Day two at Money20/20 moved the conversation from what AI can do to what it takes to make it work. On the main stages and in closed-door sessions, executives agreed that access to intelligence is no longer the hard part. What’s difficult is redesigning workflows, governing agents and earning customer trust.
The operating model is the advantage. One large bank’s AI leader put it plainly: “AI is everywhere, but value is not.” His bank’s AI assistant has handled one billion customer interactions in under three years, and AI-generated pre-meeting briefs for advisors have driven 40 percent more product sales and nearly twice as many referrals. He credited the operating model, a central strategy, a short list of high-impact use cases and a workforce trained to use the tools, rather than the technology. His comparison was the shift from steam to electric power, when the value only became apparent once workflows were redesigned around it.
AI has to connect across the enterprise. Credit union leaders described the same challenge from the perspective of a smaller institution. Putting AI into underwriting or the call center is straightforward. Aligning every tool across the organization with a regulator who asks how each model is backed is much harder. One innovation lab tested agentic AI with 16 credit unions and found more than 60 use cases. Its recommended first step is simple. Take inventory of the AI agents already in use, including those built into vendor platforms.
Consumers will let AI shop before they let it pay. One card issuer’s survey found 59 percent of eU.S. consumers are comfortable using AI for research and recommendations, but only 36 percent would let an agent complete the purchase. Card network research shared in another session put that figure at 23 percent. On a panel with a card issuer, a large retailer and a big tech wallet provider, the consensus was that trust comes from clarity and control.
Brands still matter. Nearly three in four consumers said shopping with a preferred merchant is important even when AI does the shopping, but AI will increasingly decide which brands get shown. One of the world’s largest retailers stressed the agent should facilitate the customer relationship, not replace it. One credit union CEO noted that few financial websites are built for AI agents to find and read, and urged peers to build for both humans and agents.
Payments are moving to the center of the relationship. A payments technology executive argued that every card transaction is now an opportunity to strengthen customer bonds. New industry research found that most institutions see their processing infrastructure as the biggest barrier to launching new products, and only 8 percent expect to stay competitive with their current setup over the next few years. Budget wasn’t the obstacle; fear and inertia were.
Smaller institutions are choosing a different race. Regional banks and credit unions were candid that they can’t outspend the largest banks. Their edge is collaboration, local relationships and member ownership, which frees them from quarterly earnings pressure. One example is a local influencer partnership that cost less than $4,500 drove a third of a credit union’s inbound traffic on launch day.
What it means for financial brands. Day one showed that trust is the new infrastructure. Day two showed that trust has to be built into how a company operates, then explained clearly to customers, regulators and, increasingly, the AI agents acting for them. Firms that can show how their AI is governed and what value it delivers will lead in the agentic AI era.
