This post originally appeared on the Finn AI blog, which is now part of Glia.
Successful AI applications have enhanced many areas of business from lending to fraud detection, anti-money laundering efforts, and cybersecurity. Today, banks are putting their money where their mouth is and integrating conversational AI banking to many of their frontline services.
As banks take this next step, they’re weighing up the risks and returns to develop thoughtful AI strategies for these critical business use cases. AI-powered chatbots can enhance the banking customer experience and reduce the cost of customer service—they can also make mistakes and offend people. Virtual financial advisors can help people become more financially literate—they can also develop unintended biases that exclude people from services based on incorrect assumptions. AI can help frontline staff spend more time on valuable activities and less time on repetitive tasks—but if implemented poorly, it can create uncertainty and contribute to loss of productivity and employee churn.
Implementing AI into frontline services is a daunting task…but it doesn’t have to be. Here are five steps you can take to prepare your bank for customer-facing AI applications:
1. Develop an API readiness / health strategy
API readiness is key to developing a personalized experience for the end user. Think about the APIs you already have in place for existing digital initiatives. How can you leverage these within your AI strategy to make your bots more personalized or intelligent?
Authentication is key here—your APIs must correctly permit or restrict access to your bot. By using a partner like Auth0, you can give your bot appropriate access to your existing APIs on behalf of the application itself. You can easily support different data flows in your own APIs without worrying about the many technical aspects of API authorization.
2. Align with your existing channel strategy
Think about the channels where you’re having meaningful conversations with customers already. Do you use Facebook for customer care? Do you have a sophisticated digital strategy with a stack of native apps, websites, and social platforms that you already use to communicate with customers? Weigh up the benefits and drawbacks of each channel to help you decide the best fit for your first foray into customer-facing conversational AI banking.
Once you consider your key channels, you can make more informed decisions about where to implement your chatbot. Try not to tackle every channel at once. Start with one, scale from there, and learn as you go. The power of AI comes from having that one central brain that can deliver different experiences in different channels—it will get smarter as you scale.
3. Get your customer support house in order
Customer care is often the first customer-facing application for AI banking, with many experts estimating a reduction in cost of anywhere from 30 to 70 percent.
Unfortunately, many bots have a less-than-ideal workflow in the customer support use case. For example, when the banking chatbot finds itself in unfamiliar territory and is unable to answer a question, the customer is often passed to a human via a phone call.
To prevent this clunky experience, consider incorporating customer chat within your banking call center so your teams are familiar with how it works before you launch AI. That way, when the chatbot is in place and the chatbot comes to a dead end, the customer query can be passed to a human via chat, delivering a seamless workflow and a better experience to the end user.
4. Involve frontline staff in the AI implementation
AI has a mixed reputation among frontline staff, with some workers excited to learn about the new technology and others concerned for their job security. But based on what we’ve seen, chatbots often free up time so people can work on more high-value tasks, which is a win-win for everyone.
Since your customer-facing staff could be the difference between a successful AI implementation and a PR nightmare, you should initiate an open communication strategy prior to launch. This will ensure they are informed and excited about your banking chatbot initiative from very early in the process.
5. Know your use case
Tackle AI one use case at a time. Know the problems you want to solve with AI and plan a logical, phased approach. For example, do you want to leverage your bot to drive product acquisition, to deflect calls from your contact center, to deliver a financial assistant to your clients, or to provide financial literature to make banking accessible to everyone? Prioritize what’s most important to your business and map out an AI timeline to help you achieve your AI goals.
Wherever you are on the AI journey, there is a long road ahead—even organizations at the cutting-edge of innovation have yet to scratch the surface on the full potential of conversational AI. It’s certainly the most exciting technology track to watch in banking as we move into 2019 and beyond.
Interested in learning more about conversational AI banking? Request a demonstration today.