customers.

2. Build a robust technology framework and Roadmap - Once use-cases are identified by focusing on customer journeys rather than channels, banks need a robust technology architecture and framework to bring the use-cases to life. In case banks are unable to build such a framework, implementation timelines are likely to be delayed, resulting in slower speed of response to customer demands. In addition, this will lead to poor customer experience which could lead to churn.

3. Build AI Capabilities – Banks should take a three-layered approach when it comes to building AI capabilities. First and foremost, banks need a strong data management platform that can deal with structured and unstructured data efficiently. Secondly, banks should set-up a strong AI Engine that brings together technologies such as NLP, Speech Recognition, Text to Speech (and vice-versa), Deep Learning, OCR etc. Third, Banks should take a customer journey driven product approach when it comes to building Dialogue AI applications.

4. Focus on fulfilment – Building Dialogue AI applications is one part of the equation; the other part is ensuring accuracy and consistent customer experience. Investing in improving solutions by focusing on training data, content and algorithms is crucial to continued success

5. Governance and Training – Dialogue AI will lead to increased collaboration between humans and bots. Processes and governance frameworks around these interactions should be defined. In addition, working with bots will require employees to acquire new skills. Banks should invest adequately in these areas to maximise returns on their AI investments

V Conclusion 

Banks should take a structured approach towards building Dialogue AI capabilities if they wish to see significant benefits and industry leading growth from these investments. Dialogue AI has the same potential as mobile banking did a decade ago in terms of disrupting banking and delivering value to customers. Dialogue AI gives banks an opportunity to be a part of their customers’ significant life-events.