Skip to main content [aditude-amp id="stickyleaderboard" targeting='{"env":"staging","page_type":"article","post_id":2053435,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,commerce,enterprise,entrepreneur,security,","session":"C"}']
Guest

Why banks need chatbots (eventually)

Bank

Image Credit: Shutterstock

I’ve wanted to have a purposeful business bot for years now.

We at Tochka hesitated to launch the world’s first Facebook banking bot — with full payment functionality — for almost two years. The bot itself is only a moderate success, but it has certainly been a useful tool to explore how customers behave with such a toy. We have been trying to answer questions about whether a bank might need a bot in the first place, what platform it should run on, whether artificial intelligence would be useful, and whether our bot should have a character of its own.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":2053435,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,commerce,enterprise,entrepreneur,security,","session":"C"}']

From our work on these questions, I would say that for a bot to be as useful as possible, it should have a personality and it should run on the platform that is the most popular with clients. But A.I. might be too early in its development to work here.

So, why would a bank need a payment bot in the first place? I first started thinking about a chat-enabled interface when I was receiving more chat dialogues than emails every day. No need to count the number of messages I received – I think they were running into the thousands.

AI Weekly

The must-read newsletter for AI and Big Data industry written by Khari Johnson, Kyle Wiggers, and Seth Colaner.

Included with VentureBeat Insider and VentureBeat VIP memberships.

Today, dialogues and group chats in LinkedIn, Facebook, Telegram, and others are becoming standard practice.

Another area that got us thinking about bot applications was the evolution of the personal business assistant. Our typical client, an entrepreneur, might not use a web interface or a mobile app. If they wanted to find out their bank balance, check on incoming payments, or give orders to wire funds, their pattern might be to text their business assistant or an accountant. So, we reasoned, if they can chat with an assistant, why can’t they chat with a bot?

Of course, the platform is an issue. I don’t think that a separate app just for a chat with a banking service, or any service, for that matter, would work. People appear to be fed up with apps and, for most users, it will be extremely difficult to find a place for another one on their default screen.

Telegram had a great API, but when Facebook came up with a new API and authorization solution, it pretty much set the scene. Most of our clients have a Facebook Messenger icon on their default screen, and many have liked our business page or talked to us on our Facebook account for support or feedback. So, I voted for Facebook as a starting point to play with bots.

Bots should be useful. In our case, the bot had to emulate the basic tasks a business assistant would perform, that is giving advice, directing someone to a nearby ATM, telling them their balance, and helping them make a payment.

Bots should have a personality, as well. We learned a lot from Microsoft’s experience in this area. If you remember, Microsoft had to pull its Tay chatbot from Twitter after human users managed to train the innocent A.I. into become a swearing racist xenophobe in less than 24 hours from launch.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":2053435,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,commerce,enterprise,entrepreneur,security,","session":"C"}']

We definitely did not want our bot to follow in Tay’s steps and learn to yell at customers like an annoyed teller. I thought about how I would not want my mom to talk to a bot equipped with artificial intelligence, even if it was a lot like Siri. Artificial stupidity instead of artificial intelligence annoys me more often than it helps me, and we could all do with less annoyance in the world. It’s better to give R&D departments a couple more years to bring A.I. to the level my mom would be able to deal with. However, embedding banking commands in Siri and the like might be a great development in the near future.

Tochka’s bot has a mild personality. It is not a machine pretending to be human but rather a machine pretending to be helpful. For better or worse, Tochka’s bot is not personalized and cannot tell jokes yet, like “Why do bankers make good cyclists? Because they are good at keeping their balance.”

I believe that the future will go far beyond what we did at Tochka. New developments will spring from A.I. and from building into existing ecosystems, be they Facebook chats or Siri.

It might be cool to have bots involved in a Facebook chat between two business partners, making sure that one is sending the money and that the other one gets it. A bot might remind a user about a past due invoice. A bot might one day be like an attorney or a banker who are invited into the room to seal the deal. We shall see.

[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":2053435,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,commerce,enterprise,entrepreneur,security,","session":"C"}']

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn More