Bots (and articles about bots – oops!) are popping up everywhere these days. Indeed, very few concepts have shot up the hype curve so fast, especially considering that the idea of conversational UI has been with us for a long time in various forms, and under different denominations. It looks like a jungle, but all bots are essentially doing the same thing: adding conversational capabilities to FAQ and search.
You can read a good evolution and state of the art here and here. Personally, I had the pleasure of working for two companies that developed their own virtual assistant (basically a chatbot with a cute face). You may remember Anna, Ikea’s robo-assistant – just one of a fairly comprehensive chatbot database here – although all of them have somewhat faded away now without significantly transforming our lives. So, could it be different now that the worlds of bots and banking are colliding?
Perhaps. Big fish such as Facebook and Microsoft are betting high on the concept of AI-driven user interfaces, and not only for the utility of the interface itself, but because bots represent a second chance to gain a strategic foothold in the mobile space.
This focus is already heating up the development landscape as chatbot technology advances at a rapid clip. Google and Microsoft already offer digital assistants on smartphones, called Google Now and Cortana respectively, which gain increasingly deep knowledge of their users’ habits and schedules. Amazon sells a standalone device called Echo that features Alexa, who can (among other things) play music, read books aloud and help buy items through Amazon. And Siri of course reigns over the Apple universe.
According to Gartner, about 38% of American consumers have used virtual-assistant services on their smartphones recently. By the end of 2016, an estimated two thirds of consumers in developed markets will use them daily. So all signs point towards an AI-induced change in the way we interact with … well, probably everything.
Bots for banks – AI and financial services
Let’s take a step back and look at artificial intelligence from a financial services perspective. Although some interesting examples emerged from a recent Mondo hackathon, and some non-bank fintech startups are starting to introduce clever apps such as Cleo and Penny, few others signals show these reaching the retail banking mass market anytime soon. As far as I know, no bank yet offers a fully conversational interface outside their own app. But here are a few promising examples:
- Santander UK recently added the ability for some account holders to ask questions directly to their bank accounts using the Santander SmartBank App. The voice assistant service can pull transactions and process user requests to find details on a particular charge. Future versions promise to include voice-enabled payments, account alerts, stolen and lost card reporting and deeper insights into users’ spending.
- Atom Bank, the UK-based mobile-only bank, announced that it is incorporating WDS Virtual Agent software from Xeros into its mobile app. The machine learning software will give customers an agent-like option for assisted self-service on the app.
- Caixabank has implanted IBM Watson to help its business customers in the area of international commerce and expansion.
- RBS has launched Luvo to support staff in helping them answer business customer queries quicker and more easily.
- Swedbank Group, present in Sweden, Estonia, Latvia and Lithuania, is using Nina, an intelligent virtual assistant that delivers a human-like, conversational customer service experience to enable self-service capabilities and quick and easy access to information for Swedbank customers and service agents.
- Digibank by DBS probably offers the most exciting example. Its mobile app is powered by Kasisto’s AI platform, Kai, which is unique in its ability to entertain multiple channels of conversation simultaneously.
Despite these examples, the fact remains that traditional banks with an online presence tend to force customers to dig around in order to reach customer service representatives – generally via a call or live chat. Why is the current focus of these banking technologies not so much around self-serve information finding (FAQs, searches, and so on) but rather customer service? Because, despite its inherent complexity, there are a relatively finite number of customer problems and solutions for a given set of products combined with large amounts of historical data.
Doing the robot
So, you’re thinking about bringing artificial intelligence technology to your financial institution. What do you need to take into consideration?First, decide if you’re betting for a chatbot or conversational UI in your app, or both. Regardless of the previous choice, the successful uptake of a natural language interface will largely depend on the culture of the market in which you operate. Messaging may become the dominant model of interacting with the world, as Uber DevEx lead Chris Messina predicts in his clarion call to #ConvComm. However, this culture isn’t yet universal by any means, and there are of course major generational differences around messaging behaviours. Bot discovery could be an issue for generic brands trying to get user attention, but I don’t think discoverability will be a problem for a bank bot because, as a user, I will look for the service in the same way I follow companies’ Twitter accounts for customer service purposes. It will need some solid marketing, though. Third: a bot may seem a good and easy way to keep in touch with your customer base. However, prepare for some loss of control. As with any media publisher, you will be giving power to someone else’s platform. This fits a “single use” app such as Uber (which already has an in-app messenger, as well as a Facebook Messenger integrated bot), but if you’re trying to increase the quantity and quality of interactions (as banks desperately need to be doing), diverting users to another platform may not be the wisest move.
A bot also amplifies the security problem. How should it identify and verify the user before providing information or allowing for more complex actions such as money transfers? Yet, the biggest hurdle, as with most digital transformation problems facing banks, will be in core systems integration.
Function vs conversation
Interactions via natural language processing must be quick and simple – in a word, functional. From a bot perspective, one key differentiator is the capacity for banks to allow richer “mini apps” as part of their messaging experience, in which each message has the potential to become an atomic application.
This means functionality must be broken down into manageable chunks supported by services or, better said, micro-services, in the integration layers of core systems. Sadly, if you’re a banker, these micro-services are unlikely to exist in your organisation.
Second, the MVP agile approach that works so well for mobile development could bring difficulties when it comes to implementing conversational AI in banking. Here’s why: if we limit the choice of what users can do in a chat, we will need to somehow train the users or offer “menu” choices, much more obvious in a traditional interface, which reduces usability and defeats the purpose of a “conversation” in the first place. Check out this example from BI Intelligence:Some predict that 2016 will be the year of conversational commerce, but the most effective chatbots have little to do with conversation. Chatbots are not our friends, nor should we have to speak to them as such. WeChat is a good example of what functional chatbot interactions should look like:
Banks must think about the efficacy of their interface, which means making interactions functional before worrying if they’re conversational. A well-thought domain of functions and good design is key for adoption.
The bots of today, regardless of the industry, are indeed a bit of a pain, and with very limited added value for me as user in comparison to alternative UI. But IMHO, they’ll undergo exponential improvement and will definitely become part of our overall banking experience, not as an either/or choice, but almost certainly as some kind of hybrid model.
So, in spite of the hype and overload of bot-related articles (including this one), banks would do best to start thinking about not when, but how to design a user experience of a natural language interface that fits their institution, because AI technology for finserv is nearly ready for primetime.This post was co-produced together with Strands content manager Nicole Harper.
– This article is reproduced with kind permission. Some minor changes have been made to reflect BankNXT style considerations. Read more here. Main image: a-image, Shutterstock.com