Most banks have dabbled with natural language in some form. A simple virtual assistant on a website or a mobile app that allows basic voice commands, but these are just point solutions to specific problems that don’t explore the value of human/machine conversation to the wider business.

In a recent article Banks, In Uncharted Territories, Francisco González of BBVA points to the explosion of technology in the financial services market that has the potential to destabilise unless greater action is taken. He notes that the tech giants are staying on the fringes for the moment, but I would say that they could soon become your biggest competitor and to wait until they start encroaching further would be a mistake. The mobile network operators did the same and look how well that worked out for them, marginalized to the sidelines of the least profitable side of the business.

The banking wannabes have several advantages. Their systems aren’t steeped in over forty years of computing history. The regulators are still trying to determine just what does and doesn’t come under their jurisdiction. And the big names – Apple, Google, Facebook etc, they come with a liberal dusting of trust and product familiarity that puts them miles ahead of newcomers. They also have one thing in common. In their different ways they are all trying to create an ecosystem where they own the customer.

It’s time for the banks to fight back.

Conversation may seem an odd place to start, but if your business is to be driven by knowledge that enables you to predict and anticipate your customers’ individual needs, then it’s time to get personal again. Years ago a simple chat with the cashier about the weather and holidays may have resulted in a travel insurance sale or a new savings account to pay for the flights. Pushing transactions online streamlined processes, but it also reduced the customer interaction to a series of clicks.

It’s a strange thing, but despite knowing that they are talking to a machine, when a conversation is human-like people can’t help but reveal all kinds of information. Even when the medium is text, not voice. Information that can be used in real-time to resolve the initial reason for the interaction using data from back-end systems; expand on any potential cross sale indicators; and engage the customer into another conversation using additional questions, triggered by the original conversation, prior behaviour and even data acquired from a third party.

This massive amount of raw, unstructured conversational data can then extrapolated and analysed alongside context, metadata and other criteria to deliver actionable data to the entire business, from customer experience and product marketing to the board itself. Once you’re a part of the conversation, it’s easier to bring in trusted third party partners. An NLI interface doesn’t have to mean opening up your own systems. It’s perfectly feasible that your application could retrieve the information from a third party or that the interaction is handed off to the partner and then returned back to the original application.

Can you avoid having a bot on Facebook? Eventually probably not, and if your customer is there why would you want to? The important concept is to ensure your conversational applications easily port to whatever comes along in the future, regardless of how ancient your back-end systems are. Build once to deploy on multiple services and devices, in any language, should be the key goal.

In a recent survey of over one thousand respondents from around the globe, nearly 30% expressed a desire to use their voice for banking activities. Your customers are out there ready to talk. You just need to take charge of the conversation before someone else does.