A freelancer with irregular income asks Perplexity which checking account would suit them. The answer names a few providers with a reason for each, and the freelancer has a shortlist before opening a single bank's website. The bank that has held their savings for years may not be on it, and it has little direct say in how the answer describes the options.
That shift moves the first comparison into a place the bank does not control. A strong brand and good rankings help, yet the answer is assembled from many sources the bank does not run. To maintain its presence in these answers, the bank needs to measure visibility by product and platform and improve both its own pages and the outside sources AI systems use.
Customers are using AI answers to make the initial comparisons they used to make across several search results and comparison pages. McKinsey's survey of US consumers in August 2025 found that about half now intentionally use AI-powered search, and a majority of those users say it is the top digital source they use for buying decisions. Among AI search users, 44% called it their primary source of insight, ahead of traditional search at 31% and retailer or brand websites at 9%.
Financial services is part of this shift. McKinsey reports that around 40% to 55% of consumers in its top sectors, a group that includes financial services alongside categories such as travel and consumer electronics, use AI-based search to make purchasing decisions. The range covers all of those sectors together, so it indicates a broader trend rather than providing a banking-specific measure.
For a bank, the practical change is where the comparison happens. A parent who asks which bank offers a good savings account for a teenager gets a short list of named options and reasons, often before visiting any provider. The answer also depends on who is asking. New Media Advisors, a firm that studied more than 14,000 AI answers across 129 US markets and calls its findings directional, describes AI systems running a live web search, reading the pages they trust most and returning a shortlist. It found that answers change sharply when the question names a specific person, place and need. Much of the first conversation is now tailored to the customer, and it can happen before the bank knows the customer is looking.
Published measurements give a partial picture, and each is narrower than its headline suggests. EMARKETER's AI Visibility Index for financial services, which measures ChatGPT only, found Ally, an online bank, named most often for bank accounts in its first 2026 edition. An index of banks run by 5W, a communications firm, and published in The Financial Brand found that ChatGPT and Gemini favored Chase, Perplexity leaned toward Capital One, and Claude was the most balanced of the engines it tested.
Taken together, these studies describe a pattern more than a league table. In our reading, AI engines tend to name banks that are obvious choices for a product or whose information appears consistently across many sources. The banks named vary by engine.
For a CMO, a ranking of other banks is less useful than the same question asked about the bank's own business. For each priority product, in the markets that matter and on each AI platform, which banks are named, and why? That view shows where the bank competes and what it would take to be named, which a national ranking cannot.
Many customer questions are comparisons, and comparison sites are built to answer them. Conductor's benchmark of AI citations in the financial industry, based on data from 2025, found financial education and comparison sites winning the largest share, with NerdWallet at about 10% and Bankrate at about 8.5% of citations in its data. McKinsey found a similar pattern from a different angle. In industries such as financial services and consumer packaged goods, more than 65% of the sources AI search draws on are publishers, user-generated content and affiliate sites. Across industries, a brand's own sites often make up only 5% to 10% of those sources.
The same Conductor data shows the picture reversing when the question is about banks themselves. In its banks subindustry, bank domains led the citations, with TD and Chase at the top. For questions about which option is best, the sites that compare options led.
That split tells a bank where to put each kind of effort. Its own site should be the definitive source for everything only the bank can state with authority, such as current terms, eligibility, availability and how to apply. For comparison questions, the aim is an accurate and complete listing on the comparison sites and publications that AI systems already cite, since those pages will keep answering the question. For most banks, the practical goal is a correct and current description of their savings account on NerdWallet, rather than a higher ranking than NerdWallet on "best savings account".
AI search poses a real risk to whether customers consider the bank, though its effect on the number of new accounts is less certain today. Independent evidence on how AI answers affect bank acquisition specifically is still thin. The clearest estimates are cross-industry. McKinsey estimates that brands unprepared for AI search may see traffic from traditional search channels fall by anywhere from 20% to 50%.
If a bank is missing from the answer that forms a customer's shortlist, it may never get the visit that its product pages, rates and offers were built to convert. That loss is hard to see in analytics, because a customer who never arrives leaves no session behind.
A drop in traffic can also come from seasonality, rate changes, a site migration or a competitor's campaign, so a bank should rule those out before blaming AI. A more reliable way to judge the risk is to read the answers themselves. We suggest reading them the way an editor reads coverage of their own organization. Is the bank named and linked? How is it described, and where did the answer get its information? Our guide to measuring a bank's AI search visibility sets out how to read the gap between being named and being linked.
Nobody can put a reliable date on it. The Financial Brand ran the 5W piece under the headline "Your Bank Has Two Years, Max, to Become Visible to AI". It was written by the founder of the firm that produced the index, and the article presents no data behind that deadline.
What can be said with more confidence is that customer behavior has already moved, and the work a bank needs to do is slow to organize. Assigning an owner to each priority product's information, building compliance review into the content workflow, correcting descriptions on outside sites and building a measurement baseline all take quarters rather than weeks. In our view, the more useful deadline is how long this preparation takes inside the bank.
That argues for starting early rather than waiting for a full plan. Our article on strategy, ownership and operating model covers how to size the first cycle.
Large banks usually have credibility to show. The harder part is publishing product information clearly, on their own sites and in the sources AI systems read. For a CMO planning 2026, that turns into a few concrete decisions.
Start with ownership. One person in marketing should be accountable for AI visibility, with a named owner for each priority product who can get the product page, comparison listings and press wording changed. Our article on strategy, ownership and operating model covers how to divide the work.
Measure visibility by product and platform. Our guide to measuring a bank's AI search visibility explains how.
The budget should cover work on the bank's website and with the publishers and comparison sites that answer comparison questions. A plan that funds only the website leaves those outside sources out.
For a bank, CB/I Digital works on both sides of the answer. With the bank's partnerships team, we correct how comparison sites, partner listings and press material describe its priority products, and we make the bank's own pages the clearest source for the facts only the bank can state. Each change goes through the bank's compliance review before it is published.
Our SEO & AI Search services cover the listings work and the page fixes that go with it. If you want to know which outside sources are shaping answers about your products, we can show your team where they come from.
Do business banking customers use AI search the same way as consumers?
Most published research covers consumers, so there is less evidence for business banking. The retrieval mechanics are the same, and business owners ask detailed questions about cash handling, lending and treasury services that AI systems answer from the sources they can find. Measure business banking questions as a separate set rather than assuming consumer findings apply.
Does Google search still matter if customers are moving to AI answers?
Yes. Google's AI Overviews and AI Mode sit inside Google Search, so a page Google cannot index will not be linked from Google's AI features either. Our business case article explains why that overlap lowers the cost of the work.
Does being cited in an AI answer mean customers visit the bank's site?
Not necessarily. An AI answer can name a bank without a link or cite it without a click, and a customer may visit only after deciding. Search Console now reports impressions from Google's AI features in a separate report, though it does not show how the bank was described, so the bank has to read it alongside its own measurement of AI answers and its application data. Our guide to measuring a bank's AI search visibility covers what each source can and cannot show.
Copyright © 2026 CB/I DIGITAL INC. All rights reserved | Privacy Policy