AI-powered customer journeys need to be centered in the human experience

“When all you have is a hammer, everything looks like a nail.”

This is a phrase that frequently comes to mind when thinking about how best to use—and not use—the power of Artificial Intelligence (AI) to move our business forward. To be sure, AI is a powerful instrument, and when it comes to improving the customer experience for the end users of our products, it is among the most critical tools Mastercard has at its disposal today. Our deployment of AI allowed us to prevent $20 billion in cyber attacks in the past three years and a further $20 billion in fraud in 2020 alone. When you have 75 billion transactions passing through your network annually, having systems that can identify fraud, and learn to identify emerging forms of fraud, is exceedingly valuable.

However, AI alone is rarely a silver bullet, and when it comes to something as nuanced and as complex as the customer journey, this technology is likely to be just one component of the solution.

Thinking holistically about customer needs

In my role with Mastercard, and more broadly throughout the company, we’re increasingly focused on the wider spectrum of the customer journey, and helping businesses to understand how best to serve their clientele throughout it. It’s because of the centrality of payments to these journeys that we’re able to expand outwards and provide valuable insights into things such as where customers are finding and interacting with businesses, the nature of broader purchasing patterns, and whether loyalty programs are having a real impact on sales.

If you’re only thinking about how AI can directly increase sales or cut costs, you’re likely missing potential opportunities to apply this technology to various touchpoints that customers have with your business—and that not all of them are going to translate immediately into sales. AI may result in more personalized, meaningful customer interactions, ultimately building stronger relationships with them moving forward. The beauty of AI is that it’s not just capable of predicting what a customer might want to buy from you and show it to them, but also help predict where, how, and why they want to buy. This, in turn, can improve the trust your customers place in you and your business. There are ever improving solutions that can be seamlessly integrated into a customers current experience but provide them with peace of mind they didn’t realise they needed. One example of this is using AI to proactively identify and remedy non-compliant merchant codes to reduce customer friction, enhancing their satisfaction and both banks and merchants revenues.

Your AI is only as good as your data

A real focus of how we use AI needs to be the quality of the inputs we’re providing to power these algorithms. While machine learning can offer incredible insights that can help to deepen the customer relationship, the nature of the technology is such that it can only work with the data it’s been given. Limited information may lead to misleading conclusions, while even vast datasets need to be scrutinized closely to account for potential biases.

To this end, it’s possible to see situations where machine learning is able to help you better predict the needs of individuals that roughly align with your existing customer base, but it fails to help you to understand how you’re underserving some parts of the population; for instance, it’s unlikely that machine learning is going to be able to help you to build in assistance for customers who suffer from visual impairment. Furthermore, even a large dataset may be misleading if your customers are largely homogenous, and if you rely too heavily on AI you may fail to see new opportunities, or you might construct a rigid customer journey that doesn’t account for potentially greater diversity in who you can cater to. Without care, it’s possible that an overreliance on the technology could both hurt your business and exclude groups of people.

AI as a team member rather than a manager

As technology continues to play a greater role in our lives, it’s going to become more critical to understand what its limitations are. At Mastercard, AI and machine learning have proven invaluable when it comes to important challenges like preventing fraud, but even then, we have vast numbers of people who are constantly monitoring the algorithms, checking for blind spots, and trying to account for the idiosyncrasies of normal human behavior. Ensuring that the payment process is secure is just a small part of the overall payment experience, which in turn, is only a fraction of the customer journey. We have both machines and human beings looking at a whole spectrum of business challenges, and it’s incumbent upon people such as myself to work out how best to blend these resources to ensure that we’re doing right by our customers and end users.

We use the phrase customer journey because of the people at the center of it: these are individuals with complex needs, backgrounds, and interests. While AI can definitely help us to understand these stories better and shape our operations accordingly, humans need to continue playing an intrinsic part in helping to make smart decisions about where and how to apply these technologies, and perhaps most importantly, to know when an algorithm actually can’t give us the picture we need. Often, human intelligence and a sense of empathy really are the best solutions.