How fintech companies can adopt AI without losing their human touch

It’s not entirely surprising to see the fintech industry embrace artificial intelligence with the fervor that it has for the last several years. Adopting the latest technology is part and parcel of what the most innovative fintech companies have been doing since day one – in fact, it’s one of the characteristics that has most distinguished these companies from the established financial institutions that long dominated the banking space. Fintech companies are early adopters by default and tend to be true believers once that tech hits critical mass, as we’re starting to see with AI.

Fintech’s embrace of AI is, at its base, purely practical. There is no shortage of benefits and advantages AI brings to fintech, beginning with the most basic elimination of human error, which is nothing short of essential when dealing with people’s money. By now AI integration within fintech processes has already progressed well beyond just catching errors, and it’s changing everything. We’re seeing the proliferation of chatbots, virtual assistants, machine learning algorithms, and natural language processing (NLP) across the industry, directed toward a range of processes: fraud detection, underwriting, loan approvals, algorithmic trading, market forecasting, and customer-facing interactions such as onboarding new users and supporting existing ones.

When seeking to find the right balance between AI and human power, it’s those directly customer-facing processes where things tend to get complicated. AI can bolster customer loyalty through functionality – approving a loan and distributing funds in a matter of minutes, for instance – and it can just as easily erode it by making customers feel like a fintech company doesn’t care about them and their business enough to devote anything beyond a standard chatbot to addressing their issues.

So where is the line between AI as a boon to customers and AI as a potential pitfall? Returning to the days of call centers and exclusively human-on-human service interactions is out of the question. It’s a waste of everyone’s time, especially the customers who just want a quick answer to a quick question or a workable solution to a minor problem. Ditching chatbots and automated service portals would also completely undo the initial benefit of freeing up customer service representatives for more complex and high-value interactions that require dedicated personal attention and analysis.

That being said (and even in an era like ours where passing the Turing test is child’s play for most AI chatbots), there’s still no replacing a human element and dedicated personal attention. Companies with the best of intentions can turn off customers if every interaction seems to be with a chatbot or if getting to a real person seems even marginally cumbersome. Particularly when you are dealing with people’s money and businesses, they want to feel as though they are not just another anonymous customer at the tail end of some data-accumulating pipeline. It’s one of the best ways to lose loyal customers, regardless of how useful they find your product.

Sidestepping this begins with recognizing that one size chatbot does not fit all. Chatbots need to be designed to respond to the problems and issues faced by your specific customers using your specific product. A fintech focused specifically on FICO scores can and should have a different basic pipeline than a lending platform that naturally uses credit checks as just one part of the lending process. Chatbots should also be extensively tested, molded to fit the brand voice, and able to recognize and access returning customers’ individual data seamlessly. Chatbots should make it clear that customers can get in touch with a real person, either by forwarding them to a representative or providing a manned email, for instance. Keep in mind, however, that if your chatbot works and gets people the answers they need, they’re much less likely to feel disillusioned over not speaking to a real person in the first place.

Chatbots tend to take up most of the oxygen in the customer service and AI conversation, but it goes beyond that too, particularly for fintech companies focused on lending. Alt-finance lending companies have broadened the idea of applicable credit and collateral in recent years, accepting other factors and criteria besides credit strength when evaluating applications. As this process evolves, it needs to be driven by humans with real-world lending experience. AI can’t anticipate everything and can’t experiment with new factors on its own, even when it has an enormous amount of data to learn from. Entirely new scenarios in the market and new forms of financial strength and viability are popping up all the time, including ones that run counter to established market logic. Real people need to be on hand to address that.

This same logic applies to borrowers on a much more individual level. When evaluating a loan application, can an AI tell if the business in question is digging themselves out of a hole or investing in a future? Not necessarily. A straightforward conversation is more productive for all parties than simply leaving a customer at the mercy of an AI that can’t recognize the intricacies of their situation. Make it easy for your customers to have that conversation.

When we talk about the so-called human element, what we’re referring to is a trifecta of right time, right solution, and right interaction – addressed, it bears mentioning, by the right person. This human element also acts as a bulwark against fraud and shores up customer satisfaction. This is good for any number of reasons, including marketing ones. Consider it an investment in which you’re conveying to your customers that your company is not going to treat them like just another anonymous customer, which is how many people feel with big banks. You’re tech savvy and you’re not a big bank and you’re on their side. In other words: AIs don’t generate loyalty; people do!