Pascal Oparada
Zenith Bank’s AI-enable ZIVA chatbot, launched recently, underscores Nigerian banks’ race to adopt AI technology to drive growth in an industry playing catch up with the global competition.
The bank was the first Nigerian company to have a functional website in 1995, it then smoothly migrated its operations from analogue to digital, and has since remained at the forefront of using tech to provide first-class customer service.
It is not surprising, therefore, that the bank is also tagging along in the move to drive growth and digital expansion.
Like Zenith Bank, which actually came late to the party, many banks in Nigerian banks are racing to make Artificial Intelligence (AI) a big part of their strategies. They have been wary all along but now it seems they are diving in headlong.
Tellers and customer relations are being replaced by chatbots, pieces of software that can have a conversation with a person in real-time as banks seek to decongest the banking halls.
UBA’s Leo, introduced in 2018, uses a harvest of customers’ data gleaned from Facebook Messenger to interact with customers.
Interacting with Leo was a mixed experience; test of the technology has shown. This is in contrast to Keystone’s gender-neutral chatbot which gets stuck at some point in the conversation.
Customers in Nigeria are distrustful or rather cautious of allowing chatbots to relate with them. They still prefer the conventional and traditional method of walking in, face a customer assistant whom they can look in the face, and intelligently engage and have them solve their problems.
“The technology is quite new but it is going to have quick adoption in Nigeria if it is made as simple as possible,” says Chibuike Goodnews, a technology analyst and Co-founder, Dochase ADX.
According to him, the technology is targeted at the millennials who spend so much time on their phones and desire quick feedback on their banking relationships.
In the next five years, most chatbot users would rely on the messaging app for commerce, technology analysts predict as Facebook spearheads the big data, an app that reveals human behaviour and interests.
“AI relies heavily on social data. And social data that is being released on social media is not enough to offer full banking services and there are certain things that you would ask and they would not be able to harvest on social data”, Goodnews said.
French bank, Societe Generale, is developing chatbots that could answer questions about equity funds in its Romanian banking unit.
TransferWise, a money transfer startup released a chatbot that allows customers to send money to friends and family internationally via Facebook messenger.
But it seems Chatbots in Nigeria are only limited to simple banking like account opening, account balance and simple inquiries.
Aside from the issue of simple banking, Goodnews believes that Chatbots should be able to both wow and inform users which would keep them coming back and make them repeated user of the bots.
“If those things are integrated, it should be able to attract and also wow them with some of those features.
“For example, somebody’s name is Jane and the bot should be able to say, ‘Hi Jane, your birthday is on this date and then say happy birthday and the person would be like wow we don’t know you are doing this kind of thing”, Goodnews said.
Dodgy with data
But it has been problematic because many of the customers do not use their real identities on social media.
“Identity issues have been a major problem in the adoption of AI. An example is the TrueCaller where people give different names.
“The flip side of it is that the banks can give users opportunity to update their details with the right information and those information have to sync with what they have in their accounts or else there would still be a problem”, Goodnews stated.
Security challenge
Part of the reason users are cautious is the potential threat of compromising their data. They want to be assured that their data are safe.
“Until when users are able to come out clean and say this is my data, AI will still be a major problem in Nigeria and the fact that the data is being picked from the bank into social media kind of poses a risk.
“But they have to make users understand that it is social data they are releasing and not any security data they have in order to give users the confidence that AI can merge social and security data and not divulge the information”, Goodnews said.
The level of sophistication that has been put into AI is very big and it would take some time before it becomes saturated.
But suffice it to say it is a right step in the right direction and they should just keep improving on it and with time they will get better. Other technologies started that way but they keep improving until they get to where they are today.