Revolutionizing the Industry: Exploring AI in Payments With NMI Tech Leaders

Gen AI: A Conversation with NMI Experts

Greg Myers: Welcome back to NMI Payment Playbook. This is the second episode in our three-part series on trends affecting payment providers. Today, we’ll focus on AI and how it’s impacting the payments industry.

We’re joined by special guests Phillip Goericke and Spero Langaditis. Phillip, will you start by introducing yourself, your background and your role here at NMI?

Phillip Goericke: Sure. I’m the Chief Technology Officer here at NMI. I focus on advancing our technology strategy and enhancing operational efficiency across multiple areas. I lead the entire engineering group, ranging from our site reliability engineering team to our product engineering team, platform architecture group and agile operations team. I also oversee our security measures, ensuring that our payment solutions are innovative, robust and dependable.

Before joining NMI, I gained a lot of invaluable experience building and scaling complex systems. That background has been instrumental as we work to enhance NMI’s global payment platform, particularly through this new frontier where we’re integrating AI tech to streamline our operations and improve the customer experience and payments.

Myers: Spero, what about you?

Spero Langaditis: I’ve been with NMI for a little over seven years. I started off as an engineer; I’m an engineer at heart and always have been. Until recently, I was one of the directors of product engineering here at NMI. And then over the last year, I made the pivot into the AI space.

As NMI’s Director of AI and Automation, I’m essentially in charge of the AI strategy, both from an operational and strategic standpoint. That includes how we leverage AI internally to streamline certain processes and operations and, from a product standpoint, how we can better leverage AI technologies and embed them into our products.

Myers: AI is one of the biggest buzzwords I’ve heard recently; this topic comes up all the time. However, I think AI means different things to different people. Spero, can you give us a high-level overview of what “AI” really means?

Langaditis: So AI, or artificial intelligence, is a very broad term that, at the end of the day, depends on what your context is and what you’re doing. Generally speaking, AI refers to the simulation of human intelligence in machines.

For instance, AI is useful for things like learning, reasoning, problem solving and understanding natural language. AI as a broad term encompasses a range of technologies, which can include machine learning, natural language processing, computer vision and even robotics.

Myers: When people use the term “AI,” do they typically mean generative AI?

Langaditis: It really depends on who you are and what you’re doing with AI. Over the last two or three years, Gen AI really brought the technology into the media and made AI a popular subject. But to answer your question, not necessarily. AI, in and of itself, is a very broad term. Generative AI essentially refers to an AI tool or product that can create new content such as text, images, audio, etc. It’s probably the most popular subset of AI.

Myers: What’s the difference between machine learning and AI?

Langaditis: That’s a great question. AI is focused on creating systems capable of performing tasks that require human-like intelligence. At a very broad level, machine learning is a subset of AI. It typically involves training algorithms on certain types of data so they can make predictions and decisions based on that data without explicitly programming anything. So, in other words, machine learning is about systems learning from data and is a subset of AI.

AI at NMI

Myers: Phillip, from a technology perspective, how is AI being used at NMI today?

Goericke: At NMI, we’re embracing AI in a variety of ways, from enhancing our daily operations and making us more efficient to improving our customer interactions. We’re also beginning to embed it into our products. For example, we’ve used AI tools like ChatGPT, Github and Copilot within our engineering group to speed up our coding tasks, help us debug issues and save time on projects.

We’re also using AI to answer customer questions. Our chatbot, Penny, has led to a significant reduction in support calls and ticket volumes. We’re also in the process of automating a lot of our underwriting routines and certain basic underwriting decisions so our staff can focus on more complex and nuanced issues, and we’re evaluating how AI can make us more effective in those critical workflows.

Something I’m particularly excited about is the possibility of using AI to simplify API integrations. We’d like to make our APIs so simple that even a non-technical user (someone other than a developer) can interact with them. That’s directly supporting and improving our embedded payment solution, which has been shown to improve our customer experience and positively impact user loyalty and retention.

Myers: Spero, do you have anything to add to that?

Langaditis: To echo what Phillip said, internally, here at NMI, we’re approaching AI from two facets. One involves internal operational efficiency, like how we can start streamlining some of our internal processes to make them faster and more efficient. For that, our teams are starting to really use tools like ChatGPT, Copilot (for our engineers) and others.

Then, from a product and strategy perspective, we’re starting to look into introducing Gen AI technology into some of our products. One of our main focuses is making our products easier to use by decreasing the initial learning curve that users face when adopting new software.

Considerations for Adopting AI

Myers: Phillip, how would you advise other payments professionals to approach AI? What kinds of conversations are you having with NMI customers and stakeholders?

Goericke: We tend to discuss how AI can reduce operational costs, enhance security and create new revenue opportunities. Frankly, if you’re not using and leveraging AI yet, it’s not too late. But, if you don’t start soon, you might be left behind by the competition.

I encourage others to be thoughtful and considerate about how they secure their data. It’s important to develop robust AI usage policies and limit how data can be used with AI. There are a myriad of concerns on this frontier — especially with regard to sensitive data. For instance, it’s important to consider your specific data privacy needs, and understand the differences between open-source and closed-source AI solutions. If your use case involves training an AI model on sensitive data, a closed-source option like ChatGPT may not make sense, and you might consider using an open-source LLM (large language model) that runs in your environment without  worrying about critical data leaving your network.

All in all, it’s really about leveraging AI to not just make your processes better and faster but also unlock new customer insights and engagement strategies.

Langaditis: To add to Phillip’s point, one of the biggest pieces of advice I have is to invest in AI capabilities internally. Make sure your employees are educated and trained on how to use AI effectively, whether that involves training engineering to use better prompts or teaching best practices around how to use tools like ChatGPT, Claude or Google Gemini. Education is paramount.

It’s also important to keep ethical considerations in mind. Make sure your employees adopt internal ethical AI practices to ensure transparency, fairness and accountability — especially when interacting with or using third-party systems.

The Future of AI in Payments

Myers: Let’s talk about the future of AI. Where is AI heading? Can you provide any specific insights into AI in the payments industry?

Langaditis: Within the payments industry, I think there’s going to be large strides in using AI for things like fraud prevention and risk management. That could include real-time analysis of transaction patterns to detect and prevent fraudulent activities or using AI to manage credit risk by analyzing vast amounts of data. 

Other use cases include offering personalized customer service. For instance, here at NMI, we have an AI-powered support bot that can help our customers resolve issues faster. AI will also be useful in really gauging and developing customer insights and analytics that we may not have had access to before. So providing deep insights into customer behavior to improve satisfaction and tailor offerings based on that data. Being able to extract data and analyze it with AI tooling will be a powerful use case within the payments industry.

Myers: Phillip, what are your thoughts on the future of AI?

Goericke: Well, AI could advance in a million different ways that none of us can predict. But I really do think that it’s set to revolutionize payments in a lot of areas. 

Payments is a highly complex ecosystem that has evolved over decades. It’s very complicated and requires a ton of knowledge to understand how everything is interconnected and works together. So, my bet is that AI will simplify payments, make integrations more accessible, and enable seamless financial operations in everyday business tools. 

Previously, developers would have to manually code and integrate their apps to other systems. When the associated API changed, they had to update it and maintain it. In the future, I believe that AI will lower the barrier of entry for those looking to embed payments into software by automatically and dynamically updating APIs so developers don’t have to manually code or maintain those integrations.

Myers: Phillip, would you mind summarizing what we’ve spoken about for our readers?

Goericke: Sure. So, at NMI, we’re actively leveraging AI in a variety of ways, such as enhancing operational efficiency, improving customer service, and enabling partners to embed payments more easily than ever. As we look into the future, the integration of AI with payments systems really holds a tremendous potential to transform the entire landscape. It has the potential to make payment processing, which is historically very complex, very intuitive, simple and accessible within the services and systems that customers use every day.

It’s important to remember that security with AI has to stay front and center; you need to have robust usage policies. Consider open versus closed-source AI models depending on the level of control you need and the type of data sensitivity you have. But software, in general, is moving towards simpler interactions with deeper, more dynamic integrations in systems all over the world. If you aren’t already, you need to start using these tools within your business. I personally recommend that you have a dedicated person or team that is exploring use cases and advancing how you use AI.

Tune in to the full episode here.