More of your future members will never run a search. They’ll ask an assistant and act on the answer. Plenty of tools measure how often ChatGPT mentions you. We’ve also run 90+ GEO experiments to learn what actually raises it, so we can grow your mentions instead of guessing.
After its GEO experimentMentioned in 71.43% of answers
BeforeMentioned in 0% of answers
Pearl Hawaii Federal Credit Union’s GEO experiment, ChatGPT answers
On the question of the best CD rates in Oʻahu, Pearl Hawaii Federal Credit Union went from mentioned in 0% of answers to 71.43% after restructuring its CD pages.
Why it matters now
Search gives you a results page. AI just gives you the answer.
This isn’t search engine optimization with a new name. The mechanics of being chosen are different, and the penalty for being left out is steeper.
The question moved.
People who used to open ten tabs now ask one assistant and act on the answer. If you’re not in that answer, you were never in the running.
There’s no page two.
A search engine gave ten institutions a chance. An AI answer typically mentions two or three, so the gap between being mentioned and being invisible is enormous.
Early positions harden.
The sources assistants trust are slow to change. Institutions that become the answer now are harder to displace later.
What’s in the suite
Measure it. Then actually move it.
Plenty of tools will give you an AI visibility score. A score you can’t act on is trivia, so this suite runs all the way from measurement to a published page.
Track
Where you get mentioned
How often ChatGPT mentions your institution, overall and for the specific questions your members actually ask.
Rank
Who gets mentioned instead
A ranked list of the institutions ChatGPT mentions in your market, and where you fall on it.
Segment
By market and product
Grouped by branch market and by product, so you can see that you win on checking and vanish on mortgages.
Diagnose
Which pages get cited
The sources an answer draws on when it mentions someone else, which is usually the clearest picture of what your site is missing.
Traffic
Visits from AI assistants
How often AI assistants and their crawlers visit your site, from ChatGPT to Claude and Perplexity.
Publish
The content that changes it
Content aimed at the questions you can win, published through the same platform that builds your site, then measured again.
AI assistant
Which credit union is best for a used car loan in [your county]?
Share of answers mentioning Pearl Hawaii: “Best CD rates in Oʻahu”
Before0%
After71.43%
1Ask2Rank3Cite4Publish5Measure
Step 1 of 5
Ask what your members ask. Naming your market.
Questions are written the way a person asks an assistant, and each one names its town, city or county, because an assistant only sees the words of the question. Every one is backed by measured search demand.
Step 2 of 5
See who gets mentioned. And who doesn’t.
The same questions run again and again, because a single AI answer is noise. You see the share of answers that mention you, and every institution mentioned instead, ranked.
Step 3 of 5
See which pages get cited. That’s the gap.
When an answer mentions someone else, the platform records the page it drew on. That is usually the clearest picture of what your own site is missing.
Step 4 of 5
Publish what’s missing. Checked four ways.
A page aimed at a question you can win is written from what your live site already says, and checked for writing quality, compliance, accessibility and facts. Your team signs off.
Step 5 of 5
Measure again. Against where you started.
The same questions run for a set window after the page is live, compared with the baseline. We’ve run 90+ GEO experiments this way, so you know what actually raised your mentions.
Results
Others measure and guess. We measure and test.
Lots of tools track AI mentions. Few run a disciplined experimentation program to find out, statistically, what raises them and what doesn’t. That’s what our GEO experiments do. A few of the results:
0% → 27.38%
Pearl Hawaii FCU
Mentions across six tracked questions about Oʻahu deposit products, over three experiments.
Each one measured against a baseline taken before the work, so every new engagement starts from what has been proven to raise mentions.
Go deeper
Your site has two audiences now. People, and their AI assistants.
Getting mentioned comes from publishing the content assistants draw on, written so they can read it: facts stated plainly, rates kept current and dated, and the real question answered early.
We measure mentions on ChatGPT because it has by far the most volume. It’s like SEO: most teams focus on Google results and never give Bing a thought. Visits from AI assistants are a separate measure, and those come from several assistants.
Is this different from SEO?
While there is some overlap, a page can rank well on Google and still never be mentioned by ChatGPT. SEO competes for a position in a list of links, and GEO competes for a place in one written answer that mentions a few institutions. Different goals call for different tools and methods.
Do we need a new website first?
No. Most of the gains, by far, have come from publishing new long-form, text-based content aimed at the questions people ask, and that works on the website you already have.
Let’s talk
Find out who AI mentions instead of you.
We’ll run the questions your future members ask in your market and show you which institutions get mentioned, and whether you’re one of them.