A landmark finding buried inside a recent McKinsey report should be keeping bank executives awake at night: nearly three in four consumers who turn to artificial intelligence for banking-related tasks are reaching for ChatGPT, Claude, or Gemini — not their bank's own chatbot. That figure represents almost double the proportion of users who rely on a financial institution's proprietary AI tool, and it encapsulates one of the most consequential competitive failures unfolding in modern banking.
The numbers are stark, but the underlying dynamic is even more sobering. Banks have spent years and hundreds of millions of dollars building digital infrastructure, compliance frameworks, and customer-facing technology. Yet when consumers need help navigating a financial decision — whether that means understanding a loan product, reconciling account activity, or planning a budget — roughly 75% are bypassing that infrastructure entirely and opening a third-party application built by a technology company with no banking license, no fiduciary obligation, and no access to their actual account data. The market has rendered its verdict, and banks are losing.
The Data Silo Problem at the Heart of the Crisis
The root cause of this exodus is not a lack of investment or intention. Most major retail banks have launched AI initiatives, deployed chatbot interfaces, and issued press releases heralding their digital transformation credentials. The fundamental problem is structural: data silos. Decades of legacy system architecture mean that customer information — transaction histories, credit behavior, product holdings, service interactions — remains fragmented across incompatible platforms, unable to flow freely into the kind of unified intelligence layer that would make a bank's AI genuinely useful.
When a consumer asks ChatGPT a banking question, the experience is fluid, conversational, and contextually rich — even without access to real account data. The model draws on vast training sets and sophisticated reasoning to deliver responses that feel authoritative and helpful. When that same consumer asks their bank's chatbot a comparable question, they are frequently met with scripted deflections, limited query recognition, or a prompt to call a branch. The experience gap is not marginal; it is vast, and consumers have noticed.
Time Is Running Out for Proprietary AI to Matter
What makes the McKinsey findings particularly alarming for the industry is the velocity of entrenchment. Consumer habits formed around third-party AI tools are not easily reversed. Claude, Gemini, and ChatGPT are iterating at a pace no bank technology team can realistically match, and each successive improvement deepens user loyalty to those platforms rather than to the bank's own digital touchpoints. If banks do not close the experience gap now — not in the next product cycle, not after the next core system migration — they risk becoming the data custodian behind someone else's customer relationship.
That scenario carries profound implications for revenue. The customer relationship has always been banking's most durable moat. Banks that understand their customers' financial lives in granular detail can cross-sell products, price risk accurately, and anticipate needs before competitors do. If a third-party AI platform becomes the primary interface through which a customer interprets and acts on their financial life, the bank is demoted from trusted adviser to utility pipe — still essential for holding deposits and processing payments, but stripped of the relationship depth that generates fee income, lending revenue, and long-term loyalty.
What Banks Must Do — and Quickly
Solving the data silo problem is not a technology challenge alone; it is a governance and organizational priority. Banks must urgently invest in unified data architectures — often described as data mesh or data lakehouse frameworks — that allow customer information to be accessed, permissioned, and deployed across AI systems in real time. Without that foundation, any AI product a bank launches will remain a diminished version of what third-party tools already offer, because those tools at least benefit from coherent design even without proprietary financial data.
Some institutions are pursuing partnership models, integrating large language model capabilities from established providers rather than building from scratch. Others are exploring consent-driven data-sharing arrangements that could eventually allow third-party AI platforms to surface bank-verified financial context, turning a competitive threat into a distribution opportunity. Neither path is without risk — regulatory scrutiny around AI explainability, data privacy under frameworks such as the Consumer Financial Protection Bureau's open banking rules, and cybersecurity exposure all complicate any strategy that moves customer data more fluidly through systems.
Yet the alternative — continuing to invest in isolated chatbot experiences that capture barely half the market share of consumer-grade AI tools — is demonstrably not working. The McKinsey data makes clear that the window for banks to establish AI credibility with their own customer base is narrowing. Institutions that treat this as an infrastructure procurement decision rather than a strategic emergency are misreading the moment entirely.
What This Means for the Industry
The 75% figure is not merely a product metric. It is a signal that the financial services industry is experiencing a slow-motion disintermediation of one of its most valuable assets: the advisory relationship. Banks that move decisively to dismantle internal data silos, invest in AI experiences that leverage proprietary customer context, and close the usability gap with consumer AI platforms stand a genuine chance of reclaiming that relationship. Those that delay face a future in which their brand is present on the account statement but absent from the moment that actually shapes financial decisions. In the AI era, that is a precarious place to be.
Written by the editorial team — independent journalism powered by Codego Press.