Powering the payments experience with AI
AI. We’ve read about the hype and the promise – as well as some amazing applications in many industries. Here we look at how it’s impacting payments in particular.
Cloud has ignited widespread use of AI
Cloud computing has made it easier than ever to build, test, and scale services in a more flexible way. From storage to security, it’s helped businesses operate with agility while experimenting with a new-found freedom armed with powerful technologies.
A prime example of this is its enablement of artificial intelligence (AI). The availability and affordability of near-limitless compute resources in the cloud compared to on-premises has spurred the development of AI use cases across all industries. According to IDC, AI spend is forecast to grow 18.8% in 2022 and reach $500 billion by 2024. And it’s easy to see that bolstered by the sheer scale, speed, and affordability of data processing beyond what’s been achievable before, more companies than ever can take advantage of AI.
If you wanted to develop a prototype for a self-driving car 5 or 10 years ago, for instance, you’d be looking at spending tens of millions of dollars. The hardware would fill an entire data centre. Fast forward to today, and thanks to AI you could probably do the same for $100,000.
When to comes to the payments realm, AI is proving to be a game-changer in both risk and compliance, and consumer engagement.
The deep connections that draw insight and de-risk
The evolving payments landscape – accelerated by open banking, PSD2, and faster payments – has been driving benefits for merchants and consumers for some time in the UK and Europe. But, as with all advancements there come new challenges. Being able to run credit scores and Know Your Customer (KYC) checks, and assess risk in real-time or near real-time is more important than even.
Payments providers are pivotal in this process. False positives, for instance, are a major source of revenue loss for merchants. So running all the necessary checks not only faster but smarter is key. AI is an incredibly powerful tool in doing this, with the application of machine learning (ML) enabling more effective fraud detection and increased authorisation rates.
Its advantage lies in its ability to evaluate huge amounts of transaction data and recognise patterns that wouldn’t be possible manually. So it doesn’t just see a single transaction going through a digital wallet. It looks at the patterns of usage for this person’s account. Whether the current transaction is typical in terms of time, frequency, and value. But more than that, it looks at the entire ecosystem around that transaction – all the interconnects and relationships it has with other accounts and merchants. It can even make links between accounts that on the surface appear unrelated, but through deeper analysis are found to share an IP address.
When stacked against each other, an ML algorithm has many advantages over manual processing in terms of the depth of the analysis it can achieve and the connections it can uncover. In doing so, it can deliver the best of both worlds: convenience and protection for the consumer, as well as faster approval rates, fewer declined transactions, and increased revenue for the merchant.
Enriching engagement, the virtual way
AI can also be used to develop higher levels of engagement. One aspect of this is the communication between consumers and their merchants, banks, and digital wallet providers. While we’ve all probably had a questionable experience with a chatbot in the past, the technology is becoming more sophisticated and getting better at delivering service that matches, and sometimes even exceeds, that of a human agent.
Again, the differentiator is the technology’s ability to draw on a mass of consumer information – including historic transactions and behaviours – and use that to inform how the chatbot responds to the request. With a more complete view of customer requirements and expectations, the virtual agent can respond in a more tailored way and, in some instances, even predict what the customer will need next. Applied to issue resolution, this is a huge benefit. Over time, AI models get smarter too. They continue learning and improving, enhancing their usefulness and becoming more accurate and efficient.
Giving the consumer what they want, when they want it
Perhaps the more interesting application of AI in the realm of customer engagement is mass personalization or mass customisation. In a way, it’s applying real-time analytics and AI and ML capabilities to create the type of close consumer–merchant relationship that used to exist before big, impersonal stores took over the world.
For example, you’re on holiday in Las Vegas. The brand from which you normally buy jeans in your home town realises you’re in the area, using your location information. The branch manager pings you a text message with an invitation to come into the store to have a coffee, and attaches a promotion code for 20% off your favourite style of jeans.
In this scenario, the merchant knows you’re there, they know what jeans you usually buy, and they know they have your favoured style and size in stock. So by putting all this together, making those connections and tailoring outreach, they can achieve a pretty slick and highly personalised interaction. And whether you take the offer or not, you’ll definitely remember it, probably feel quite special, and maybe even get a sense of being closer to the brand as a result.
An intelligent ecosystem complemented by AI
AI provides a powerful set of tools to change the consumer experience. Its ability to quickly uncover links that wouldn’t otherwise be identified means merchants can both know and provide for their customers better, thereby increasing revenue. And they can protect their own businesses more effectively.
As we’ve seen, the role of payments providers is central to the whole process – using AI to combat fraud faster and more efficiently, and increasing approval rates by addressing key sticking points for merchants in the transaction process.
Of course, the technology is still evolving – and its use comes with ethical considerations – but the intelligent application of AI in areas where it can lend efficiency, insight, and security will result in an improved ecosystem for all.