Visa didn’t just dabble in AI—they rebuilt their house around it. With over 2,500 engineers focused solely on AI and more than 100 internal AI apps in production, the company’s approach spans everything from software development to customer dispute resolution. AI is part of the operating system at Visa—literally. Even before large language models became cocktail-party conversation, Visa was preparing for this shift. Every employee now has access to generative AI tools, and the company was among the first to adopt Microsoft’s Copilot enterprise-wide.
The company’s AI roadmap started over a decade ago with a bold bet: rewrite the data platform to juice predictive models and lay the foundation for neural networks long before “transformer” meant anything beyond a type of toy. When ChatGPT crashed into the zeitgeist, Visa didn’t scramble—they accelerated. By early 2023, generative AI was infused across daily workflows. Now they’re looking ahead to “agentic commerce,” giving AI the tools (plus guardrails) to act on users’ behalf—think intelligent assistants that can buy stuff for you, securely.
All of this sits atop a hefty $3.5 billion investment in AI infrastructure, including an in-house “AI observatory” that monitors every model in production. This internal dashboard lets execs track risks, compliance, and model drift in real time. Visa’s careful dance between algorithmic horsepower and policy boundaries highlights a bigger trend: the companies winning with AI are getting good at balancing innovation with governance.
Full story at Fortune.