“Good morning Sir. This is Siri. I have a reminder from your bank that your loan payment of $356.89 is due tomorrow. Would you like me to make the payment for you?” – “Yes, Siri.” – “Thank you, Sir, your payment has been made.”
Far-fetched? Not really. I believe that we are only 3-4 years away from this.
Artificial Intelligence and other technologies are fundamentally changing the way that consumer lenders are interacting with their customers. Banks employ huge Ops “factories”, where thousands of employees deal with various aspects of servicing the customers. But according to a recent Deloitte study, 75% of the contact center jobs are to be replaced by bots within the next 20 years¹. Let’s take a look at how this is happening, for example, within the functions of onboarding, KYC, call centers, and collection.
To begin with, banks no longer need to physically see the customer establish their identities. Various technologies have emerged in the so-called “e-KYC” (Know-Your-Customer) space to ensure easy customer identification using digital tools. Among the emerging markets, India today is at the forefront of this development with its Aadhar “India stack”. As early as 2014, banks in India started e-KYC services that enabled customers to walk into the bank with an Aadhaar national ID number and open an account only by getting his fingerprint scanned. The bank would then print out the account opening form with all the details of the customers already in it.
Other popular forms of e-KYC being deployed by lenders nowadays include “selfie banking” that leverages facial and other image recognition technology to confirm customer’s identity from a photo and independent third-party databases of IDs. For example, HSBC introduced this concept in the US market in 2016, and in Southeast Asia UnionBank in the Philippines became a pioneer, rolling out this technology in its EON account proposition in 2017. These approaches typically work best in combination with powerful anti-fraud verification tools enabled by the analysis of the customers’ digital footprint, such as those provided in SEA by CredoLab and Lenddo.
Moving on to the call center operations, when thinking of robots, most of us visualize Asimov’s android-type machines. However, in the contact center space, the robotic revolution is already here, but the Bots are entirely virtual. Their names are “IVMR” (Interactive Voice & Message Response), “SMS Gateway”, “Chatbot” and “Emailbot”. These technologies are dramatically changing the economics of the contact centre and collection industries. For example, Trueaccord in the US is now doing the full collection cycle of non-performing loans using only Bots. Closer to home, in SEA, AsiaCollect is using Bots for over 60% of its collection actions, and any human interaction that it runs is also heavily robotized through the utilization of predictive dialing and predictive analytics.
What makes Bots so attractive is not only the fact that they are cheaper but also more efficient. A lot of the contact centre work has to do with sales, and even collections calls can be thought of as a sales job, selling the customer on a repayment solution. One of the key drivers behind sales efficiency is that people tend to fall into broad psychological types, with different types responding better or worse to different types of sales arguments. Good salespeople know this and adjust their script depending on what psychological profile they are talking to.
Machines are now actually starting to do this better than people. For example, AsiaCollect is now piloting technology that recognizes the psychological profiles of the non-performing loan customers based on the words they use in a conversation with a collector. The system then suggests to the collector a change in the script and tactics based on this psychometric analysis. It won’t be long before the layering of these techniques onto the bot communication front-end will drive even the best call center salesmen and collectors out of their jobs.
Other interesting developments have to do with voice processing technology. It is an obvious fact that people respond differently to different voice types, including whether it is the opposite or same-sex. There are now solutions on the market that can make you sound in real-time like Brad Pitt or Julia Roberts at the push of a button. And these technologies are starting to be deployed by call center operators. Imagine that if the conversation type with someone calls for a certain gravitas you could make yourself sound like Sean Connery or President of the United States. Would that have an impact on your ability to convince the other side? You better believe it!
Yet another highly promising technology is Emotion Recognition. A well-known driver of success in the collection calls is the ability to keep the emotional tone of the conversation at a low enough level that the customer is kept in a comfort zone because a negotiated solution is only possible while in this zone. The emotion recognition software can analyze the tonal content of a conversation and initiate auto-shutoff if the conversation moves out of the comfort zone. Combine that with a voice recognition listening for certain blacklisted words, and you have a 100% automated collection call quality control solution, that makes sure that a collection operator will never actually pressure or threaten the customer – technology will simply not allow it.
A lot of these technologies perform best when deployed at scale. So large financial institutions are waking up to the fact that they may be able to turn some of their mid- and back-office processes into profit centers. The top example of this trend is OneConnect company owned by Ping An, the largest insurer in the world. A few years ago, Ping An spun out OneConnect to provide services to other financial institutions. OneConnect has attracted hundreds of millions of VC capital to pursue this vision and has recently announced a decision to list on Hong Kong stock exchange at a valuation of ca. USD 8 billion.
While the technology to deliver Ops solutions for financial institutions continues to evolve, one thing is for sure – within a decade, interacting with an actual human at a financial institution will most likely be considered a high luxury. Bots will get more pervasive, smart, and efficient. While some may lament the elimination of human interaction, I actually see the unit cost efficiencies that are brought by this as one of the key drivers that will make consumer finance available to all, even the poorest, parts of the world’s population. And that will be a good thing.