Downtime can be costly for banks and financial institutions. AI can help.
AI-powered tools can help banks beat downtime, protect brand and avoid losses.
Spells of downtime, however short, are unavoidable for most businesses. But that doesn’t make them any less costly — or disruptive to the lives of clients.
Outages can have a material impact on business, particularly for financial institutions. Every second of an outage can mean lost transactions, frustrated customers and potential damage to brand and reputation.
According to a Ponemon Institute study, just one minute of downtime can cost a business an average of $9,000, bringing the cost per hour to over $500,000.
But there’s even more at stake than losses adding up by the minute.
Unplanned downtime can impact brand, consumer trust
Trust is crucial to financial institutions, yet outages — whether caused by scheduled maintenance, local power outages, or other unexpected events — are often disruptive for consumers, making it hard for them to make payments or access their funds when they need them. And few things panic a consumer more than being suddenly unable to make a payment — no matter how short the outage might be.
For banks and financial institutions themselves, downtime can mean a loss of revenue, call volume spikes to customer service centers, angry customer posts trending on social media and even scrutiny from regulators.
To avoid the worst outcomes, banks and financial institutions are investing in ways to strengthen their downtime solutions.
AI can mean fewer false declines, help beat downtime
When unexpected downtime occurs, traditional solutions are often unable to keep up with the vast amount of data needed to get back up and running. They tend to rely on static, rules-based models as a backup method to manage transactions, which can make for a large number of unnecessary declines — and the angry consumers, merchants and issuers who might come with them.
AI can improve the payment experiences during outages by mirroring issuers’ uptime approval decisions with a high degree of accuracy.1
When a bank experiences downtime, AI-powered solutions like Visa’s Smarter Stand-In Processing (Smarter STIP), whose model is trained on billions of data points of cardholder activity, can step in to define correlations and emulate likely issuer decisioning behavior on its own. The model can then generate informed decisions to approve transactions on the issuer’s behalf — up to 50% more than might have been approved with existing STIP solutions2 — providing far more consumers with a frictionless payment experience than traditional solutions can.
Where static rules applied across an entire card portfolio can result in a larger percentage of unnecessary declines, Smarter STIP’s AI-driven payment model can automatically tell, with a high degree of accuracy,3 when a transaction would likely be approved by the issuer — for instance, by recognizing the location of the merchant in relation to the cardholder and the time of day they are shopping.
AI can help banks and financial institutions better serve clients
Ultimately, AI can help banks and financial institutions with multiple challenges — from real-time visibility, to better forecasting, to more optimized operations. AI can enhance financial institutions’ ability to serve clients — especially during outages — by harnessing data and deep learning to provide consumers with a smarter, stronger, and more dynamic experience. And it can give today’s digital-native consumers the kind of real-time visibility into their finances they increasingly expect, as more and more people conduct banking on mobile and online platforms. All of this together makes for a better experience — and happier clients and consumers.
Learn more about how Visa uses AI in payments at VisaNet+AI.
1 95% average accuracy observed in internal simulation of the Smarter STIP model offline on transactions in Q4 2019 for all Visa BINs globally.
2 Based on Visa internal analysis comparing existing and Smarter STIP approval rates of all transactions in Q1 CY2020 for a single US Issuer.
3 95% average accuracy observed in internal simulation of the Smarter STIP model offline on transactions in Q4 2019 for all Visa BINs globally.