Facebook Inc. enabled payments via so-called chatbots back in 2016, and now the massive social network is acting to standardize and streamline consumers’ use of the digital gadgets on sellers’ pages.
The United States Patent And Trademark Office published on June 28 an application from Facebook for a method for “processing payment transactions using artificial intelligence messaging services.” With the application, which Facebook filed Dec. 22, 2016, the Menlo Park, Calif.-based social network describes a system in which a Facebook Messenger user can pay merchants using a card on file by typing instructions on his or her device as part of a conversation within the Messenger app.
A key part of the method is that users can start the payment process and receive confirmation without leaving the app, according to the application with the USPTO. “One or more embodiments initiate a payment transaction on behalf of the user based on a natural language conversation and without redirecting the user away from the communications session by typing instructions on his or her device as part of a conversation within the Messenger app,” the application says.
With natural-language processing, bits of code known as chatbots or simply bots can carry on a conversation with a consumer using a more or less normal or “natural” vocabulary and syntax.
Facebook’s application outlines a method that offers several advantages to merchants and bot developers, observers say. For one thing, it presents a single platform “for messenger-bot developers to hook into,” notes Steffen Sorrell, a principal analyst at Juniper Research, a United Kingdom-based firm that follows bots and other forms of artificial intelligence in payments.
But beyond that, it eases the workload for developers in other ways, as well. “It provides context analysis, tokenization, and appears to handle the back-end payment process, so in essence will simplify the development process for merchants/bot developers,” Sorrell says in an email message. “Additionally, it is integrated with the social graph, so will be able to handle product suggestions automatically.”
The “social graph” refers to a map of connections from one Facebook user to others in the user’s extended network. The more such connections the better, as bots “learn” more swiftly with more interaction, experts say.
Bots supporting so-called conversational commerce have found their way onto social networks ranging from Kik and Slack to Viber and WhatsApp (which is part of Facebook). But it’s on Facebook, and particularly Facebook’s Messenger app, where the technology has made the most progress.
Messaging apps appeal to bot deployers not only because of their conversational character but also because of their rapid growth. In 2016, there were 1.61 billion users of messaging apps worldwide, or 62% of all mobile-Web users. By the end of next year, those numbers will swell to nearly 2.2 billion users, or 68% of all those on the mobile Web, according to eMarketer.
Specific details about Facebook’s plan for the methods outlined in its patent application remain unclear. Without commenting further, a company spokesman cautions against reading too much into the filing. “We often seek patents for technology we never implement, and patent applications should not be taken as an indication of future plans,” he tells Digital Transactions News via email. If anything, it seems likely development work is in its very early stages, according to Sorrell. “Looking at the developer documentation, it appears as though a) payments are still in beta, and b) the platform as described in the patent is not yet fully fleshed out in the live product,” he says.