This, of course, is beneficial to the bank too. In fact, a new report from Juniper Research titled “Chatbots: Banking, eCommerce, Retail & Healthcare 2018-2023,” found that chatbots could save banks up to $11 billion annually by 2023. According to the report, by end of this year, chatbots will save banks up to $6 billion.
The report also states chatbots will also save 2.5 billion customer service hours by reducing the amount of time spent on customer service inquiries, “dramatically” cutting response and interaction times via phone and social channels.
But chatbots have applications beyond customer service. Using these virtual assistants or some form of AI internally to free bank employees from simple routine tasks so that they could tackle more advance issues is also something that certain banks have started exploring. Urooj Khan, head of strategy and innovation at TD Bank Canada, told Bank Innovation that AI solutions, including chatbots- can be “incredibly helpful for employees and improve efficiency on the internal side of things.” Khan, along with his team at TD Bank, is currently researching various methods of AI applications for internal use. He could not disclose any specific project.
Another example is French bank Crédit Mutuel, which is already working with IBM to adapt artificial intelligence to improve employee efficiency. Jean-Philippe Desbiolles, VP of cognitive solutions France at IBM, previously told Bank Innovation, that the French bank plans to use AI to free employees of the task of having to process 350,000 plus emails that they receive daily. IBM’s AI-driven solution will evaluate and categorize these emails, and once that’s done, it will either be able to respond to the email on its own, and for more complex matters, the email will get forwarded to the suitable bank person.
But the Juniper report shows that 70% of chatbots will be used by the retail industry (retail banks, healthcare and ecommerce) for customer services. Nevertheless, to achieve such a wide level of adoption, chatbots have a long way to go.
Right now, most of the bank chatbots can understand and respond to only fixed set of commands, add natural language quirks and diverse accents to the mix, and that only confuses the chatbots, not to mention various novel security concerns. And Juniper’s report agrees that these chatbots have to be more sophisticated than the current lot to make such an impact. See how BofA’s Erica works right now.
In the words of the report, “Improvement in AI, as shown recently by Google’s Duplex, to create a more personable user experience would be fundamental in creating a ‘pull’ factor for chatbots.”
So let the work begin.