When a Swedbank customer has a question, they can log on to the website of the Swedish retail bank and submit their query online.

Behind the scenes, the bank’s virtual assistant Nina whirs into action, searching for the right answer, a link to a relevant web page, or potentially a follow up question to clarify exactly what they want.

It’s the first iteration in the use of artificial intelligence (AI) that will revolutionise the way that banks and other financial institutions interact with their customers.

Only a handful of financial institutions around the world are using AI to service their customers at the moment, but this will increase rapidly, says Robert Schwarz, managing director Australia and New Zealand at Nuance Communications, the software company which built the Swedbank customer service system.

“In the next five years all the very large financial institutions will have it,” he says. “The amount of interaction that’s happening between consumers and financial institutions is going up rather than going down – we need to be able to handle that.”

Along with taking online queries, AI can also handle phone queries and provide spoken answers to customers’ queries.

Swedbank introduced Nina because of the 3.6 million phone calls it received each year, about 2 million related to simple transactional issues. Nina is able to help many customers and obviate the need for them to call to speak to human agents.

“From a banking perspective, the cost to serve and the cost to sell comes down because you don’t have very expensive agents handling calls and can contain that cost with the digital system,” says Schwarz.

Machine learning to boost capability

There are two types of questions a virtual assistant can answer for a bank customer. First there are general questions, such as ‘what is the home mortgage interest rate?’ To answer these sorts of questions, the bank doesn’t need to identify the customer.

However, for questions such as ‘what is my bank balance’ the bank needs to ‘authenticate’ the customer to confirm that they are who they claim to be. Schwarz says PIN numbers and passwords are becoming redundant and will eventually be replaced with voice biometrics, such as voice recognition.

Swedbank is currently using the first and developing capability for the second type of question.

Questions Nina can answer are categorised by intent – that is by what the customer wants to do. In the early days of AI they could handle tens or perhaps hundreds of intents but the number was limited because human involvement was needed in teaching them.

However, advances in AI mean that tens of thousands will soon be possible. The system can comb through the bank’s own knowledge content base to see what sort of answers are possible and can also ‘listen in’ on human agents’ calls to understand what the right answers to some questions are.

More and more companies are springing up to help businesses take advantage of the capabilities offered by AI.

From 2011 through May 2014, over US$2 billion dollars in venture capital funds have flowed to companies building products and services based on cognitive technologies, according to a report by Deloitte. During this same period, over 100 companies merged or were acquired, some by technology giants such as Amazon, Apple, IBM, Facebook, and Google.

Automated report writing

At Chicago-based start-up Narrative Science, artificial intelligence is deployed to help automate portfolio commentary for investment firms.

The company produces enterprise software for advanced natural language generation, which ingests structured data – anything that’s numeric, symbolic, or unambiguous text – and from that creates natural language.

The company works with fund managers to help them produce personalized investment reports for the funds’ investors, which contain portfolio commentary such as a description of what happened in markets over the quarter, details of the drivers of performance or underperformance, and the important decisions taken by the portfolio management team.

“Narrative Science is actually able to help those firms completely automate that process,” say Kim Neuwirth, product manager at Narrative Science. “We ingest data from the performance attribution system such as Bloomberg or FactSet and deliver to their investment communications team the narratives that ultimately go to market.”

The system improves operational efficiency, enhances client communications and helps with compliance, she says.

Neuwirth says that along with saving time and costs for financial services companies when they produce these reports, AI will also bring benefits to the customer.

“The ability to communicate in a way that consumers ultimately respond to and understand is through personalised language. Compliance teams are also satisfied as the technology consistently meets their requirements. I think it’s going to really help revolutionize the customer relationship and engagement models that a lot of financial services firms have,” she says.
This article represents the views and opinions of the author and do not necessarily reflect the opinions of BPAY.
Published by BPAY Pty Ltd.  BPAY is offered by over 150 Financial Institutions. Contact your Financial Institution to see if it offers BPAY and to get the terms and conditions. This is general advice – before using BPAY please review the terms and conditions and consider whether BPAY is appropriate for your personal circumstances.