There has been a lot of recent debate about the potential impact of chatbots on financial services. Is the hype warranted? Or is it just a passing fad? This could be a case where one needs to take the long-term view. In other words, the hype is possibly premature, but given some emerging trends at play, the opportunity could be significant over time. There are several key factors worth evaluating.
• Rapid adoption of social messenger apps.
• Availability of open banking application program interfaces (APIs).
• Development of robust user authentication tools.
• Advances in machine learning technology.
Rapid adoption of social messenger apps
The use of social messenger apps on smart devices has exploded over the past few years. Time spent within such apps dwarfs that spent on other types of apps, as the richness of features and functionality continues to increase. What started as chat apps are now much, much more. In fact, usage on social messenger apps is outpacing traditional social networks, as consumers move away from public social broadcasting towards more private means of communication.
Ultimately, these apps are becoming platforms for the delivery of relevant and interactive content to, from and between users. This capitalises on an entire generation of millennials who have grown up almost exclusively on social media platforms, and are not as conditioned to using a plethora of standalone apps to consume disparate services – for example, mobile banking apps. In a way, social messenger apps have become the new operating system.
Availability of open banking APIs
One of the hurdles in democratising financial services has been the difficulty for third parties of obtaining access to consumer banking data from financial institutions. Select companies have already done some heavy lifting to gather basic account information, for example Yodlee and Plaid. However, full control of accounts via third parties is still not generally available, though we are starting to see some regulatory intervention to facilitate such innovation – notably in Europe through PSD2 (the revised Payment Services Directive).
PSD2 obligates financial institutions to provide open APIs so that third parties can more readily access and leverage consumer account information, enabling developers to build new services on top of core banking infrastructure and data. There has been a recent trickle-down effect in the US, as a lobby group was recently formed called the Consumer Financial Data Rights coalition (CFDR) to promote a similar concept of open banking.
Development of robust user authentication tools
The ability to authenticate users to access their financial accounts through third -party applications is becoming both more secure and convenient. Third parties storing sensitive financial data are now using bank-grade security. For example, this is the basic premise when a consumer links financial accounts through Mint’s personal financial management service.
Furthermore, clunky redirects from third-party apps to separate websites for financial account authentication are getting replaced by fully native in-app experiences, and usernames and passwords are being replaced by biometrics. The net result from all of these progressions is an increase in consumer trust and interest in using third-party applications to conduct financial transactions.
Advances in machine-learning technology
Machine-learning technology – the driver behind artificial intelligence – is becoming sophisticated enough to address real-world consumer use cases. In a nutshell, machine learning – and in particular, a sub-segment of machine learning called deep learning – is a set of complex, interrelated algorithms that become more accurate in generating relevant responses as more data runs through them. It is experiential, just like the human brain.
Until recently, use cases for machine learning in fintech were typically found behind the scenes, for example fraud detection. The more transactions that run through the fraud system, the better the system becomes at identifying fraudulent behaviour. But the technology has advanced, and is starting to interact directly with consumers.
Increasingly, machine-learning technology will be able to process greater levels of conversational data rather than a predefined set of variables. For example, several types of institution are starting to deploy machine-learning tools to supplement customer service on websites and mobile apps.
So what does this all mean?
As social messaging apps continue to proliferate, banking APIs become more standardised and readily available, and authentication becomes more frictionless, we should see an increasing number of financial services offered through these messaging channels via chatbots. Machine learning therefore becomes the wildcard – a potentially key differentiator that can catapult one institution ahead of the pack, whether it is a bank, fintech startup or digital giant. The big question is, how fast can this technology progress so that truly free-flow conversational financial services can be offered instantaneously and flawlessly through messaging platforms?
Based on some recent announcements, there is potential, but we still have some way to go. For example, TransferWise, a digital remittances provider, announced it now enabled consumers to send money through its chatbot on Facebook Messenger. While this is a good start, the features and functionality are limited to a predefined set of options. Several other firms – banks and non-banks – have also recently released chatbots via Facebook Messenger with a similar scale and scope.
To accelerate advances in this space, both financial and strategic investors are starting to take notice and put money to work. For example, Kasisto – a conversational artificial intelligence platform provider with a focus on financial services – recently raised a series A round led by Propel Venture Partners with participation from Mastercard and Commerce Ventures. Existing seed investors Two Sigma Ventures, DBS Bank, Partnership Fund for New York City, New York Angels and Harvard Business School Alumni Angels of New York also participated in the round.
So is the hype on chatbots justified? While it may be premature to jump on the bandwagon, it is not a stretch to view chatbots as more than a passing fad. It is just going to take time to mature as a new digital channel, the same way mobile apps themselves took time to overtake the web. Of course, primary benefactors – other than consumers – are messaging app providers themselves – Facebook, for example. Their last-mile relationships with consumers continue to strengthen as they offer an increasing number of services within their ecosystems across many aspects of our lives.
This is an edited version of an article first published on LinkedIn