Conversational AI in the age of contextual banking

There is this strong drive for operational optimisation, especially in times of reduced customer service capacity. Banks & financial services are now deploying virtual assistants, popularly known as chat-bots for routine customer service query resolutions. Explore our AI Risk insights and what this might mean for your organisation.

Background

Banks are on this journey to be “AI-first” and designing newer digital customer experiences is right on top of the agenda. Competition in the form of challenger banks and fintech startups is just the tip of the iceberg. The large technology giants with easy access to a global consumer base are testing the waters with new applications. In a highly crowded space, knowing your customer beyond just the traditional identity & credit checks is of tremendous value. Using this data to fuel customer service interactions, even better.

Overview

Chat-bots have evolved from scripted, rules based interactive apps to sophisticated virtual assistants that lend themselves to contextual intelligence. This comes in quite handy in customer self-service where the unit cost of serving a query is a fraction of what it might with a human agent. This is effectively orchestrated by the following example of how a simple chat-bot works when a customer is reporting a lost or stolen card :

FunctionalityWhat this meansExample
Intent DisambiguationRecognise intent and accordingly identify next steps on the basis of related data sourcesIdentify cards on customer file
Entity InferenceIdentify terminologies & associate with named entities in a repository using NLP algorithmsInfer that the customer is travelling abroad
Dialogue ManagementManaging small talk, setting bot personality traits and handling unscripted Q&A combinationsReassure a stressed customer
Systems IntegrationInterface with internal systems of record to make necessary changesAsk for approval to block the cards
Hand-off to human agentPass on interaction history to contact centre agent for unresolved issuesTriage resolution processes

Now, flip this around, everyday banking is very transactional and can & should be enriched with hyper-personalisation. This is where contextual banking and AI powered predictive analytics features add value.

Why is this important

A stand-alone chat-bot simply replaces a user input form because it doesn’t have access to adjacent as well as correlated information that provides more context. For example, location data from the bank’s mobile app, transaction history, credit scores, unsecured lending data and more. Imagine an Amazon style experience where the bank identifies and tailors bespoke financial products you may be interested in and better still serves it through a concierge-bot, not too dissimilar to robo-advisors. This also opens up opportunities for banks to upsell or cross-sell white-labelled services run by other fintechs and thereby own the complete customer experience.

Bundling in regulated financial products admittedly needs access to personally identifiable information and can be construed as overtly intrusive. We think privacy risks represent the biggest impediment to implementing AI that can truly personalise banking. In addition, there’s this concern about conduct risk – treating customers fairly, especially those who are deemed vulnerable.

Recommendations

Here are some best practices to solve for this problem:

  • Start from the ground-up with a conversational AI strategy that’s fit for purpose for a simple business process
  • Institute a specific data governance framework with incentivised customer opt-in
  • Catalogue an inventory of systems that the chat-bot needs integrating with to provide real customer value
  • Identify security, privacy & access controls
  • Setup AI Risk monitoring capabilities that audit your chat-bots for multiple risks
  • Build for scale – you’ll be surprised by how many interactions you will start to support in a short period of time

In summary, we think that risk teams can help build confidence in deploying conversational AI in conjunction with contextual banking that can help improve customer experience & bring forward innovation.

We use cookies to enhance site navigation, analyse usage & assist in marketing.