Insights
Why a bank’s AI visibility score comes back low.
When an institution scores poorly, the cause is usually structural rather than editorial. These are the four we see most often.
An AI visibility score is a summary of a simple exercise: ask ChatGPT, Perplexity, Google AI Overviews the questions your customers ask, and record whether your institution is named. A low score means the assistants had little to work with, or had reason to name somebody else.
Nothing on the page states the obvious
Assistants answer from what a page says plainly. Rates, locations, product names and eligibility often live in an image, a slider or a PDF, where a reader that only has text finds nothing to quote.
No structured data
Schema markup is how a page tells a machine what kind of thing it is describing. Without it, an assistant is guessing from prose — and it guesses conservatively, which usually means leaving you out.
AI crawlers are excluded
Plenty of robots.txt files were written when the only crawlers that mattered were search engines. If the assistants are disallowed, the rest of the work cannot help.
The site answers questions nobody asked
Pages are often organised around internal product naming rather than the question a customer would type. An assistant matches the question, not the brochure.
None of these are quick wins in the marketing sense; they are ordinary site work. The useful part is knowing which of them is actually costing you, which is what a benchmarked reading is for.
How we measure it →