Mer. Feb 12th, 2025

With the recent launch of DeepSeek, GenAI has hit the headlines again.Mastercard has adopted a structured approach to informed, perceptive, and proactive AIIt seems that AI is following a well-trodden trajectory. Computing history confirms that key principles such as ‘costs decreasing while compute increases’ still hold true.Yet, despite the excitement, concrete examples of how financial services firms use large language models (LLMs) and embrace the potential of GenAI remain sparse.With this in mind, Mastercard’s latest report, GenAI Rising Superpowers, caught my eye. In it, the company introduces a framework for understanding GenAI’s evolution, categorising it into three types: informed, perceptive, and proactive AI. This provides a neat, structured way to assess AI’s impact on financial services.Informed AI is the most mature and widely used. It relies on LLMs trained with specific knowledge and real-time data to generate relevant insights, enhancing decision-making.
Perceptive AI, or multi-modal AI, is in its early stages. It integrates various inputs—audio, visual, and sensory data—opening new possibilities for advanced customer service and marketing.
Proactive AI, also known as agentic AI, represents the cutting edge. It has the potential to reason, make decisions, and act semi-autonomously. Or, in the future, fully autonomously.
This framework resonated with me, and I was interested in learning more, so I contacted Ken Moore, Mastercard’s Chief Innovation Officer. Our exchange provided a fascinating look into how the company is leveraging AI today and where it sees the technology heading next.Mastercard’s ‘north star’ is powering economies and empowering people, and it sees AI as a key enabler. AI has been integral to Mastercard for over two decades, securing and intelligently managing its network to protect over 143 billion transactions annually.According to Moore, Mastercard is now using its AI expertise to ensure safer, smarter, and more personalised commerce and strengthen its internal operations.“Traditional AI excels at processing structured data, enabling applications like fraud detection and personalised experiences,” he explains. “Generative AI allows us to take this further, enhancing existing products while unlocking new possibilities – ranging from software development and implementation to customer support, sales, and service delivery.”Talking to Ken, you can see that with a strong foundation of AI engineers and data scientists, Mastercard is well-positioned to harness GenAI’s potential for refining processes and creating new products.I asked Ken which of the three AI types – informed, perceptive, and proactive – would be most relevant to Mastercard in 2025. He answered, “All three matter, but their readiness and scale vary.”Mastercard already employs informed AI extensively. One standout example is Decision Intelligence Pro, which builds on the existing Decision Intelligence (DI) solu