Banking · 17 November 2026
Data silos weaken banks' AI edge as customers turn to AI assistants
A McKinsey report found nearly three in four consumers using AI for banking tasks choose ChatGPT, Claude or Gemini, almost double the share using a bank's chatbot.
What happened
Consumers are increasingly turning to general-purpose AI assistants rather than their own banks for help with financial tasks. A McKinsey report found that nearly three in four consumers who use AI for banking tasks choose ChatGPT, Claude or Gemini, a share almost double that of customers who use a bank's chatbot.
CIO Dive, which reported on the findings, placed them within a broader warning: data silos are limiting banks' ability to turn their AI investments into a competitive edge. The headline framing suggests that the problem is less about whether banks have AI tools and more about whether those tools can draw on the joined-up data needed to serve customers well.
Why it matters
The finding points to a shift in where customers start their AI journeys. When people reach for a widely used assistant first, that tool sets the bar for what a helpful, fast and conversational answer looks like. A bank's chatbot is then judged against that benchmark, often within the same session, and it has to work harder to earn trust and repeat use.
For leaders in digital transformation, the underlying issue is architecture. An assistant can only personalise a response if it can see the relevant account history, product holdings, service interactions and preferences in one place. Where that information is split across legacy systems, siloed channels and separate teams, even well-built AI features will struggle to feel coherent, which is precisely the competitive risk the coverage highlights.
By the numbers
- Nearly three in four consumers using AI for banking tasks choose ChatGPT, Claude or Gemini.
- Almost double the share of those using a bank's chatbot for the same purposes.
The Renascence take
Most commentary will treat this as a contest between banks and AI labs over customer attention. We think the more useful reading is that customers are telling banks which experience standard they now expect, and most institutions are not yet built to meet it.
The behavioural principle at work is simple: people follow the path of least friction, and a general-purpose assistant that answers without asking them to log in, navigate menus or repeat themselves is almost always the path of least friction. Banks that respond by bolting another chatbot onto the same fragmented data estate will get the same result. A customer-obsessed operator should start by mapping the top ten banking tasks customers already attempt with outside AI tools, then fix the data joins those tasks depend on before investing in another interface.
Sources
This briefing was written by our Newsdesk, synthesising reporting from the outlets below. Follow the links for the original coverage.
FAQ
Questions we get on this topic
More in Banking
Stay ahead of CX
Get the signal, not the noise.
The stories shaping customer experience — plus the Journal and Experience Loom — in your inbox.


