What Fit is your AI – Off the Rack or Tailor-made?

The benefits of incorporating artificial intelligence across the payment process are well-documented, from optimizing payment routes to anticipating and mitigating downtime; and the impact is being felt by retailers and their customers alike.   

For instance, we are using AI to track bank response and fulfillment times, against their SLAs and typical response times, any deviation from the norm is logged and – in the event of a deterioration of service levels – our retail partners are automatically alerted and payments re-routed. 

Even a small decrease in bank response times can represent a precursor for potential downtime (due to over-loading or maintenance, for instance); AI’s ability to anticipate and mitigate such occurrences represents a distinct competitive advantage.  Our own research highlights 28% of Latin American online retailers are experiencing payment downtime at least once a week, while 82% currently have no possibility of alerting their customers of the same.

AI’s capacity to track and analyze terabytes of data in real time makes it particularly relevant for spotting suspicious or fraudulent behavior, either by a consumer or an intermediary.  We apply AI to set up (and mitigate) attack scenarios to help recognize, forecast and prevent such interventions.  

These are crucial considerations for retailers – Mastercard estimates that around 20% of total e-commerce revenues in Latin America are lost to fraud – and the impacts of such strategies can be almost immediate.  In one example, the use of AI enabled us to reduce the level of fraudulent transactions in Colombia for a client from 8% of daily transactions to 1%, in just one month.  

The crucial detail

All great examples of AI transforming the payment experience.  But there is one small but important detail that is rarely mentioned:  there is AI and ‘payment AI’ - and when it comes to online payments, it’s only the latter which counts. 

Each of the above examples was based on actual, specific and local data; the real benefits of AI in payments will not be found in generic data or models.  Take the example of route optimization in anticipation of bank or intermediary performance issues or downtime.  Our company manages payments to and from more than 30 countries (including pay-outs as well as pay-ins); so, comparing actual bank response times with stated SLAs and acting in the event of a divergence is not a ‘generic’ process.  It requires real time data from multiple banks and intermediaries in each country; and not just for those involved in the particular transaction; alternative fulfillment partners are tracked in the event that the transaction requires re-routing.  

This is particularly true when it comes to fraud detection; behavioral analytics are specific to each sector (transaction amounts and user expectations can differ markedly between a retail shopper and an online trader, for instance), and each country.  With respect to the latter, country differences range from preferred payment methods, to compliance and regulation requirements for each market – all of which should be factored into the analysis when identifying probable fraud.

The presumption that AI is somehow generic and should simply be ‘applied’ to the payment process is a false one – when it comes to payment processing, not all AI is the same. The only AI that makes a difference, is AI that is specific and taylor-made for your customers, their location and their behaviors. 

Like your customers, each AI-driven insight is distinct

Customer experience provides another clear example of the impact of genuinely ‘proprietary’ AI, specific to each retailer. Working with a Brazilian retailer, for instance, our AI has been able to, literally, slash response times to customer service tickets.

In May, 50% customer service tickets for this client were resolved within 23 minutes, while 75% of the same were closed within 111 minutes (nearly 2 hours).  Today, those numbers have literally been decimated: 50% of tickets are now resolved within a minute, while 75% are closed within 12 minutes (the new SLA established).   

These reductions were achieved through a combination of analysis and automation; AI is now wholly or partly involved in resolving all customer service tickets, with 65% of resolutions being entirely automated. 

When it comes to payment AI, get specific or go home

Today, we are also leveraging AI to identify individual customers’ preferred payment methods, and to streamline their experience accordingly; such preferences are based not only on the sector but the country of transaction and – even – the nationality of the end user.  

Each country’s citizens have specific payment preferences and expectations, many of which will also vary with the sector they are purchasing from, as well as the amount of that purchase.  In each case, delivering the appropriate payment experience for each transaction matters: according to our research, almost nine out of ten (87.5%) consumers consider their preferred payment process and option to be either ‘important’ or ‘fundamental’ to their purchase decision.   

Such insights and competitive advantage are not generic; they are specific and particular to each sector, brand, country and, even, user.  It’s only through round-the-clock and round-the-globe monitoring of every step of every transaction that we are able to deliver such results.  The resulting data enables us to track the performance of each bank – and, potentially, each transaction – in real-time. In the event of anomalies, we can further investigate by observing the metrics of the users of that particular financial institution and alert both the bank and relevant retailers.  This is the benefit of genuine payment AI, based on real transactions, real payment trends, and real customer behavior.  

That’s why, if you want to deliver these types of impacts - your payment AI should always be ‘tailor made’ and never simply ‘off-the-shelf’!  Anything else is simply missing the point. 

About D24

D24 is a leading payment service provider that empowers companies in more than 27 countries across 3 continents to access comprehensive ‘deep payment’ processing solutions including cards and alternative payment methods. Securely adhering to strict industry governance, and providing access to more than 250 APIs via a single API for seamless integration, our customers benefit from efficient pay-ins and pay-outs with specialized solutions and payment gateways tailored to merchants’ individual local needs. 

Disclaimer: D24’a cross border payment services are not applicable or available in the UK, US, nor further jurisdictions mentioned throughout this publication.