Visa, Mastercard accelerate AI tools for businesses | PaymentsSource

• Key insights: Visa and Mastercard released agentic AI programs to court business users. 
• What’s at stake: The card networks are looking to expand revenue beyond card swipes, while businesses try to streamline processes. 
• Forward look: Both card networks are trying to recruit new users for agentic AI.  

With banks betting on agentic artificial intelligence to automate complicated tasks, Visa and Mastercard are sniffing an opportunity to sell emerging AI tools to card issuers and businesses. 

Processing Content

“We’ve seen a lot of point solutions for call centers and there are lots of applications,” Vidya Balakrishnan, general manager and vice president of financial services at artificial intelligence firm ServiceNow, told American Banker. “This increases complexity and these solutions are disjointed across banks.” 

ServiceNow has developed an agentic artificial intelligence payment dispute resolution system with Visa, contending AI agents can help human agents vet and resolve disputes faster. “AI can automate the whole intake and resolution process,” Balakrishnan said. Visa is upping its bet on new forms of AI at the same time rival Mastercard is making its own moves. Mastercard on Tuesday launched what it’s calling a “Virtual C-Suite,” or a set of agentic AI tools designed for small to medium sized businesses. 

Both Visa and Mastercard have made major investments over the past two years in agentic AI, which plays a key role in the card networks’ need to build new sources of revenue beyond card swipes. 

Vetting disputes

Visa will use ServiceNow’s AI platform to support the entire payment dispute lifecycle, automating triage, prioritization and documentation. 

AI agents are “trained” to spot low-value disputes and prioritize cases by using merchant data and historical trends, in an attempt to spot compliance risk earlier in investigations. “We’re looking at stuff like, how do you flag a dispute that doesn’t need a chargeback?” Balakrishnan said. “How do we make sure banks have the intelligence they need up front?”

ServiceNow is using its internal large language model, adding historical data such as the number of disputes a consumer may have filed over a certain period of time, and other aspects of the consumer’s tenure with the bank.

The Visa/ServiceNow service will not use AI agents to resolve disputes but to analyze current and historical data on behalf of the human agent. Each bank that uses the service will create instructions or questions for the AI agent to answer.  

“We’re looking at consumer sentiment and history to make a recommendation,” Balakrishnan said.ServiceNow did not release the names of banks using the Visa product, though Balakrishnan said there were banks onboarding the product.To support the release, Visa and ServiceNow reported the results of their own research that found issuers write off an average of $3.3 million per year and spend more than 360 hours annually on audit preparation. 

“The results were alarming, and the environment is super volatile,” Balakrishnan said. 

Mastercard research also reported a boost in chargebacks, predicting global chargeback volume will expand 24% from 2025 to 2028, reaching 324 million transactions annually. 

And each dispute costs financial institutions $9.08 to $10.32 to process on average, according to Mastercard. 

 “Historically, fraud and dispute claims have been fully outsourced or managed through cumbersome, error‑prone processes involving spreadsheets and labor‑intensive investigations by bank employees,” Lindsay Hooks, a director at Cornerstone Advisors, told American Banker. Some institutions have made incremental improvements, such as automating intake or implementing workflow tools to boost efficiency, reporting, and compliance adherence, according to Hooks. “Unfortunately, without proper attention those legacy solutions quickly become outdated and cannot keep up with the sheer volume of first-party fraud cases to sift through,” Hooks said. 

Emerging AI‑driven solutions are transforming dispute management, according to Hooks, adding new platforms leverage various types of AI to rapidly analyze transaction data, historical customer behavior, and supporting documentation to assess claim legitimacy, determine chargeback rights and recommend next actions. 

“With both true fraud and friendly fraud on the rise, and with strict regulatory and network requirements governing timely and appropriate responses, technologies that leverage AI are becoming the leading way for banks to regain control of losses, streamline operations, and ensure consistent compliance,” Hooks said.

AI exec

Mastercard this week also launched new agentic AI tools, with a focus on small to medium sized businesses. 

Mastercard’s Virtual C‑Suite creates “AI agents” that aid in responsibilities such as finance, security and marketing. Similar to Visa’s dispute resolution, Mastercard is positioning AI as a way to help human staff, rather than make decisions. The Virtual C-Suite analyzes business performance, identifies risks and opportunities, predicts likely outcomes and recommends prompt steps and long-term moves. The service features dashboards and interfaces that enable business owners to ask the AI agent direct questions. A virtual CFO will be the first piece of Mastercard’s Virtual C-Suite, with other job categories to follow. 

As Visa and Mastercard seek clients, there are signs businesses are warming to using agentic AI, with nearly 80% adopting the technology in some form, according to PwC. 

Mastercard’s new release is focused on businesses that are not likely to have large management teams. 

“Small businesses are the cornerstones of communities, but it’s easy for owners to lose sight of the passions that inspired them when they’re buried in spreadsheets and stretched across multiple roles,” said Mark Barnett, global head of small and medium enterprises at Mastercard,  in a release. “We hear these pressures from entrepreneurs every day.