Accertify’s Conrad Kennington Breaks Down Questions Retailers Should Be Asking About AI in Total Retail
Retailers evaluating AI-powered fraud prevention often start by asking about model accuracy and data volume. For retail teams, the better question is whether the AI can approve more legitimate customers, reduce unnecessary friction, and improve financial performance compared with the systems already in place.
As retailers continue investing in artificial intelligence, evaluating the technology often starts with questions about model performance and the data behind it. But according to Accertify’s Conrad Kennington, Vice President of AI, those questions can overlook what retailers ultimately need to understand: whether AI will deliver better financial outcomes than the systems they already have in place.
In a new article for Total Retail, Retailers Are Asking the Wrong Questions About AI, Kennington draws on dozens of conversations with retailers evaluating AI tools to explain why model performance shouldn’t be considered in isolation. Different retailers operate with different customers, risk profiles, and business objectives, making a universal definition of a “good” AI model difficult to apply.
Instead, Kennington argues that retailers should evaluate AI based on the outcomes it produces. In fraud prevention, that means looking across approval rates, fraud detection, false declines and, ultimately, financial performance. A model that prevents more fraud may not necessarily be delivering greater value if it also turns away more legitimate customers than the system it replaced.
The same principle applies to data. While large datasets can strengthen AI models, volume alone does not determine performance. What matters is whether the data provides the diversity, relevance, and context needed to distinguish legitimate customer behavior from fraud.
As retailers weigh new AI investments, Kennington encourages them to shift the conversation away from technical benchmarks and toward a more practical question: Will this technology help us approve more legitimate customers, reduce unnecessary friction, and improve financial performance?
Read Conrad Kennington’s full article in Total Retail.
Frequently Asked Questions
How should retailers evaluate AI fraud prevention?
Retailers should ask whether the AI will improve business outcomes, not just whether the model looks impressive in a demo. The most important questions are: Can it approve more legitimate customers? Can it reduce false declines and unnecessary friction? Can it adapt to the retailer’s customer behavior, fraud patterns, and risk tolerance? And can it prove a measurable impact on fraud losses, approval rates, manual review, and revenue?
What is a false decline?
A false decline happens when a legitimate customer is incorrectly rejected because a transaction is mistakenly identified as risky or fraudulent. False declines matter because they directly cost retailers revenue, create avoidable customer frustration, and can weaken long-term loyalty. In fraud prevention, the goal is not just to stop bad transactions. It is to make more precise decisions so good customers can complete their purchases with confidence.
How does fraud prevention affect approval rates?
Fraud prevention directly affects approval rates because every decisioning strategy determines which transactions are accepted, reviewed, or declined. Overly conservative systems may block more fraud, but they can also reject legitimate customers and reduce revenue. Stronger fraud prevention should help retailers make more precise decisions, approving more good customers while still controlling fraud risk.
How can Accertify help?
Accertify helps retailers move beyond narrow fraud blocking and toward smarter risk decisioning that supports growth. By combining AI-powered decisioning, machine learning, global network intelligence, device and behavioral signals, and configurable rules, Accertify helps retailers identify fraud more precisely while keeping good customers moving. That means teams can reduce false declines, improve approval rates, limit unnecessary manual review, and protect revenue without adding friction to the customer experience. Learn more at accertify.com