Build or buy AI fintech? For Mastercard the answer is “both”

Enterprise IT organizations have tons of questions about whether to build or buy AI tools to create content, manage workflows, perform searches and summaries, monitor network activity for security purposes, find sales and marketing opportunities. .. the list goes on and on. Calculating the costs of customizing a ready-made tool versus coding it yourself becomes a difficult, multivariate mathematical problem that IT must solve with the finance team.

George Maddaloni, executive vice president and CTO of Mastercard operations, finds himself in such a position. Mastercard has built some of its own fintech AI to power its massive global B2B financial transaction network spanning 800 different products in 186 countries and 73 languages; more technology is purchased. We spoke with Maddaloni about how Mastercard makes these decisions and what large enterprise customers like his want to see from their cloud providers.

Let’s talk about the Mastercard AI project created to support customer experience.

George Maddaloni: Mastercard is a great network provider. We serve B2B clients, banks, financial institutions. Through these customers, we also serve consumers around the world. Therefore, we handle tons of calls from end consumers that we handle on behalf of our clients.

The application of AI in this tool was about routing these requests – how we handle them in the myriad ways they affect us. Since they come through different channels, how can we direct them to the right places and therefore make our contact centers much more effective? We created a routing engine, enabled by AI, that would allow us to get things to the right place and eliminate a lot of time spent sorting through request handling. It gives the customer time back to get an answer or resolve a problem.

So it’s more rules based and machine learning AI’s Generative AI.

Maddaloni: Generative AI was incorporated for language interpretation, because in some cases we were dealing with emails and needed to handle multiple languages. There was an element of rules and an AI model that allowed those rules to be applied. And then map all of that into a contact center that has both new and legacy components. We actually patented it in terms of applying both generative and AI techniques [other] Artificial intelligence techniques and applying them to an existing contact center infrastructure.

Describe the scope of your contact center operations.

Maddaloni: We have a part in the United States and a part in Asia. Whenever you get the size of a Mastercard, the technology will be in a state of “modernizing as you go.” So, there are many different components depending on the service and the call center that manages it.

What your technology vendors have done Not Has this kept you on the path of building and not buying?

Maddaloni: Every company will have its own nuances in terms of what it needs to do with the technology. We definitely still buy a lot of the commercially available features, without a doubt. But the types of challenges and requests we receive are B2C: in some cases, we receive B2B. Creating something that can interact between these really helps create a path [service] cases.

We have to support a number of languages, and we use this type of language for the business, in terms of credit cards, debit cards, transaction types, sometimes loyalty program requests we receive.

I would expect that when you talk to someone, they would say, “Yes, we’d like to use these technologies and put them together.”

But we need to have control over their actual application, especially when it comes to artificial intelligence, especially when you’re dealing with a customer. You need that context and there’s no way to buy it out of the box.

You patented this process. Will Mastercard enter the software business?

Maddaloni: We patent all the time in our industry just to make sure we protect the technology we currently have. We have a decades-long history of developing work with technology. Protecting our intellectual property is critical. So, we have a long-standing patent program. We are happy to have been able to apply for this too and we will see what comes of it.

What do AI technology providers need to develop to help B2B companies resolve customer service cases, which are less straightforward than B2C?

Maddaloni: You’re right about the complexity of what we’re trying to solve in the B2B market. So understanding how to route error calls and requests is critical. How you apply business language to how cases can be handled, because it becomes less about the customer and their transaction request at that time, and more about how you handle a case and the complexity associated with that activity. And you need that context and that information to be able to feed it back into the problem that you’re solving or the request that you’re about to fulfill.

Will generative AI make it easier to solve these more complex B2B questions with automation?

Maddaloni: I think the generative part of this is really about getting the language right and applicability: how can I incorporate it into a product of some kind.

We then used the generative capabilities in a product that we launched separately from the contact center called Decision Intelligence Pro, which goes beyond fraud prevention. Fraud prevention has always been based on artificial intelligence, but we have been able to use generative artificial intelligence techniques and neural network to reduce the time it takes to detect fraud – and detect more cases – which we integrate into the system [fraud] scoring service we offer at our [banking and financial services] clients.

This is an example where we can start creating products with generative AI techniques and make a service or product more effective.

What advice would you give to companies that are considering AI right now, so that they don’t buy something they’ll later regret?

The other big thing we’ve been thinking about from an AI perspective is using our data and technology responsibility principles in any new technology we’re bringing forward.

Giorgio MaddaloniExecutive Vice President and CTO of Operations, Mastercard

Maddaloni: I think the way we’ve approached technology decisions is to make sure we keep them in a layered approach. To handle all these calls, you need a contact center platform, you need a case management platform, you need those capabilities. As we thought about using AI in that context, we didn’t lock ourselves completely into one particular AI platform for a long period of time. We have our data layer, we have our call routing layer and certainly our AI approach on top of that, so we’re really making sure that we have this way of approaching our architecture to avoid sort of [AI vendor] block from which we could not escape very easily.

The other important thing we’ve thought about from an AI perspective is using our own Data and technology responsibility principleit’s in every new technology we’re bringing forward. One of the key principles is customer data privacy – rules that we need to adhere to and make sure we meet all these obligations, as well as eliminate bias in all the data we are consuming. All data principles are fundamental, especially these two.

When it comes to generative AI, what do you expect from technology providers that you haven’t seen yet?

Maddaloni: Over time it gets better and better. We’ve seen results where vendors have evaluated their AI capabilities against benchmark tests and are improving them over time. So certainly performance is one thing. And the second is controls: making sure we have the ability to apply our limits and data lineage techniques. As we run a global business, this is extremely important.

So while we talk to a lot of vendors about what they might want to do with a subscription for a data SaaS model, we really insist on those data principles that we have an obligation to our customers to meet. We make sure they incorporate them. Most of them are very willing to listen to us and take this into account. And they receive the same type of requests from other customers too.

These questions and answers have been edited for clarity and brevity.

Don Fluckinger is a senior news writer for TechTarget Editorial. He deals with customer experience, digital experience management and end-user IT. Do you have advice? Send him an email.

Source