Becoming an ‘AI Powerhouse’ Means Going All In

Reading Time: 6 min 


AI in Action

This column series looks at the biggest data and analytics challenges facing modern companies and dives deep into successful use cases that can help other organizations accelerate their AI progress.
More in this series

There are plenty of organizations that are dabbling with AI, but relatively few have decided to go all in on the technology. One that is decidedly on that path is Mastercard. Employing a combination of acquisitions and internal capabilities, Mastercard has the clear objective of becoming an AI powerhouse. Just what does that term mean, and how is it being applied at the company?

What’s an AI Powerhouse?

Some refer to the idea of aggressive, pervasive adoption of AI as being “AI first.” Others use the term “AI fueled” or “all in on AI” (that’s Tom’s favorite, since it’s the title of his forthcoming book on the subject). Mastercard prefers “AI powerhouse.” Regardless of the descriptor, the concept refers to organizations that apply AI throughout their businesses, using it to either transform operations or develop new products, services, and business models, or both. These companies employ multiple AI technologies, develop many use cases, and put them all into production deployment.

As we’ve noted in previous columns, this level of engagement is unusual. In a 2019 MIT Sloan Management Review and Boston Consulting Group AI survey, 65% of respondents said they hadn’t seen value from their recent AI investments, in part because relatively few projects had been put into production. Survey respondents reported being skittish about the technology, with 45% saying they perceived some risk to their business from AI (up from 37% two years earlier).

To succeed in becoming an AI powerhouse, companies need to have a set of prerequisites in place. Senior leaders of the organization have to be on board and engaged with the effort because AI use cases must enable new ways of doing business and AI powerhouses inevitably invest at least a fair amount in AI. A lot of high-quality data is needed because that’s how machine learning models are trained. From a talent perspective, companies need smart individuals who can develop and apply AI systems. And it’s important that an organization have an ethical compass about how it uses AI and other information resources.

Mastercard has all of those essentials. Its CEO, Michael Miebach, is a strong supporter of AI at Mastercard. The company processes 110 billion payment transactions annually, it has hired or acquired many data scientists, and its executives are data and digital savvy. We’ve also written about how Mastercard was one of the first organizations to devote a senior executive to data ethics issues.

Building Mastercard’s AI Capabilities

Mastercard has acquired AI capabilities in recent years for both its internal use and to benefit its customers, including banks, merchants, and consumers.

To enhance its Data & Services business unit, which provides a variety of data-, analytics-, and AI-related services to external customers, it acquired Dynamic Yield, an AI-based personalization software company.

To support its Cyber & Intelligence (C&I) unit, which focuses on providing safe, secure, and frictionless interactions for Mastercard’s consumers, customers, and partners around the world, it acquired Brighterion, a company that uses AI to assist in fraud prevention; NuData Security, another fraud prevention company that uses behavioral biometrics to recognize users; and RiskRecon, a company that detects and ranks cybersecurity risk. Each acquisition was designed to help advance trust and inclusion in the digital ecosystem.

C&I is the home of Mastercard’s global AI organization. It’s headed by Rohit Chauhan, executive vice president for artificial intelligence. Chauhan is no stranger to data, having led the customer-focused Data & Services activities before taking on the AI role. It’s Chauhan who is charged with making Mastercard an AI powerhouse.

Chauhan and his team are working across a variety of fronts to make AI pervasive — “like electricity,” he said. There are five basic pillars of the company’s AI efforts:

  1. Powering products and services using AI. The first foray into this area was fraud detection, but Mastercard has leveraged these efforts into many different products and services for customers, including authentication and chargebacks. This work will ultimately encompass all components of the payment cycle.
  2. Powering internal business operations using AI. Chauhan said he is proudest of an AI-based forecasting application that predicts how much business Mastercard will do 90 days in advance with 99% accuracy. The company has also created AI-based predictive maintenance modules running in merchant-initiated transaction servers, through which all transaction messages are transmitted. The modules predict which servers will go down before they actually do.
  3. Supporting customers in their own AI journeys. Chauhan said that companies often get bogged down in big data and struggle with the adoption of new technologies. In addition to offering customers AI-based products, Mastercard has created a program called AI Express that works with customers to identify their own AI use cases and create proofs of concept in six to eight weeks. The C&I team has already done so for many different banks across the globe.
  4. Pursing AI for good. Mastercard is trying to use AI for sustainability and quality-of-life improvements. It has multiple AI projects underway related to inclusive growth and community development. It is partnering with the global nonprofit Accion to improve microfinance credit decisioning with AI. The Mastercard Center for Inclusive Growth is partnering with the University of Chicago’s Data Science Institute and to increase social impact through data science. The goal is to build the capacity of data science talent in underserved communities around the world.
  5. Prioritizing ethical AI. In 2019, Mastercard staked out a strong position on data responsibility that emphasizes customers’ ownership of, control over, and ability to benefit from their data. In AI, a product design process ensures that data and models meet the company’s criteria regarding transparency, lack of bias, and regulatory compliance. Chief data officer JoAnn Stonier has a strong focus on those issues, as described in a recent MIT SMR podcast on AI.

Chauhan believes that AI should eventually permeate the company and reside within every one of its divisions, but for now, his centralized group within C&I is necessary to convince business units of how important AI is, help them develop use cases, develop proofs of concept with them, and train their people in AI skills.

Chauhan’s group is trying to build organizational capability in a variety of areas, which includes training people in business units. “I truly believe it’s possible to train existing staff,” he told us. The AI group is identifying the best tool sets and vendor relationships for the company, building data sets and AI “feature stores” (collections of reusable, well-documented variables for use in machine learning models), and creating an infrastructure to support production deployment. The group is also developing a robust governance process for AI, including both human capabilities and those involving automation tools such as machine learning operations. And Mastercard is experimenting with automated machine learning, which Chauhan feels will eventually empower the work of many “citizen data scientists” around the company, although it is early in that process.

The greatest challenge for many organizations these days is acquiring the right talent to achieve their goals. Mastercard has already made great strides in that regard, in part through its acquisitions. But Chauhan is confident that Mastercard can continue to grow its capabilities through new hires and internal training as well. Organizations rarely define their goal as becoming a powerhouse of any kind without confidence in their ability to achieve it.


AI in Action

This column series looks at the biggest data and analytics challenges facing modern companies and dives deep into successful use cases that can help other organizations accelerate their AI progress.
More in this series

More Like This

Add a comment

You must to post a comment.

First time here? Sign up for a free account: Comment on articles and get access to many more articles.