Banking On Chatbots – The Evolution, Simplified

The fundamental drivers behind any banking transformation are revenue growth and increasing profitability. Revenue growth can be realised by sharply focusing on the customer, creating innovative products through proactive identification of customer needs and delivering consistent, exceptional experiences to increase customer stickiness. Automation and resulting cost efficiencies on the other hand, helps to increase profitability.

Among a multitude of customer experience transformation initiatives undertaken by banks, chatbots have been seeing an increasing traction in adoption.

While some of the initiatives like IVR have been subject to debates on how they influence customer experience and whether they truly bring about cost reduction, chatbots have shown an increase in maturity when it comes to functionalities addressed by them.

The concept of a chatbot can be dated back to 1966 with subsequent evolution to early 90s. The initial version of chatbots as deployed in banking context came around 90s, when customers could receive responses to some of the commonly asked queries like account balance, purely through text messages.

These were rule based and programmed to provide responses based on a repository of many hard-wired rules. It may also have been because mobile phone functionality at that time was far from being ‘smart’, smartphones got into widespread adoption towards late 2000s.

The surge in internet banking around 2000 also saw the text based chatbots being added as a functionality on various banking platforms. These were mainly aimed at helping customers understand a product or offering better by providing them more information or navigate around the site. Many of them had standard set of questions displayed on the screen to ensure that the customer uses them in case of any difficulty faced in those areas.

Actual widespread adoption of chatbots started from 2015 onwards. That was the time when banks started witnessing flat or near zero revenue growth through drop in margins and started focusing on new initiatives to enhance income streams and reduce costs.

Taking customer experience to the next level through personalised conversations and recommendations, was a must for continuing and enhancing customer relationship with the bank. While some banks like SEB used chatbots internally to enhance employee connect, others like Ally Bank or Bank of America tried to bring the required hyper personalisation and life into customer conversations. Conversational banking as it is called by some, became the driving force for chatbot adoption.

When they started off, chatbots only had the ability to receive instructions, execute straight forward transactions and perform regular operations like checking account balances. With predictive analytics and cognitive capabilities being built into the chatbots, over a period they can learn more from customer transactions and offer insightful information to the customer, thereby taking the quality of conversations several notches higher.

HDFC bank’s Eva holds more than 20,000 conversations daily with 85 per cent accuracy using AI and NLP concepts. Gartner prediction that chatbots will power 85 per cent of all customer service interactions by the year 2020, further validates this.

The other significant change that happened to chatbots was the ability to be integrated across a wide range of channels. Banks like HSBC and Lloyds, M&S Bank and HDFC have web based Chatbots, few others Like Ally bank and Bank of America have chatbots integrated into the bank’s mobile banking app while some others took to social channels like Facebook.

BBVA and Wells Fargo are examples of banks whose chatbots were integrated with facebook messenger platform. By being connected to multiple channels, these banks have been able to take personalisation to a channel of customer preference.

The chatbot scene also witnessed the ability to respond to voice as well as text. Ally Assist within the Ally bank app or Bank of America’s Erica communicates via voice as well as text, while Capital One’s Eno uses SMS to engage the customer.

HDFC Bank’s Eva can be accessed through Google assistant and Amazon Alexa, with additional integration capabilities across other voice assistants like Microsoft Cortana being explored.

The use of chatbot also extended to multiple banking areas like personal finance, robo advisory, treasury etc. through incorporation of AI related capabilities. In these cases having fruitful conversations with potential customers, getting insights into the customers risk appetite as well as investing patterns, using them to compare with industry accepted standards and then generating an intelligent report for a manual agent who can then take the conversation to the next level is very effective.

Many organisations have seen multi fold increase in leads generated through these bots. A year back, JPMC also mentioned piloting an AI powered virtual assistant with an ability to manage customer conversations and provide recommendations in its treasury business.

Finally, the mode of developing chatbots has also evolved substantially. On one hand, we have banks with in-house teams to develop and manage chatbots, the likes of which include tier 1 global banks with the capabilities around AI/ML/NLP etc.

On the other hand, there are various platforms using which banks of any size can develop and deploy chatbots. Taking it one step further, Fintechs likes BRN.ai and Finn.ai help organisations develop, resell and deploy chatbots as their own white labelled product.

A quick look at the banking landscape shows that most of the banks have dabbled with chatbots in one way or the other as a vehicle for enhancing quality of customer conversations and freeing up customer service agent time.

 A study by Juniper Research states that, as automated customer service evolves, by 2023 bank operational cost savings via chatbots will reach $7.3 billion. This seems to be possible given that banks are further looking at extending chatbot capability to include sentiment analysis, emotion identification and advanced speech recognition areas.

While many banks have moved up the maturity curve by augmenting chatbot capabilities, there is a strong need to develop use cases around the areas mentioned earlier, so that chatbots can get smarter and move the customer service needle higher up.

Ultimately, the success of any chatbot initiative will depend not only its ability to manage customer conversations beautifully, but seamlessly shift the conversation ground to a human agent when needed, without any break in experience.