Mastercard is building AI agents for banks to enhance product recommendations, highlighting the payments network’s heightened focus on agentic advising solutions.  

The agents, which are being rolled out this quarter, use Mastercard’s data to help banks better understand customer behavior and predict cash-flow requirements, Kaushik Gopal, executive vice president for business and market insights, told FinAi News.  

This leads to more accurate product recommendations, he said, including:

• Loans;
• Credit card rewards programs; and
• Savings accounts.  

The recommendation tool reflects how Mastercard is expanding beyond a traditional payments network as it looks to strengthen its advisory business and data solutions with agentic AI, Gopal said.  

“We have this huge proprietary data set that has powered our advisory business for our customers, but now we’re just taking that into the agentic age,” he said.  

“Think of an always-on agent that is able to analyze [consumer-consented] data and then say, ‘Hey, we’ve noticed this change in your behavior; we think this will be of value. Would you like to hear more?'”

Mastercard expects the recommendation agent to become “more sophisticated in spaces like portfolio management,” highlighting agentic AI’s overall potential, Gopal said. 

“As the models get more and more sophisticated, particularly led by the large LLMs, we will get into a world where certain decisions and recommendations can become far more deterministic,” he said. 

Mastercard has launched other agentic tools beyond payments this year as part of its Agent Suite. For example, the company’s Virtual C-Suite is designed to support small businesses with tasks such as performance analysis, risk management and market predictions.  

Lower unit of expense 

In addition to improved accuracy, Mastercard expects its product-recommendation agents to drive efficiencies at banks by operating 24/7, minimizing disputes and reducing the need to commission researchers to analyze consumer behavior, Gopal said.  

Disputes, in particular, create friction in banking operations and can shrink profits, he added. 

“This is not about reduction in labor force; it’s about a lower unit of expense for growth,” Gopal said. “There’s net new engagement, more transactions that are taking place, but against a set of security standards.” 

The degree with which these agents engage with customers will depend on banks’ consumer-consent policies, he said, adding that the overall success of agentic AI ultimately comes down to trust. 

Mastercard’s largest banking partners include:

• $4.9 trillion JPMorgan;
• $3.5 trillion Bank of America;
• $2.8 trillion Citi; and
• $683 billion Capital One.  

Register here for the FinAi Lending Summit, set for Oct. 7-8 in Las Vegas.