Nationwide Building Society has used its experienced frontline staff to train its digital agent, known as Arti, to answer coronavirus-related questions.
The ramifications of the pandemic has created a spike in customer queries. Covid-19 has hit peoples’ livelihoods and increased demand for information from providers of mortgages and other financial products.
For example, millions of people have applied to have mortgage payments and interest deferred for up to three months, when their earnings were hit as the UK went into lockdown.
Jag Bassi, head of digital propositions at Nationwide Building Society, said the challenge faced by Nationwide during the Covid-19 pandemic is no different to that experienced by all financial services institutions.
“As a large proportion of the population struggle and find themselves in a difficult financial circumstances, to be able to provide payment holidays across all our product ranges became an absolute priority,” he said.
Demand for support came in a sudden spike and there was a major operational challenge. “We had to support our call centres because members couldn’t go into branches, so Arti was developed to understand many questions around things like payment holidays,” said Bassi.
Nationwide already had Arti, which uses IBM Watson technology to help it understand what customers are looking for, but had to adapt it to support customers with specific questions about things related to Covid-19, such as how to apply for mortgage payment holidays.
The building society used its experienced frontline staff in the development to ensure Arti provided the level of support expected. “Dialogue design and chat bot development is really a new and nascent capability,” said Bassi. “The approach we took was not to search the market for people that had chat bot expertise, but instead we took our frontline staff and gave them the training they needed to do dialogue design.
“Arti’s conversations are a reflection of that expertise that we already have inside our organisation,” he said.
People like former mortgage advisers are shaping what customer queries might be and helping Arti understand the type of language that members would use, and to provide answers back in a simple, concise and easy to understand way. “With the right people in the room we can complete an entire dialogue journey in a few days.”
Bassi said the service is probably a “tiny bit slower” than doing a web page, but “hugely more interactive, targeted and specific”. Arti has had 10,000 conversations about mortgage holidays alone since March 27, via online chat.
In the long term, Arti will be used across the business, connect with digital services and create personalise support for each customer. “We have simple use cases and we will move onto more complex use cases as we go along,” said Bassi.
Beverley Bartlett, senior digital manager at Nationwide, said Training Arti is just one part of a bigger picture. “With an 89% increase in new registrations for Nationwide’s online bank since lockdown began, it’s important we provide our members with more choice to engage with us digitally, alongside our branch network and contact centres, should they need them.
“The speed at which we delivered this new capability meant that frontline teams could spend their time dealing with more complex queries from our members,” said Bartlett. “It means we can continue to serve our members in more ways, more quickly and effectively, but still with the added reassurance of the human touch, if needed, that our members value.”
Investment paying off
Investment in artificial intelligence (AI) in the financial services sector is paying off as organisations quickly automate customer services to cope with huge spikes in demand.
Businesses are using automation technology investments to quickly scale up capacity to cope with this spike in demand, without having to recruit or outsource. These technology projects will save money and improve customer services in the short term and will become useful proofs of concept in the wider rollout of automation technology in customer services operations.
In the past if banks received sudden surges in demand for customer support they could increase the number of contact agents they have. This could be done through recruitment or increasing the number of staff they use at outsourcing service providers. But automation technology has transformed how business react, with Covid-19 challenges bringing this change to the fore.
For example Leeds Building Society, already using repetitive process automation software, increased the number of software robots in its mortgage processing operation to cope with high demand for mortgage holidays.
The building Society said requests for mortgage deferrals had suddenly increased to more than 2,000 a day, overloading call centres. After the building society created a web form using, calls to the call centre reduced by 75%, with answers given in 21 seconds.
Similar success was reported by high street bank TSB. It developed a smart digital agent to serve customers in the absence of bank branch support and overloaded call centres during pandemic. Within five days of the agent going live on 25 March, through a virtual assistant, it had answered more than 40,000 customer requests.
According to a recent study by EY, more than one-third of businesses accelerated their automation strategies when the extent of the disruption caused by the Covid-19 pandemic became clear.
It found that 36% of businesses are accelerating investment in automation, and 41% are re-evaluating their automation plans.
Customers also appear open to automation technology, even when it comes to their finances. In the past, one of the concerns holding companies back from moving to automation was fear in the customer base, but attitudes are changing. According to a survey by Pinsent Masons and Innovate Finance last year, nearly two-thirds (63%) of people are happy to engage with a chatbot.
The research also found that 35% of people prefer an instant response from AI than a delayed reply from a person.