News and resources on payments systems, innovations and initiatives worldwide.
MPE 2024: How are merchants implementing AI in payments?

MPE 2024: How are merchants implementing AI in payments?

At the Merchants Payments Ecosystem conference 2024, Chairwoman Anna Maj moderated the session ‘Next-Generation Technologies for Merchant Payments’, speaking to a panel of experts including Bartosz Skwarczek of GZA Capital Group, Galit Shani-Michel from Forter, Bono van Nijnatten from ASOS Payments, and Francesco Burelli of Arkwright Consulting. Maj questioned the panelists on how AI is being used in merchant payments and where AI can take their companies in the future.

Shani-Michel started off by explaining that automation, analytics, and identity is where AI can make a huge difference, and merchants should concentrate on how AI can be utilised to help monitoring and generate revenue.

General manager at ASOS Payments, van Nijnatten described how the major online fashion retailer is using AI for fraud. He stated that progressing for rule-based to machine-learning allowed them to cut a staggering 60% of their fraud department in favour of AI.

He continued: “However, for the people who are still working in the fraud department, they don't have to work around the clock anymore. Instead of working 24/7, they can work regular business hours. At the same time, AI will always go hand in hand with the decisions of humans; there's always a person who will be the final decision maker on fraud.”

Skwarczek called AI a “double-edged sword”, which, on the positive side, can be used for copywriting and creating content, for streamlining efficiency in the AI department, and developing cybersecurity tools. Yet, on the other side, he said that AI can be “extremely stupid” in that it can make simple, common mistakes in instances where there is conflicting data when it comes to simple facts. There are several stages of development before AI is at a stage where it is operating efficiently.

Maj opposed Skwarczek’s last viewpoint, saying that AI can only be 'stupid' when it is programmed that way, and there are limits to AI, such as bias and the knowledge base available. However, with GenAI and better coding, AI possesses a wealth of potential.

Burelli implied that there is a need for a strategic approach to AI implementation, and businesses should take a proactive and human-centered approach that includes engaging with employees to use AI in order to leverage the potential of the technology.

Discussing the deployment of AI for retailers, van Nijnatten noted ASOS’s launch of an AI stylist that can collect clothing items and accessories relevant for any occasion, curating a personalised experience for the consumer.

Skwarczek stated that there should always be a human core in the process. AI solutions should be implemented to make the client experience as seamless as possible, but a human factor is essential to be between all these technologies to make them available for clients. He used an example of his company:

“The AI solutions we use gave to us back time from process of dealing with chargebacks, with disputes and everything related to that, so they could be more involved with other things. The AI solution that we implemented provided support and gave us the time to be more focused on other parts of the job which are extremely important.”

Discussing fraud management and prevention in AI, Shani-Michel explained that AI has complex solutions and cannot just be a one-stop-shop to solving all the issues in a company.

Why is trust essential?

In a keynote titled ‘Trust is the New Gold’ Chris Reid, EVP of identity solutions at Mastercard, explored how fraud occurs and how cybercrime has evolved with new technology.

He explained that online fraud markets are efficient, vibrant, and competitive. They are operated by a variety of supply chains: factories, raw materials, and distribution. Criminals are stealing identity documents which can be used to falsify documents. He detailed that the quality of these false passports, driving licenses, BRPs and more have evolved to become more detailed and accurate.

Additionally, false bank accounts that are KYC-approved are available for sale on the dark web, including being able to ship false cards that are sometimes aged and have gone through bank checks. Looking to solutions, Reid focused on how businesses are building a trusted identity in the merchant and payments side, and forming a network taking learnings from the dark web to create a solution.

“What you get within our ecosystem is good growth, sustainable growth within commerce and merchant payments. That's what we're aiming to do. Yes, we're focused on commerce, but we're raising the ceiling for growth by driving a greater level of trust.”

How can AI be leveraged to combat fraud?

On a keynote titled ‘From Cost Reduction to Revenue Generation’, VP of payments at Forter, Galit Shani-Michel touched on how AI can be leveraged in fraud and payments. She started by asking the audience how many of their companies were using AI for fraud detection, of which a majority responded no.

Shani-Michel explained that AI and machine learning is where bad actors can be identified and stopped, citing statistics that customers will never buy from a site again after being falsely declined and 60% of consumers will abandon the checkout process if it takes longer than two minutes. The point is that customer experience is key, trust is vital, and that being a reliable retailer will increase revenue — that AI can help generate.

She continued that AI can also make accurate fraud decisions based on the identity behind the transactions, removing friction for customers. Shani-Michel emphasised how AI needs to be leveraged to understand issuer behaviour and automate payment testing to determine what is the best process for each transaction, so it is all about using AI for the right purposes.

She concluded: “Right now, if you ask the right questions, with the right tools, you can get very quick answers and then spend your time fixing it. We spoke about how employees are afraid of not being needed anymore — you will still be needed. You will just be able to use your time to actually fix the problem rather than dig into your data and try to understand what went wrong. So I will also encourage you to think about monitoring and reporting. But more than anything else, remember that if you don't stay ahead of these fraudsters, they will win the game.”

Comments: (0)