From 0 to 100%: How Three Credit Unions Increased Their AI Search Visibility
- Derik Krauss
- Oct 9
- 7 min read
Guest Editorial by Derik Krauss, Co-Founder, MetriFi
Today, many website visitors aren't people. They're AI agents.
Over a billion people are using AI tools like ChatGPT, Google AI Overviews, and Perplexity to search the internet. As Google searches decline and AI searches grow in popularity, this changes how people find and interact with your brand online. The shift toward the Agentic Web has major implications for credit union website design. Credit unions need to show up in AI responses or risk becoming invisible to billions of AI users.
My AI Visibility Research and Hypothesis
I recently published groundbreaking AI visibility research for the credit union industry, based on an analysis of 7,500 AI search responses. The goal of my analysis was to understand what causes brands to show up in AI responses. I wanted to identify patterns, trends, and actionable strategies for Generative Engine Optimization (GEO).
From my research, what stood out most is that AI tends to reference detailed, objective web pages that are highly relevant to a user's prompt. (Prompts are messages users send to AI, kind of like keywords in Google search.) This conclusion became a basis for a hypothesis that I wanted to put to the test.
My hypothesis: Publishing long-form, authoritative content that directly answers specific prompts will increase your AI visibility.
Testing My Hypothesis in the Real World
Working with my team, we recruited several credit unions to run tests with us, including the three featured in this article: NW Preferred FCU, Lone Star CU, and HFS FCU. The objective of the tests was straightforward:
Increase AI visibility for one prompt that currently has low visibility.
We believed that publishing authoritative, targeted content would boost their visibility. Below are the results of our tests. Each credit union saw significant increases in visibility for their selected prompts.
Case Study #1 – Northwest Preferred Credit Union
One test was conducted with Northwest Preferred Federal Credit Union (NWP). They chose to focus on the prompt, "Who has the best high yield savings account in Stayton, Oregon?"
Before taking any actions, we used a Generative Engine Optimization (GEO) tool to send the prompt to AI 25 times. NWP appeared in the responses 0% of the time. AI consistently listed competitors like Axos, Lending Club, Rising Bank, and Maps Credit Union, but never NWP.
Then, the credit union worked with the same GEO tool (plus ChatGPT) to create and edit a targeted article called "Who has the best high yield savings account in Stayton, Oregon? NW Preferred Federal Credit Union". The article is a thorough, 3,600-word piece of content that gives factual details about NWP's high-yield savings accounts and addresses the exact wording of the prompt.
NWP published the article as a page on their website and made sure it was indexed by Google. Then, we reran their AI visibility for the prompt and the results were clear:
NWP FCU went from 0% to 100% visibility for their targeted prompt.
Importantly, the article we published began appearing in AI-generated responses for the targeted prompt—a clear signal that the strategy was working. This demonstrated not only improved visibility but also validation that the content was being recognized as authoritative by AI systems.
Subsequently, NWP published another piece of content on their website: a blog post that compared their high-yield savings accounts to offerings from their competitors.
As a result of publishing these two pieces of content, they not only saw an increase in visibility for their selected prompt, they also saw increases across several other related prompts.
Case Study #2 – Lone Star Credit Union
A similar test was conducted with Lone Star Credit Union, this time targeting the prompt, "Where in East Texas could I refinance my auto loan?"
Before any changes, Lone Star’s visibility for this prompt was 0%.
Similar to the process with NWP, we used a GEO tool and ChatGPT to draft a detailed, long-form article. The credit union edited it and published it on their website.
A few days after publication, we reran the prompt and their AI visibility had increased to 100%.
(Note: During the testing period with Lone Star CU, the agent search tool changed from Google to OpenAI's native search, which may have impacted the results. Even so, visibility rose from 0% to 100% and we can see in the data that the new content is being featured by AI in responses, indicating that the new content is the most likely driver of the improvement in visibility for the prompt, “Where in East Texas could I refinance my auto loan?”)
Case Study #3 – HFS Federal Credit Union
We also applied this process with HFS FCU. The targeted prompt was: “Where in Kona can I find the best credit cards?”
Before publishing new content, HFS had 0% visibility for this prompt. We used the same GEO approach as in earlier tests, creating a long-form article specifically targeted to answer this prompt. After HFS edited and published the article on their website, visibility jumped to 100% between August 26th and September 2nd. We also confirmed that the new article is being surfaced in AI responses.
Looking at the full period from August 1 to September 2nd, the prompt’s overall visibility averaged 88.89%. Even with some natural fluctuations, this shows that the article consistently positioned HFS as a top answer for credit card searches in Kona.
These results for HFS, alongside the successes at NWP and Lone Star, provides further evidence that publishing thorough, prompt-specific content can reliably move a credit union from invisible to visible in AI-generated answers.
What These Wins Have in Common
Across all three credit unions—NWP, Lone Star, and HFS—the winning formula was consistent. One factor stood out as the clear, primary driver of success:
AI visibility increased substantially after publishing long-form, authoritative content that directly answered the selected prompts. After publication, AI began referencing the new content in its responses.
While the primary driver of increased visibility was targeted, long-form content, two other practices may have contributed to the results:
Clearly formatting content: The articles are structured with clear headings, lists, and Q&A sections to make it easier for AI to parse.
Ensuring indexing: We ensured the new pages were indexed with Google and strategically linked within the site to increase the chances the content could be easily found by search engines.
The results show that publishing authoritative content can move a credit union from invisible to visible in AI-generated answers. Therefore, this should be a focus of Generative Engine Optimization.
Why This Matters for Credit Union Website Design
Credit unions are now competing for AI’s attention as much as for human clicks. When planning a credit union website design, AI visibility should be a core consideration, not an afterthought.
Based on our research, here’s a strategy for making your website AI-compatible:
Rather than trying to make pages designed for humans also work for AI, it may be more effective and efficient to create dedicated pages for AI.
Your website can have two types of pages built for different purposes:
Human pages: Pages built for humans that are easy to scan and visually appealing.
AI pages: Pages built for AI that are text-based, objective, and extremely thorough.
Not only does this approach better align with the different use cases for these pages, it also makes Generative Engine Optimization more straightforward, since it only requires adding new targeted content instead of redesigning existing pages. While AI visibility is a new field that’s still developing, early results suggest that this is an effective and scalable strategy.
The Takeaway
My initial research provided a hypothesis for how Generative Engine Optimization could help credit unions gain visibility in AI search results. The tests with NWP FCU, Lone Star CU, and HFS FCU are early signs that the strategy is working. As we continue our testing, further insights will come to light.
In the new AI-first era of search, visibility is earned by providing AI with authoritative information, especially when it answers a specific question. The sooner credit union website designs adopt GEO practices, the better positioned they will be to own key answers in their markets.
For credit unions planning a website redesign, start by measuring your AI visibility. If you’d like advice on getting started, feel free to reach out to me directly at dk@metrifi.com.
Derik Krauss is the co-founder of MetriFi (metrifi.com). He didn’t set out to disrupt the status quo—his heart just led him to seek the truth. That passion has fueled over 100 A/B tests and helped credit unions generate ~$143M in new loans and deposits. Now, he’s building MetriFi and Paraloom to harness analytics and AI for measurable growth. He’s a family man who loves Jesus, liberty, basketball, fly fishing, and delicious food. He can be reached at dk@metrifi.com.
