KNect365 is part of the Knowledge and Networking Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.


The rise of the machines: AI in customer service

Banks need to pay attention to artificial intelligence as customer service becomes increasingly automated.

Bianca Ho, co-founder of Clare.AI, which builds AI digital assistance for financial institutions, told FundForum Asia that financial institutions needed to embrace AI.

She said: “On a market trends level, Gartner says 85% of customer support will be carried out by machines in 2020. That is a pretty bold statement.

“Banks are a huge component of that customer support segment. Banks receive a lot of customer enquiries everyday.”

Ho said there was also a cultural change, with 56% of people aged between 18 and 34 saying they preferred using live chat to the telephone.

“I don’t like people calling me on the phone. I don’t want to go to a bank to get things done. I do prefer using a live chat if that can solve something straightway,” she said.

Around 35% of all banking interactions are now carried out through mobile devices, the highest proportion for all channels, including online banking and traditional channels, such as ATMs and branch visits.

Ho said: “If you look at how banks are servicing their clients, it started 150 years ago with a bricks and mortar experience, then it was telephone, then when the web became more prevalent it was online banking and mobile banking.

“Now there is another channel, which enables customers to use a chatting interface to talk to their bank.”

She said enabling a new channel represented a lot of challenges for banks, but banks were willing to embrace it as chat was the preferred interface for many people.

“There is also an enhancement of customer experience,” she said.

“If you can do a customer experience at scale with your mass retail customers the way that you can do it with ultra high net worth, you are getting a lot more data and you can personalise that data to your users.”

She said banks were also likely to be able to make considerable cost savings by using chatbots, with potential annual savings in the US running to US$23 billion a year, out of a total annual salary expenditure of US$79 billion.

“With AI there will be people who need to change their job responsibility, but there will be savings that will be harnessed through that,” she said.

Introducing AI

Ho said it was important for a bank to define its strategy and goals when introducing AI.

She said: “You need to define your objective and think why you want it and how you want automation done in your company.

“AI will only work with a lot of data, so you need to give up a lot of data to build something relevant.”

She said it was also important to select the scope of an AI project, with customer support often a good place to start, as its impact was measurable.

Finally she said companies should be prepared to get their hands dirty.

“You need to understand the challenges of building it. You don’t really understand AI until you have used it in your own context,” she said.

Ho said chatbots could help to solve some of the problems banks faced.

She said customers often complained that banks did not know what they needed, and they did not proactively provide them with the information they wanted.

They also said banks offered them irrelevant product promotions.

By contrast, banks said they did not know exactly what customers wanted, and they could not give them predictive information.

She said: “Without data banks cannot create the personalised product that customers want, that is the disconnect we see.”

She said chatbots could be used initially to raise awareness of marketing campaigns. They could then nurture the customer relationship by providing relevant product information.

If a customer made a purchase, they could act as a sales assistant, helping with customer onboarding, before taking on a support role, answering customer questions and checking their exposure and holdings.

They could then develop an advocacy role, introducing relevant investment ideas or outlook information based on a customer survey, which in turn could raise customer awareness.

Realistic expectations

But Ho warned that the hype surrounding AI had created unrealistic expectations of what AI could do.

“It is our responsibility to educate financial institutions and our partners,” she said.

She said the group usually told banks to treat introducing a chatbot like hiring a new person, as it would take three to six months for it to get up to speed.

“It is like a baby. You need to feed it the right things, and teach it. AI is similar,” she said.

Ho believes banks can use AI to help rebuild trust.

“The trust in financial institutions has been broken, the online banking experience has not been the best. We want to help repair that trust,” she said.

Get articles like this by email