People tell assistants about money before they tell anyone else, and banking questions follow: where a first paycheck should live, which bank suits a new business, whether to switch after a fee. The answers name institutions. Here is what puts a bank into them.
Banking moments are life-shaped: the first real job and the first real account, the new business needing banking that will not fight it, the move to a new city, the fee that finally exhausted patience, the savings goal hunting for yield. Askers include the constraint that matters to them, no fees, branch nearby, good app, and the assistant matches names to constraints.
For bank recommendations, assistants verify against regulator standing and deposit insurance, rate and fee data as published and as aggregated by comparison sites, review sentiment about service and apps, press coverage, and bank pages that state products plainly. Rates are checkable facts and assistants treat them as such; service reputation arrives through reviewers and gets repeated in framing.
The banks that keep appearing publish plain, current product facts and are corroborated on the comparison sites assistants cite; credit unions and regionals win local and fee-sensitive moments over national brands regularly, because the constraint was local or cost, and their evidence fits it. Contradictory rate or fee data across pages and aggregators quietly disqualifies, precision is the entry fee in this category.
The test costs nothing: ask ChatGPT, Gemini, Claude and Perplexity the questions above, phrased the way a real person would say them, and write down who gets named and what sources appear. Run each question twice on different days, since answers vary and patterns matter more than single runs.
For bank marketers, the fixes: product pages with current, plain facts, aggressive consistency between your site and the aggregators, and attention to app-store and service reviews, because their language becomes your framing. For banks that want the systematic version, the free Aethon audit runs your real buyer questions across all four assistants with screenshots, and the banking playbook turns the gaps into a work plan.
By matching stated constraints, fees, rates, location, app quality, to verifiable data: regulator standing, published rates, comparison-site aggregation, and service reviews. Checkable facts dominate; slogans contribute nothing.
No. Answer placement is earned through verifiable sources. Advertising exists separately and is labeled; it does not write the recommendation.
When the asker's constraint is fees, rates or local service, the credit union's citable evidence often matches better, and assistants score the match, not the marketing budget.
Current, plain product facts on parseable pages, perfect consistency with rate aggregators, and service-review hygiene. Then monthly measurement across the four assistants; the free Aethon audit provides the baseline.