Most websites are very good at talking. They have pages about checking accounts, auto loans, mortgages, credit cards and membership. They publish articles, promote products and put calls to action in all the right places.
They're also very good at measuring. They know how many people visited a page. They know which pages get the most traffic. They can tell you how many people clicked a button or used a calculator.
But there's a gap between measuring what visitors do and understanding what visitors are telling you.
And I think that's where websites are headed next.
Every interaction is a signal
Consider a visitor who arrives at a credit union website and:
-
searches for auto loans
-
reads the auto loan page
-
looks at rates
-
uses the payment calculator
-
comes back two days later
-
clicks "Apply"
Traditional analytics sees a collection of events.
A smarter website sees a journey.
The visitor isn't simply generating page views. They're telling the website something:
I'm shopping for an auto loan.
That's useful information.
The question is: what does the website do with it?
From analytics to understanding
Analytics has traditionally been about reporting.
- How many visitors?
- Which pages?
- How long did they stay?
- How many conversions?
Those numbers are valuable, but they mostly tell us what happened.
The next step is using those signals to understand what may be happening and what might help the visitor next.
For example, suppose a visitor repeatedly interacts with mortgage content and then uses a mortgage calculator. That's stronger evidence of intent than simply visiting the mortgage page.
Someone who searches for "boat loan rates," reads the boat loan page, and uses a calculator is giving the website even more information.
And a successful online banking login tells us something else entirely: this is very likely an existing member.
None of these signals needs to identify the person by name.
The website simply needs to pay attention to what is happening.
We don't all need the same website
This is where I think the idea of personalization starts to evolve.
The goal isn't necessarily to create a completely different website for every person.
It's to make the website responsive to context.
A visitor researching their first mortgage may need education and reassurance.
A visitor who has already compared rates and used a calculator may need a clear path to an application.
An existing member shopping for an auto loan may need something different from a prospective member who arrived looking for information about joining the credit union.
The website shouldn't have to guess blindly.
The visitor is already giving it clues.
We've learned a lot from personalization
Personalization has been an important step in making websites more responsive to their visitors.
We've seen firsthand what happens when a website recognizes that someone has shown interest in a particular product or topic and uses that information to surface something more relevant.
But working with credit unions on these experiences has also raised a bigger question for us:
What if the website didn't need us to tell it everything?
What if it could learn more from the behavior already happening on the site?
What if it could recognize patterns across a visitor's journey, rather than relying on a single interaction?
And what if it could also learn from what we've seen across other credit union websites?
After working with more than 50 credit unions, we've seen a lot of different ways people navigate, search, compare and make decisions. That experience gives us a unique perspective on what those behaviors can mean and what a helpful next step might look like.
The goal isn't to make every website the same. It's to use what we've learned to help each website make better decisions for its own visitors.
And what if AI could help connect all of those signals, while keeping the credit union firmly in control?
This is where AI gets interesting
We don't need AI to simply generate more website content.
The internet already has enough content.
We need AI helping a website understand which content is useful, when it is useful, and what might be missing.
Imagine a system that can look at:
-
website behavior
-
search terms
-
content interactions
-
calculators
-
forms
-
application starts
-
successful journeys
-
returning visits
-
referral sources
-
analytics data
-
existing website content
and then recognize patterns.
It might discover that visitors who use a particular calculator are much more likely to apply. It might discover something even more interesting: the amount they're calculating matters. Maybe visitors calculating loans in a particular range are significantly more likely to apply. Maybe there's a loan "sweet spot" that correlates with higher intent. Those are the kinds of patterns that can be almost impossible to spot in traditional analytics, but incredibly valuable when a website can learn from its own behavior.
It might discover that people repeatedly search for a question the website doesn't answer well.
It might identify a common path that leads to successful applications.
Or it might recognize that a visitor has already demonstrated strong interest in a product and offer a more appropriate next step.
AI doesn't have to replace the content teams are already creating. It can help them make better use of it.
A credit union already has a tremendous amount of useful information on its website. The challenge is getting the right information in front of the right visitor at the right point in their journey.
AI could help identify those opportunities, connect existing content, and eventually propose new content when a genuine gap exists, all while working within rules and approved sources established by the credit union.
Sometimes the smartest response might be a new piece of content.
Sometimes it might simply be surfacing something that already exists.
Sometimes the best decision might be to stop showing things that aren't relevant.
That's a much more interesting use of AI than simply asking it to write another blog post.
The website should learn
I think the next generation of websites will move through a simple cycle:
Observe.
What are visitors doing?
Understand.
What might that behavior mean?
Respond.
What would be most helpful right now?
Measure.
Did that response actually help?
Learn.
What should we do differently next time?
That's behavior-driven content.
It's not about making a website feel creepy or overly personalized.
It's about making it more useful.
And credit unions have an interesting opportunity
Credit unions already have an enormous amount of information about the people they serve.
But the website often operates separately from that intelligence.
And every credit union website also has something else working in its favor: thousands of interactions happening every month.
The opportunity isn't necessarily to collect more data.
It's to make better use of the data you're already generating.
At Level 9, we've worked with more than 50 credit unions. We've seen how people navigate these sites, where they get stuck, what they search for and which experiences help move them forward.
We're increasingly interested in what happens when that experience is combined with the behavioral data a website generates every day.
Because a website shouldn't just tell visitors where to go.
It should pay attention to where they're trying to get.
We're beginning to think about websites differently.
Not a website that simply publishes information.
Not a website that simply measures behavior.
A website that learns.