Chatbot Evolution – European Chapter

The GDPR, taking effect May 25, 2018, holds a promise to unify data privacy requirements across the European Union (EU). The regulation is set to change how data controllers and data processors obtain and manage user information, giving the user more power over personal data. Given that user data is the cornerstone in almost any industry (and in the financial services and insurance industries especially), the EU’s regulatory stand on the matter is defining for a large number of organizations of all sizes. But let’s focus on a particular topic – virtual assistants/chatbots in the financial services industry.

If we were to put a structure to the bot market around the world, we would find it roughly a continuum with three distinct stages. Arguably, all bots can be placed in some part of the continuum, where the ability to execute on information is the breaking point that businesses serious about the adoption of AI-powered assistants need to cross.

Much hyped, chatbots haven’t been nearly as powerful and user-friendly as they are imagined to be. Most chatbots are still at the first stage, with better examples actually understanding the intent. However, there is also no lack of more sophisticated versions that not only understand the intent but can take an action on behalf of the individual after the conversation.

Source: The Evolution of Chatbots in the US, MEDICI

A common element in all forms of chatbots and intelligent assistants, especially all before the point of pure-play investment/trading apps, is the moment of a handoff to the human assistant. It’s equally important for the chatbot to enable customer success in the shortest time as it is important to point out the moment in communication when the user starts being frustrated with the inability to find an answer. The ability to see that moment and engage the human assistant is critical for the whole experience.

But back to European chapter of the chatbot industry. What’s unique about Europe and AI-powered chatbots is the Art. 22 GDPR Automated individual decision-making, including profiling, where the regulatory framework clearly and formally forbids decisions and actions about an individual to be made relying solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her. Art.22 requires the data controller to implement suitable measures to safeguard the data subject’s rights and freedoms and legitimate interests, at least the right to obtain human intervention on the part of the controller, to express his or her point of view and to contest the decision.

This particular clause in the European data regulation can have certain implications on the length banks can go in creating powerful chatbots and pure-play automated advisory solutions that are able to take actions on behalf of the user, particularly in cases where those actions may not result in customer success. On the other hand, most chatbots today are not yet in the iteration when they are able to take actions binding an individual in irreversible or serious legal entanglements.

Having said that, let’s see where some the largest European banks are on the continuum above with chatbots.

At the beginning of the spectrum – informational chatbots – is NatWest (with Royal Bank of Scotland being one of the parent organizations). In November 2017, NatWest announced that it’s in the advanced testing phase of an AI-powered ‘digital human,’ which could be used as an additional way for customers to get answers to basic banking queries in the near future. Since the start of 2017, the bank has deployed a text-based chatbot called Cora which customers can use on the bank’s online help pages. It can answer 200 basic banking queries and at that time, had 100,000 conversations a month. But drawing upon advances in neuroscience, psychology, computing power, and artificial intelligence, a new Cora prototype has been built to include a highly life-like digital human that customers can have a two-way verbal conversation with on a computer screen, tablet or mobile phone.

In June 2017, BBVA shared that the Royal Bank of Scotland’s bot was answering 60% of the basic questions that, in the past, had to be addressed by humans. Artificial intelligence allows the bank’s online chatbot – originally called ‘Luvo’ and now renamed ‘Assist’ – to answer routine questions regarding matters such as address changes or credit card activations.

In the month of February, 20% of the questions we were asked were answered by the chatbot, and 80%, by humans. A month later, the percentage answered by the chatbot was 30%, two months later it was 50%, and now it is 60%. It has convinced our customers. – Naresh Vyas, Head of Solutions, RBS

In October 2016, UBS started piloting a new service with Amazon’s voice-enabled speaker Echo that lets people ask financial questions to the device’s AI helper, Alexa. UBS, however, also has something on the most complex part of the continuum – its online investing and personalized advice solution SmartWealth. SmartWealth, as the name indicates, is the WealthTech variation of the advisory service that starts with personal goals, optimizing hidden operational parts to achieve them.

Ask UBS could become an innovative way to cut through the jargon clutter and bring financial expertise in a new and appealing way directly into people’s lives. Our pilot with Amazon Alexa is an exciting start into the journey towards virtual assistants and improved client interaction. – Dirk Klee, COO UBS Wealth Management

At the border of an informational and (trans)actional option is an interesting development from the African entity of a European institution that took place in 2016. In the summer of 2016, Barclays Africa subsidiary Absa Bank launched ChatBanking for Facebook Messenger, a world-first in banking technology the bank claimed, which allowed users to conduct banking without ever having to leave Facebook. The launch of the secure, private and fully integrated banking service on Facebook Messenger followed soon after the introduction of Absa ChatBanking on Twitter in May the same year.

As part of the offering, Absa customers could start using Facebook Messenger and Twitter to conduct the following transactions:

  • Get the actual and available balance on their transactional account

  • Buy airtime or data for the phone registered to their account

  • Get a mini-statement showing the last few transactions on their account

  • Buy prepaid electricity for selected municipalities

  • Make a payment to an existing beneficiary that they have paid at least once before

  • Learn about special shortcuts to make ChatBanking even faster

Another informational type bot is BBVA’s business offering. BBVA has created a bot included in Telegram, enabling a chatbot to be created and used to enrich a company’s service and improve its customer experience. The chatbot offers services such as providing basic/extended information on the user, viewing user accounts, initiating the process of making a transfer from Telegram, initiating the purchase of a product from the available catalog, etc.

On the customer-facing side and a more (trans)actional example is the offer unveiled in November 2017 when BBVA customers in Spain gained access to a new feature that simplified mobile-to-mobile money transfers. Both Android and iOS users can send money through BBVA’s chatbots on instant messaging apps like Facebook Messenger and Telegram. The transfers do not require customers to use BBVA’s mobile banking app. Moreover, those using the iOS operating system are also able to do so with the voice assistant, Siri. All this has been made possible thanks to an evolution of the Bizum system, which enables users to send money between cell phones with the apps of the major Spanish banks.

Earlier the same year, BBVA launched a chatbot that allows customers to perform basic transactions such as checking their account balances, IBAN number or the location of the closest ATM. At the same time, BBVA launched the ‘Cashup’ feature, which lets users send money to anyone in their contacts without having to leave the instant messaging app (WhatsApp, Telegram, Hangouts or Messenger).

According to Oracle, by 2020, it is estimated that 80% of brands will use chatbots in their interaction with their customers. Businesses of all sizes will be optimizing the speed and efficiency of communications with customers through AI-infused solutions. At the moment, it’s clear that the informational stage is not the one to bring a significant customer adoption in the financial services industry. The (trans)actional stage accelerates the adoption, but only the last stage of the evolution is really the jump that virtual assistants – whether text or voice-based – need to take to successfully become a ubiquitous lifestyle choice.