Agentic commerce marks a shift toward a future where AI assistants can independently search, decide, and transact on behalf of people and businesses. Mastercard’s Agent Pay which was launched in April last year enables software agents to securely complete purchases and manage transactions while maintaining strong safeguards around trust, consent, and accountability.
Mastercard has also expanded its partnerships and programs to accelerate adoption of agentic commerce technologies. The company has integrated Agent Pay into broader commerce experiences, including Microsoft’s Copilot Checkout, and is collaborating with industry partners such as PayPal to embed secure, intent-driven agentic payments across the value chain. Additionally, Mastercard’s Start Path startup engagement initiative is being expanded to support innovators building infrastructure and services for AI-assisted commerce at scale.
Google’s Universal Commerce Protocol
Earlier this year, Mastercard also announced that it has joined forces with Google on the Universal Commerce Protocol (UCP), an open and interoperable standard designed to enable seamless interaction between AI agents and merchants across discovery, purchase, and checkout phases while preserving security and transparency. Within this ecosystem, protocols such as UCP and the Agent Payments Protocol work together to protect merchants from fraud, give issuers confidence in transactions, and provide consumers with clarity and control over agent-mediated purchases.
Mastercard: Patenting Activity
Filings peaked in 2016 and 2017, a period that aligned with Mastercard’s rapid expansion of digital payments, tokenization, and cybersecurity. This was also when the company moved into biometric authentication with initiatives such as its “Selfie Pay” identity verification app in Europe, explored payment integrations with social media platforms like Twitter (now X) and Facebook, and strengthened major digital wallet partnerships including Apple Pay’s international launches. Together, these efforts reflected a broader push toward mobile, secure, and frictionless payment experiences.
Activity slowed in 2018, then picked up again in 2019 and 2020 as Mastercard invested more in contactless payments, AI-based fraud detection, and network strengthening. The high number of pending filings in 2020 reflects the impact of COVID-19, which accelerated global adoption of e-commerce and contactless payments.
From 2021 onward, filings declined sharply, reaching a low in 2022 before a partial rebound in the subsequent years. This shift followed key strategic moves by Mastercard, including the acquisition of cybersecurity and identity firms such as Ekata in 2021 and expansion of open banking initiatives after earlier deals, such as Finicity.
By 2024, Mastercard’s AI-related patents accounted for approximately 18.1 percent of its portfolio, reflecting a focus on data-driven payments, fraud detection, cybersecurity, and personalization. Growth was driven by trends like real-time payments, rising e-commerce fraud during COVID-19, and broader use of machine learning in risk scoring and identity verification, supporting AI-powered fraud prevention and decision intelligence initiatives.
Mastercard: Top Technology Areas
Mastercard’s patent activity is heavily concentrated in core digital infrastructure areas. Information and communication technology (G06Q) dominates the portfolio, highlighting the company’s focus on building secure, scalable, and interconnected payment systems. This is closely followed by technologies for transmitting digital information (H04L) and processing electronic data (G06F), which together reflect Mastercard’s emphasis on real time transactions, data handling, and system reliability across global networks.
Smaller but still important portions of the portfolio point to innovation in wireless communication (H04W), advanced computing models (G06N), and graphical data reading (G06K), all of which support faster, smarter, and more user-friendly payment experiences.
The presence of image and video recognition (G06V), coin-freed apparatus (G07F), and receipt or token registration technologies (G07G) shows a broader push toward automation and digital commerce. Overall, the distribution illustrates how Mastercard blends foundational computing and networking technologies with emerging tools that enable frictionless, intelligent, and fully digital financial ecosystems.
Building trust and transparency in AI-driven offers
Traditional recommendation systems in payment networks struggle to keep up with constantly changing cardholder data. They often require complicated preparation and significant computational resources to analyze transactions, making it hard to give timely, accurate, and personalized offers to cardholders. This limits the ability of banks and merchants to engage customers effectively.
Mastercard uses a deep learning system to provide personalized merchant offers to cardholders more efficiently. It analyzes past transaction data to predict which offers a cardholder is most likely to use. The system combines two approaches: one that looks at direct cardholder-merchant interactions and another that finds hidden patterns in the data. A separate engine matches available offers to the right cardholders, and another delivers and tracks the offers when used.
The system can also use cardholder preferences, location, and similarities between users to make recommendations smarter, allowing banks and merchants to send timely, targeted offers without heavy computing work.
U.S. Patent No. 11,250,461, titled “Deep learning systems and methods in artificial intelligence”, was filed on September 25, 2019, and was granted on February 15, 2022. Suqiang Song was listed as the inventor. Daniel Fitzgerald from Armstrong Teasdale LLP represented Mastercard in the filing.
Faster, smarter lending decisions for banks
Small and medium-sized businesses often struggle to get loans because they lack enough collateral or formal financial records. Banks are hesitant to lend to these businesses since they are more likely to default than larger companies. Traditional credit scoring methods rely on formal financial data that many small businesses do not have, leaving most under-financed or forced to rely on informal funding.
Mastercard’s application uses a computer system to predict the creditworthiness of small businesses using invoice and transaction data instead of traditional financial records. The system creates a network that links a business to other companies it works with, tracking payments, transaction amounts, and supply chain relationships. It then analyzes this network to generate a credit score that reflects the business’s ability to repay loans. This score can be automatically shared with lenders, helping small businesses access financing even if they do not have formal credit histories.
U.S. Pat. App. Pub. No. 2024/0119517, titled “Artificial intelligence based methods and systems for predicting creditworthiness of merchants”, was filed on December 6, 2022, and was published on April 11, 2024. The application lists Ayush Agarwal, Maneet Singh, Bhanupriya Pegu, and Shivshankar Anand Reddy as inventors. Barta Jones is representing Mastercard in the filing.
Scaling financial data management with AI
Financial transactions often include messy or inconsistent data, making it hard to identify and standardize merchants, payees, or account names. This complicates analyzing histories, detecting fraud, and generating insights, as traditional systems struggle with text variations and large data volumes. Manual cleaning is slow, error-prone, and not scalable.
The invention provides a computer system that trains an entity resolution model to automatically identify and standardize entities in financial transactions. The system takes historical transaction data and uses natural language processing to extract entities, create a dictionary of labels, and tag transactions with the identified entities. The text is tokenized and converted into a format suitable for a transformer-based model, which is then trained to recognize and standardize entities across transactions.
Once trained, the model can process new transaction data to automatically match and normalize entities, improving data consistency and accuracy for financial analysis, reporting, and fraud detection.
U.S. Pat. App. Pub. No. 2023/0385701, titled “Artificial intelligence engine for entity resolution and standardization”, was filed on May 26, 2023, and was published on November 30, 2023. The application lists Yogesh Sakpal, Gauri Shah Bhatnagar, Shraddha Shirke, Dean Vaz, Siddhesh Dongare, Dmitriy Kontarev, Brett Ragozzine, and Christopher Brousseau as inventors. Hovey Williams is representing Mastercard in the filing.
Mastercard: Top Law Firms
Mastercard’s patent work is handled by a group of law firms, with Armstrong Teasdale clearly emerging as the most prominent partner. Harness, Dickey & Pierce also plays a major role, followed by Buchanan Ingersoll & Rooney, Panitch Schwarze Belisario & Nadel, and Buckley, Maschoff & Talwalkar, which together form a strong secondary cluster. Hovey Williams appears as another consistent contributor, indicating an ongoing relationship in patent matters.
A smaller set of firms supports Mastercard in a more limited capacity, including Talem IP Law, Barta Jones, Jordan IP Law, and Michael Best & Friedrich. Their inclusion suggests more focused or selective engagements rather than broad portfolio management. Overall, the mix of firms reflects a strategy centered on a core group of trusted patent counsel, supplemented by additional firms for specific expertise or needs.
