Even the legacy players are exploring the space. Mastercard is piloting 'Mastercard Kai,' a chatbot designed to answer questions related to accounts, purchase history, rewards and benefits. Bank of America Corp. later this year plans to debut "Erica," which will use audio and text to carry out similar basic banking functions. Royal Bank of Canada is also testing the technology developed by the company behind Mastercard Kai to roll out its own chatbot. Meanwhile, third parties such as SimplyTapp Inc. have developed payment bots that integrate to varying degrees with platforms such as Facebook’s Messenger and the latest iMessage on iOS 10.
Chatting with bots
Chatbots harness software that uses artificial intelligence (AI) to process language from interaction with humans in chat programs and virtual assistants. Audio-based bots such as Apple's Siri, Amazon's Alexa and Google Home are all now household names. Yet the bulk of chatbots today reside in social media messenging platforms, and are populating them fast: chatbots on Facebook's Messenger grew last year to 34,000 by November since the service was opened to developers in April.
"You have to be present where consumers are," says Ted Fifelski, president and cofounder of SimplyTapp, a payments startup based in Austin, Tx. "If not, you become less valuable, less useful."
Chatbots harness software that uses artificial intelligence to process language from interaction with humans in chat programs and virtual assistants.
Companies are taking note. In this 'bot boom,' chatbots might soon become a necessary way for financial institutions to do business, since people are increasingly using chat software for social interaction and completion of basic tasks such as searches and fact checking. Meanwhile, consumers struggle to use more than a handful of apps, making the prospect of even dealing with a banking app less likely. Couple that with how Messenger averaged 1 billion users monthly by last July, the latest figures available.
"They all are starting to allow interactions that are much richer and really allow users to build new experiences," says Dror Oren, vice president of product for Kasisto, based in New York.
Bots can also make payments: PayPal's bot debuted on the platform last September. Kasisto, the company behind Mastercard's Kai chatbot, also has its own product available in messaging platforms to transact and make basic banking queries. Other messaging platforms are incorporating bots, including Skype and Slack.
Kasisto's Kai banking product can answer banking queries, but it can also predict needs and guide customers through banking conversations. It uses the same AI engine Kasisto is supplying to power virtual assistants in mobile banking apps.
A double-edged sword
For banks, it's not like the concept of communicating with customers via messaging is new, as many have used SMS texting as a format. Think of chatbots as an extension of text services such as text banking, Oren says. "It's another cycle of previous technologies that have existed; they just get better."
For banks and any business accepting payments, bots will be a big money saver. Banks and credit card portfolio owners are still saddled with people preferring to pick up the phone to pay a bill or make a loan payment. These calls cost organizations about $6 each, estimates Rurik Bradbury, global head of research at LivePerson, a New York-based provider of mobile and online messaging solutions. "Cost is a big driver here," Bradbury says. "You can cut a majority of that cost by doing messaging instead of voice calls."
For instance, T-Mobile, a client of LivePerson's, saved 48 percent in customer care costs by shifting customers away from customer-hostile voice calls to messaging and bots, according to LivePerson. "Bots can help with these efficiencies where there is an obvious, simple, repetitive process and give that time back to the agent," Bradbury adds. "The consumer won't have a worse experience because of it."
Regardless, when it comes to payments, providers are jumping on board.
In theory, most consumers have experienced the frustration of a chatbot or SMS messaging failing to answer a question. In one example, a user asked a popular virtual assistant on a smartphone, "Give me a list of Chase bank branches in Santa Monica." The assistant yielded the following response: "I can help you find a place if you turn on location services." The assistant failed to even supply a list on the smartphone's browser.
So financial institutions should beware of heightening frustration, Bradbury says. "The big mistake with chatbots has been imagining they could suddenly take the entirety of a conversation and just work by magic," Bradbury says. "If you go 'all bot' too soon, you'll shoot yourself in the foot. Yes, you'll save a lot of money but have a lot of unhappy customers."
From humans to bots to humans
LivePerson, which was a pioneer in web chats for companies, is taking an approach where it has humans behind the interaction, then builds in bot capabilities as time goes on. It's currently doing this with T-Mobile and some of its banking customers. Yet instead of having a chat session that is open and closed, LivePerson uses a 'persistent connection' where the customer and agent can converse continuously over hours, days or weeks, similar to texting a friend on your own time.
"Companies are trying to think about chatbots first, before they actually have this persistent messaging connection with their customers," Bradbury says. "Instead, you need to start messaging with customers, powered by humans, and then over time, you need to feed in the bot elements."
Regardless, when it comes to payments, providers are jumping on board. Last year, SimplyTapp introduced its Gane app, which lets users pay via near-field communication (NFC) in store to take advantage of merchant discounts and send and receive money.
What's next? Ask a bot and see what happens.