Yellow Card co-founder and CTO Justin Poiroux shares how the evolution of fraud detection, digital ecosystem requirements and emerging protocols are compelling organisations to build their own AI capabilities rather than rely on external providers.
The technology landscape is experiencing a fundamental shift as organisations increasingly choose to develop internal artificial intelligence (AI) capabilities rather than depend on third-party software-as-a-service solutions. According to Justin Poiroux, co-founder and chief technology officer of Yellow Card, this movement towards self-reliant AI infrastructure is becoming a strategic imperative for organisations serious about maintaining competitive advantage and security.
The first major catalyst pushing organisations towards in-house AI development is the evolution of fraud detection requirements, as traditional approaches to combating financial crime are proving increasingly inadequate against ever-sophisticated threats.
“The deployment of specialised fraud AI agents has fundamentally changed how we approach threat mitigation,” explains Justin. “These systems require extensive training in institutional Anti-Money Laundering frameworks, Know Your Transaction protocols, and comprehensive fraud prevention methodologies that simply cannot be effectively delivered through generic SaaS platforms.”
Unlike previous technologies confined to narrow functions, modern AI agents serve as specialised virtual experts, seamlessly integrating into specific domains.
This specialisation requirement means that organisations relying on external providers often find themselves constrained by one-size-fits-all solutions that lack an understanding of their specific risk profiles and operations.
Yellow Card, for example, has developed AI-driven chatbots to support customers around the clock, while biometric and identity verification systems leverage AI to improve efficiency and reduce friction in the user experience.
Balancing collaboration with data protection
The second trend stems from the growing need for secure, scalable digital ecosystem participation. As organisations become increasingly interconnected, they must balance collaboration with data protection.
“Strategic partnerships across fintech, AI, and infrastructure sectors are essential for bridging innovation and accessibility gaps,” notes Justin. “However, true collaboration requires maintaining control over sensitive intellectual property and customer data while still enabling meaningful ecosystem participation.”
The biggest concern is protecting company secrets. When using external AI services, sensitive data, customer information and trade secrets could end up in the wrong hands. Building your own AI systems lets companies work with partners and embrace new ideas while keeping their data safe and maintaining their competitive edge.
New tech requires systems that work together
The third factor accelerating the move towards internal AI infrastructure is the emergence of advanced integration protocols, which require deep system-level implementation. This includes technologies like Model Context Protocol (MCP) and Agent2Agent Protocol, which are far more than simple chatbots.
“These protocols are designed to facilitate comprehensive AI integration into existing platforms, essentially replacing backend systems and automating entire workflows,” explains Justin. “This level of integration cannot be achieved through surface-level SaaS connections – it requires fundamental architectural alignment.”
The implications go beyond admin functions, where AI agents will assume responsibilities currently handled by humans. This transition demands first-hand knowledge of internal processes, legacy system dependencies and workflows that external providers can’t address.
“The barrier to entry for building in-house AI infrastructure continues to decrease,” observes Justin. Companies that build their own AI systems can take full advantage of these new technologies, giving them a real edge over competitors. They will be better placed to seize new opportunities, manage costs and keep the security and control they need to grow sustainably in a complex market.
