At the recent PASA International Payments Conference (PIPC) Electrum’s Head of Commercial, Anton van Zyl took to the stage with AI expert Nick Pentreath to explore the interplay between trust, rising transaction volumes, and developments in AI. Here are some of the key takeaways from their talk.
While real-time capabilities have shaped the industry and become the baseline for payments today, the true challenge lies beyond technology; it is about delivering instant payments that users trust implicitly. Meanwhile, as AI technologies emerge and evolve, the question of how they will fundamentally reshape payments is still unfolding.
The infrastructure and instant availability of real-time payments already exist, so success is no longer about technological feasibility. Instead, it depends on addressing trust in the system as a whole as transaction volumes skyrocket.
South Africa is set to process 600 million real-time payments this year. As these volumes rise, the need for reliability and customer trust in digital payments rises with them.
“In a digital system, failure is visible. In a real-time system at exponential scale - with tens of millions of South Africans depending on these systems every day - failure is systemic trust erosion.”
In a cash economy, transaction failures often present as momentary inconveniences. However, in a digital, real-time system, failures become visible and deeply damaging to consumer confidence. Scaling trust thus requires scaling capacity while rethinking infrastructure to meet the demands of an instant, AI-driven, hyper-connected payments ecosystem.
Leading global technology companies like Amazon and Netflix generate significant revenue through AI-powered recommendations and personalisation. Payments leaders such as Stripe are harnessing large language models trained on billions of payment transactions to enhance fraud detection by over 60%. This achieves efficiency gains in months that previously took years. This application of AI at scale is not just about cost reduction but about fundamentally improving performance, security, and customer experience in real time.
However, the path to AI success is complex. Studies show that 95% of generative AI projects fail, often due to focusing on irrelevant use cases, poor data quality, lack of governance, and one-size-fits-all solutions.
Real-time data capabilities, data consolidation, governance, and compliance are mission-critical foundations for AI readiness in payments systems.
Integrating AI also means addressing ethical challenges like algorithmic bias and decision transparency. A vigilant governance approach combining diverse data sets, continuous monitoring, and human oversight must be a foundational design principle. This ensures AI-powered payments systems not only function effectively but also do so fairly and transparently, cementing public trust.
“What if we miss the boat as an industry? What would the cost of not leveraging the latest fundamental technological evolutions be, with the future state in mind?”
For many banks, legacy systems built on incremental and siloed architectures slow AI adoption. Transitioning to “AI-first” infrastructure does not require rip-and-replace changes overnight. It rather involves strategic steps such as identifying high-value use cases to prototype, building centralised data layers accessible via APIs, and fostering a culture open to experimentation and agile learning.
Collaboration across the payments industry to establish open standards and interoperability will be crucial to accelerating innovation across the entire ecosystem and ultimately bringing value - and trust - to the end user.
Looking ahead, scaling trust and capability to support billions of real-time transactions is a once-in-a-generation opportunity. It requires courage to reimagine payments infrastructure from the future back. The convergence of evolving regulation, technological advancements, data availability, and rising customer expectations creates a powerful platform to redefine the future of payments.
“Real-time isn’t a tech problem anymore. Everyone agrees instant is the baseline, and the infrastructure exists. What’s the constraint? Institutional courage.”
As the payments industry navigates this pivotal transition, embedding augmented intelligence from the ground up and amplifying human expertise rather than replacing it will be key.