Paying via Chatbots


Nowadays, we use a great many web services and apps to make our lives easier. And yet, we want things to be even easier still. Chatbots could soon be helping us to shop and access news and other services. This scenario presents both an opportunity and a challenge for the payment industry.

It feels as though each day brings a new trend and each week a fresh revolution; one which looks set to change everything. If we put aside the artificial hype and our fascination with technology for a moment, however, we realise that real revolutions are actually much less frequent. The last change to genuinely affect multitudes of people was probably the launch of the smartphone, with its apps and mobile Internet access. Sometimes, however, today’s smartphones are still too complicated, with users having to switch between a variety of apps and web services in an attempt to optimise their everyday lives. It is here that chatbots have the potential to bring us a great leap forward.

Let’s consider just how complicated it is to organise something using a smartphone. Say, for example, that you’re interested in going to a concert with a friend. First, you decide on an artist or genre via email or a messaging program. Next, you use your browser to search for concert dates. If there’s nothing in your hometown, you go to Google Maps to find out how far it is to an event in a different city. Once you’ve collected this information, you go back to your friend to get their input. After you’ve chosen the date and location, you go to the booking website. Here, of course, a variety of problems can arise, including sold-out dates or poor seating choices. By now, you’ve probably spent a great deal of time and used a wide variety of apps and services, and you still don’t have concert tickets. Then, of course, you must continue your task via a specialist booking or payment app—all this to achieve a single aim.

Chatbots can dramatically streamline this type of process.

Doing it All via Messaging Apps

Chatbots hang out exactly where users do: in messaging apps. You can use WhatsApp or a similar program to send them messages, just as you would to real people, and the chatbot will try its best to react like one. You no longer need to leave your messaging app to book a ticket. This is precisely the kind of convenience today’s users are looking for. The most popular chat apps, like WhatsApp and Facebook Messenger, are installed on almost every smartphone. Skype, too, is hugely popular, while younger users communicate using Snapchat (1). The trend is moving away from numerous apps to just a handful, in which users are spending increasing amounts of time. WhatsApp alone is used by more than a billion people (2), while Facebook Messenger has around 800 million users (3). If, therefore, people are already hanging out in messaging apps with their friends, they can also use them to do other things. Instead of switching to a different app, you send a message to a music bot which helps you find artists to suit your collective music tastes. Then it’s simply a matter of sending a query about concert dates near you and available seating, to which another chatbot will provide appropriate responses. You don’t even have to leave the messaging app to book your tickets. Instead, you can use the chatbot to book—and of course, pay for—them directly.

This is precisely the idea behind chatbots, which have been available for various messaging platforms for several months now. For industry giants, like Facebook, Google and Microsoft, chatbots are currently among the hottest topics, and many Internet retailers are champing at the bit to gain access to messaging apps’ huge target audiences.

Chatbots Today

The super-simple concert booking, however, does not yet function as described. Chatbots are still in their technological infancy. Currently, for example, Facebook Messenger contains chatbots which can, when provided with a location, show current information on the weather, or order pizza via text. There are also chatbots which function as interactive gaming partners or language trainers. These chatbots should never be confused with the botnets which have been keeping IT security experts on their toes for years. Chatbots are also known, somewhat clumsily, as “dialog programs”; the software comprises a simple user interface which can receive text and language input and provide answers. To do this, it integrates itself into messaging programs just like a human contact. In order to speed up chatbot development, companies like Facebook are providing appropriate software modules and interfaces. Not every chatbot needs to have artificial intelligence, but it must at least be able to recognise and react to typical dialogue. The hard part is creating chatbots which are self-learning and capable of retaining current context. This means that someone looking for information about an artist, then enquiring about concert dates, shouldn’t need to mention the artist’s name twice. Just like a human interlocutor, the chatbot should remember that its chat partner is still talking about the same artist.

Payment via Chatbots

The more intelligent chatbots become and the more services that can be covered by messaging apps, the more important the subject of payment becomes. Various models are conceivable: users can, for example, simply store payment data in the messaging platform and then pay with any current payment method, like a credit card or PayPal. This means that users need only share their payment information with a single provider like Facebook, Google, or Microsoft—just as they do now with app stores. This method is already in use in the USA, where Uber, the popular ride-sharing service, is fully integrated into Facebook Messenger, complete with payment functions (4). Chatbots themselves could also be designed to receive payments. In order to do this, the platform operators would have to build a kind of chatbot app store to enable users to find appropriate chatbots in the first place. These chatbots could—similarly to apps—be at least partially fee-based, or use a kind of in-app purchasing model. Payments could then be processed centrally by the platform operator using payment data. It would also be possible for chatbots to replace retailer-specific apps, and for these chatbots themselves to contain payment functions. After all, who needs yet another shopping app when you can do everything via your messaging program? Almost no-one. Using chatbots would eliminate the need for expensive and complicated app development. Instead, retailers would simply use their chatbot to approach their customers and offer them a convenient way to pay. This concept of “conversational commerce” already exists (5); after all, chatbots are designed to facilitate direct communication between customers and brands, shops and retailers. The requirements of payment services are, therefore, very similar to those of payment apps. They must be as easy to use as possible, with one-click models being the most desirable. They could also use integrated fingerprint or other biometric characteristics.

Simple P2P Payment

For a long time now, the payment industry has been discussing making peer-to-peer payments via smartphone, and there are already a variety of apps which do this very elegantly. There is, however, no standard app which facilitates direct payments via smartphone. Why, though, would people even need a separate app when they already use a messaging app to keep in touch with their friends? A chatbot could take over this function and, with a simple “Pay Tim 23 euros”, a restaurant bill could be split.

Conclusion: Exciting and Dangerous

Chatbots may well be the next big thing, primarily because they can be developed very easily and cheaply. Although chatbots are still in their infancy, the major technology companies are amassing all kinds of tools for equipping them with artificial intelligence. Soon, therefore, users may no longer be able to tell whether they are communicating with a real person or with a chatbot. Retailers will need to rethink their strategies if their customers begin avoiding store-specific apps or mobile websites and hanging out almost exclusively in messaging applications. There is, however, no such thing as an ideal world when it comes to chatbots. Although communication with virtual assistants may sound simple, it must be secure if it is to handle personal data. Just how effectively chatbots can fool people was demonstrated in the Ashley Madison case: instead of chatting with interested women, the cheating portal’s paying customers were conversing with chatbots (6).