Executive Summary
Steven Ramirez, CEO of Beyond the Arc, says the early stage of the customer journey now starts on AI platforms. People ask ChatGPT, Gemini, or Claude which institution to consider, and the ranked list of links on a Google results page matters less every year. Ramirez has spent more than 20 years as a fintech and marketing strategist. He sees two problems for banks and credit unions. The first is whether an institution appears in the answer. The second is whether the answer reflects what makes that institution different. An institution that appears with a generic description has gained little.
Measuring the effect is hard. Ramirez recently reviewed 16 months of Google Analytics 4 data and found that AI visits often arrive without the tags that identify them, so they show up as unspecified traffic. He recognizes them by their pattern. They go deeper and show higher engagement than other visits. The impact may be larger than the reports show. The ground also moves quickly. Ask the same question twice, five minutes apart, and the answers can name different institutions. Ramirez describes a bank that was the best answer yesterday and loses that position after ten competitors publish clearer content.
Visibility depends on work inside and outside the website. On the site, institutions need clear content, answer blocks that AI can extract, FAQs, and schema markup, which is code that tells large language models what a page describes, such as a rate or a service. Outside the site, models draw on what other sources say about the institution. Rates add compliance risk. Ramirez says many banks have made rate information hard to find, on the assumption that customers will talk to them. An AI platform may then rely on stale information or fail to find a rate at all.
His audit covers four areas: how well AI understands the institution’s position, content readiness, authority and reputation, and measurement. An outside firm is optional, but busy marketing teams find the work hard to do systematically. He advises focusing on products the institution wants to win. A community bank without a large auto lending business gains little from appearing in auto financing answers. His first step takes 20 minutes. Read the current content and check whether answers are easy to extract. Visitors who arrive with an AI-generated shortlist have specific expectations and other places to go, which raises the bar for personalization. How to personalize for the AI itself, and what agentic commerce means for banking, remain open questions.
Listen, watch, or subscribe on your preferred platform:
YouTube | Spotify | Apple Podcasts
We are also available on all other major podcast platforms. Simply search for “Next Best Action.”
Key Takeaways
- Discovery Starts on AI Platforms, and Showing Up Is Not Enough: Ramirez says the early stage of the customer journey now begins on AI platforms, and the list of links on a Google results page is losing its place as the starting point. The answer also has to reflect what makes the institution different. Appearing with a generic description does not help.
- AI Traffic Is Hard to Attribute, and the Impact May Be Larger Than It Looks: In a 16-month review of GA4 data, Ramirez found AI visits that arrived without identifying tags. They look like unspecified traffic, but they go deeper and show higher engagement. If the impact looks small today, he says that is a reason to take a deeper look.
- SEO Process Does Not Carry Over: Ramirez names one belief as the biggest misconception: that the same process, tools, and reports used for SEO will work for AI. Answers vary between identical questions, competitors can overtake an institution quickly, and several platforms matter instead of one. AI also weighs credibility, which Google rankings did not weigh the same way.
- Missing Rates Create Compliance Risk: Ramirez points to the compliance care that goes into how institutions communicate rates. He says many banks have made rate information hard to find, on the assumption that customers will talk to them. An AI platform may then rely on stale information or fail to find a rate. He says mortgage lenders are working out how to keep a high-touch, relationship approach while adapting to AI.
- The Audit Covers Four Areas, and Only Where the Business Competes: Ramirez’s audit looks at AI understanding, content readiness, authority and reputation, and measurement. He advises focusing on products that matter to the institution. A community bank without a large auto lending business gains little from appearing in auto financing answers. An outside firm is optional, but busy teams struggle to do the work systematically.
- Start With the Content You Control: When a competitor appears in AI answers and you do not, Ramirez recommends a side-by-side comparison of the content, not only the result. Look for extractable answer blocks, consistent FAQs, and clear key takeaways. A marketing leader can get a sense of where the content stands in 20 minutes.
- AI Shortlists Raise the Bar for Personalization: Ramirez says the customer journey moves back and forth between awareness, consideration, and decision. A visitor who arrives with a tighter shortlist has other options if the site misses their needs. He says personalization and differentiation work together. Whether institutions must personalize for the AI itself, and how agentic commerce changes banking, are questions he says have no answer yet.
To learn more about Beyond the Arc’s AI visibility work, visit https://beyondthearc.com/ai-visibility-for-banks.
Full Episode Transcript
Alex Jiménez: Welcome to Next Best Action. My guest today is Steven Ramirez, CEO of Beyond the Arc, where he leads client work at the intersection of go-to-market strategy, customer experience, and AI innovation for banks, fintechs, and financial services companies. He has spent over 20 years as a fintech and marketing strategist. And if you go to a lot of fintech conferences, you’re going to see him there. Welcome, Steven. Well, we can talk about anything in fintech. Our topic today is about AI visibility, which both of us are spending a lot of time on lately. So welcome.
Steven Ramirez: Alex, thank you. I am so much looking forward to this conversation.
Alex: Me too. And like I said, we can talk about anything, but we don’t want to bore people with two hours’ worth of content. So speaking about AI visibility, why should financial service companies start to worry about how AI describes them, and not just how Google ranks them?
Steven: So Alex, I think there’s two main considerations here, and it applies both for banks and for fintech companies. The first is perhaps the more obvious one, which is behavior is shifting, both from a consumer perspective, but also from a commercial B2B perspective, in terms of how people find information.
They are not following the blue links. The age of blue links on Google is fast declining. And so if you want to be discovered, the point of entry is increasingly the AI platforms, right? So one very compelling reason to focus on AI visibility is that the early stages of the customer journey are now beginning on the AI platform.
I think that’s first. But I think that visibility is not enough. You may show up, but are you showing up in a way that really reflects the differentiation that your company is all about? The thing that makes you special, is that appearing in these AI answers? And so I think these are the challenges that marketers now have to tackle.
Alex: Yeah, and looking at some of the data from our clients, we see that AI as a source is still a very, very small number. But on the other hand, we are seeing that anything that comes in from AI has a much higher conversion rate than those coming in from just Google. So I think you’re right that it’s becoming more and more important. It doesn’t mean you forget about SEO, but it means you have to now think about AI visibility as well.
Steven: Yeah, and I think, Alex, there’s definitely some intersections there. So I think a good place to start is with your Google Analytics data. You can look at GA4. But having just done a really deep dive into the last 16 months of GA data, one of the things you find is that the attribution is not always clear.
So you may not know that that is a ChatGPT visit because the tags didn’t make it all the way through to your analytics, but in fact it was. And so you might say, well, how do you know that? Well, you can look at the visit patterns of AI visits versus non-AI visits. Just as you’re saying, they tend to be deeper. They tend to have higher levels of engagement. So you may find that there’s an increase in traffic that you don’t know where it’s coming from. It looks like it’s just, you know, unspecified. But actually, the impact may be bigger than you think today because of the lack of attribution.
And the other aspect, and maybe it’s not so big for you, but maybe it’s growing for your competitors.
Alex: Exactly.
Steven: Which is exactly the problem. You’re not seeing that big of an impact today. That’s a good indicator that it’s time to take a deeper look.
Alex: Yeah, yeah, absolutely. So when a prospect asks ChatGPT or Gemini or Claude about mortgage rates or account options, what determines whether a bank or credit union shows up in that answer?
Steven: Well, there’s a variety of both on-page factors and off-page factors that you have to look at. The on-page factors are the ones that you have the strongest level of control over. It’s how structured is the data, the information. So it’s not just, does it appear on the page, but in the backend code, do you specify that that’s a rate? Do you specify that that’s a service that you offer? Through the use of schema, there’s sort of backend code that provides these hints to large language models. So there’s things that you can do on the technical side. There’s things that you can do on the content side.
But then there’s lots of factors that exist beyond your site. And that is the thing that I think is very different with AI visibility versus SEO. The large language model is looking at all of the information across many sites, including the sites that mention you. And so now, what others say about you is directly connected to what you say about yourself. So there are a lot of factors that come together.
And I’ll just give you one other thought. Particularly with something like mortgage rates, rates of any kind, I mean, Alex, you know that the level of compliance care that goes into how you communicate rates is really critical. So it’s not just, you know, how do you structure it, but what happens if you don’t get it right?
Alex: Yeah. Well, what happens if ChatGPT hallucinates your rate?
Steven: Yeah, or has stale information, or can’t find it. And so I think around rates for banks, this is a really challenging question, because I think many banks have gone down the path of not making that rate information so easily available, with the idea of, don’t worry, they’ll talk to us. Well, ChatGPT ain’t gonna chat with you, right? So then when it’s trying to find this information, it’s gonna be missing. And if that information is missing, to your point, does it make it more likely that there’ll be a hallucination?
Alex: Yeah. Yeah. Recently I spoke to a head of mortgage, and they said, “We don’t want our rates on the website. If they want to know our rates, give us a call, talk to our officers.” And right there, that’s a problem for various reasons.
Steven: And it’s very profound, because the industry, in many areas, mortgage being one of them, this has been the way, right? This is how it has operated. I was actually in Indiana, also talking to mortgage bankers. I was at the Mortgage Bankers Association. You mentioned getting out to conferences. So I was in Indianapolis. And I see in mortgage in particular the industry trying to get its head around how it maintains this high-touch relationship orientation, but adapt to these new challenges that AI is presenting.
Alex: Right, right. What’s the biggest misconception banking marketers have about AI visibility right now?
Steven: I think the misconception is that you can address it using the same process that you have been using for SEO, search engine optimization. That this is just one more type of optimization. You can still use your same process, your same tools, your same reports. I think that is wrong, and potentially making it more challenging for you as an institution.
Alex: And the way that large language models work, if you ask ChatGPT right now a question and then five minutes later you ask the same question again, you’re going to get vastly different answers, and you may have different organizations being quoted or presented to you. So it’s not in the SEO days or the SEO process where, okay, you know that you’re going to be in the first three or four responses. You may or may not be in that response at all, depending on what you’ve done for AI visibility.
Steven: And Alex, it moves so much faster. So you were used to looking at your ranking in Google. That, of course, moves, but there is so much focus on AI visibility and AI optimization that even if you were part of the answer yesterday, well, now there’s literally 10 banks that have published new content on their website that scored better. It was clearer. It had better detail. It was in chunks that could be extracted more easily by the AI. It had a higher value to users, according to the AI. And so while you were the best option even an hour ago, now there’s 10 new competitors that have more, that their content, that the value of the content is now stronger than yours.
Alex: Yeah, and for SEO generally, you only worried about how you show up in Google, and, you know, Bing and other search engines were really not that important. But now it’s ChatGPT, Gemini, Claude, but there’s many, many others, right? And, you know, some of us use them interchangeably.
Steven: Absolutely. Different tools for the job. And from an optimization standpoint, the other challenge is that in the past, the rankings were more mechanical. Did you cover certain search terms? Did you include certain references? Yes or no. Google wasn’t evaluating the strength of your credibility. So it’s as if, Alex, you posted a great how-to on how to do brain surgery, by Alex. Very well documented, lots of research. Perhaps you’re not the most credible source, but your overall authority probably helped that content rank. Now the LLM is like, “Yeah, I’m not going to Alex for brain surgery.” Fintech strategy, absolutely. Brain surgery, no. Those decisions, semantically, can be made by the AI.
Alex: All right. Well, there goes my medical career.
Steven: Sorry to be the bearer of bad news. You’re not going to make that career transition.
Alex: Well, given what you just said, I know that one of the things you’re helping organizations with is doing an AI visibility audit. So walk me through what that means and what it uncovers for an organization.
Steven: So we have a 10-point methodology that we go through as part of this assessment. And the good news for you, Alex, is I’m not gonna go through all 10 right now.
Alex: No, that’s it. Just go directly to Beyond the Arc and you can connect.
Steven: We’ll give you the beyondthearc.com/ai-visibility.
Alex: Exactly. There you go.
Steven: But there’s four areas to think about. So the first is we try to assess the degree to which AI understands, right? So we’re looking at AI understanding and your position in the market. That’s one piece. The second piece is a deeper dive into the content, and I sort of alluded to that earlier. It’s really an assessment both of the strength of your content, but also some of the technical readiness behind the scenes. Things like, how well do you use schema to do markup in your code? We literally, Alex, in the content that we’re creating now, we are dramatically restructuring the way that it looks. It has extractable blocks. It has much more consistent use of FAQs. We’re much clearer about what the key takeaways are. So this is part of that scorecard to evaluate, because that is what AI needs to be able to extract, right? And those are some of the on-page factors.
The other two parts of that framework, we look at authority and reputation, right? And so this really gets to, Alex, you’re a great example. You are quoted in sources like Financial Brand, American Banker. You’ve spoken at Money20/20. You have things in the industry that build your credibility that AI now uses, right? So part of it is how do you show, and part of a broader landscape that did not used to be part of your SEO considerations, or not exactly in the same way.
And then the last piece is the measurement. So we create it for your target audiences. That’s the other thing. Do you really care, if you’re a community bank, do you really care if you show up in content about auto financing, right? Particularly if you don’t have a big auto lending business, it’s great that you show up there, but it’s not moving the needle. It’s not one of the metrics. And so, AI understanding, content readiness, authority and reputation, and then measurement and improvement, that is really what goes into the audit.
And do you need to have somebody do it externally? No, Alex, you do not. But it’s hard to do it systematically internally, right? You know, every marketer that we know is already pretty busy. They already have a lot on their plate, particularly at community institutions. But ask even for fintechs, right? If you’re the head of marketing at a fintech, you have a small team. So it becomes essential. Do you have the resources to execute now? This is what you have to decide. And I think now is a perfect time, as you think about your strategic plan for marketing over the next 12 months, how much focus you’re going to put on this topic around AI visibility.
Alex: Yeah. One of the curious things is, you know, all the things that you mentioned that are now should be part of the copy, should be part of what’s on the website, a lot of those things are what help customers, members, prospects actually make a decision as well. So not only are you fixing AI visibility, but you’re also fixing the understanding for someone who comes to the website just organically. So some of these things you should have been doing already. Right.
Steven: Well, it’s a good reminder, right? And I would say that yes, you probably knew that. And we’ve taken a very sort of straightforward approach to say, let’s think about the customer journey in three stages. There’s an early awareness stage. There’s a consideration stage. There’s a decision stage, right? Everyone is going through that. Now, I’ll say as a footnote, it’s not always linear. You don’t always go from awareness, consideration, and now I’m ready to decide. Oftentimes you get to consideration, something hits your radar, you go back to awareness, you think you’re at decision, you back up, and now you’re at consideration. So there’s a bit of a back and forth. And I think that this is what creates the opportunity for personalization, right? Because what we’re realizing, what that really means, is that the content that you’re delivering is not hitting all of your target segments equally.
Alex: Yeah, yeah. It reminded me of something we just noticed with one of our personalization clients. The number of times that someone in consideration for a product comes back to the website, in some cases it’s a huge number. They may come back to the website 10, 15 times before they actually click on the apply now button, which is not something you’d expect, right? You always think, okay, they’re going to go and look at our rates, and they may look at some of our competitors, and they’ll come back, and then they’ll apply because our rates are better. But several different times, it’s quite surprising how people don’t make that decision just as you said, in a linear fashion. So it makes a lot of sense.
Steven: Yeah, and Alex, how are you using AI for your buying discovery? I mean, I don’t know about you, I use it at every stage. And so new things occur to me, and new opportunities. So I do think that back and forth is also shaped by your interaction with AI.
Alex: Yeah, absolutely, absolutely. So if a CMO finds out that a competitor shows up in AI answers and their company doesn’t, what should they fix first?
Steven: Well, I think the first part of that is, is it relevant for the business? My example, if you’re, yes, your competitors may show up, but is it a battle you need to win? That, I think, is the first point, because you have limited resources, you have limited time, you really need to be focused about where you show up. So, but let’s assume that, yes, you know what, it’s a direct competitor. They’re showing up, we’re not. I think that drives most immediately. I think the thing that you can most easily control is the strength of the content. Ultimately, the strength of the content is what is driving this. So I would look at, Alex, how clear is it? And I would do a side-by-side comparison, right? Not that you’re looking at the end result. It’s like, oh, I’m looking at why they crossed the finish line. You know, I know that you’re a runner. They crossed the finish line first. But what did they do for training? What kind of shoes do they have? You know, they probably had a lot of other decisions along the way that got them to that finish line. And so we need to start doing that sort of forensic review to see. But I would start with the content as the first part of that.
Alex: Yeah, I mean, it’s funny, because that is probably the first step when you’re looking at SEO as well. I remember working for an organization where our agency was spending tons of money on auto lending and we had no program. And it took us about two quarters to figure that out. So we were throwing money out the window.
Steven: You know, Alex, I think that does speak to a really important consideration. You probably have seen on LinkedIn, I’ve seen at least three different posts this week that say, you know, pretty much everyone can be an AI consultant these days. Like, and you know what? They’re all qualified, Alex? No, they are not. It’s not only a knowledge of the technology and the platform, but a knowledge of the business. And so that is exactly what I see with lots of agencies, is that they do not have an intrinsic understanding of the industry. And that is why they don’t know, when you say that you’re in lending, they don’t know to ask the next three questions, right?
Alex: What kind of lending?
Steven: What kind of lending? Auto? Is it direct or indirect? Oh, okay. Do you have a particular, are you doing RVs? Are you doing boats? You know, there’s questions. Does it include insurance? If you know the industry, you know the next five questions to ask, and they don’t. And I think that you wind up spending money and spending effort on areas that are not going to move the needle for that. So it goes back to that differentiation, right? That’s also about the differentiation getting squashed. And so now, particularly as a community bank, if you sound like everybody else, don’t you think you’re just dead in the water?
Alex: Yeah, yeah, yeah. So how does AI visibility connect to the personalization work that banks or credit unions are already doing on their website?
Steven: Well, if I think about it from the personalization lens, to me, I think the lessons of AI visibility say that personalization is more important than ever, because the simple generic path just is not going to get you in front of the people who really are doing the exploration. They are teasing out through AI who they should be talking to. They’re explaining their needs. They are doing a lot of homework very quickly. And so you have to be able to sort of dynamically present this information.
I do think, though, that there is also a technical challenge there, right? So you can personalize for people. I wonder to what extent we now will have to personalize for AI consumption, right, because we now have this intermediary of the AI. This, I don’t think we’re discussing yet, right? I don’t know that there’s an answer, but I think that the strength of personalization, at least, you know, ensures that by the time they make it to your website, it is really tailored to what they’re looking for.
I think it also raises the bar, because they can, through AI, get a tighter shortlist, that if you do make that list and then they come to visit your site and it doesn’t really meet their needs, well, now they clearly have three other places to go. So I think that personalization and differentiation go hand in hand, and they should now be much higher on the radar for banking CMOs.
Alex: Yeah. If you look at the retail world, there’s a lot of buzz about agentic commerce. And really, that’s what we’re talking about here, right?
Steven: Yes.
Alex: Luckily for financial services, that’s not yet hitting, but it’s definitely hitting a lot of the retailers. And so how is agentic commerce going to impact banking? And you may not be doing that personalization for the end customer, but you’re doing it for the agent. So that’s something that you and I don’t yet have an answer for. But it’s coming.
Steven: But we will. We will.
Alex: It’s coming.
Steven: Because I think that if you look at where the industry conversation around this is, what I see changing from an AI perspective is, in the past, consumers and commercial clients understood just a fairly small set of solutions. I want to be able to manage my money in a different way. I want to be able to send it more easily to a friend of mine. I want to be able to send it to my cousin in a different country. I want to be able to take it out of investment, spend it really easily. All these money movement things that make the spending part easy. Well, to execute that, I had to have a checking account, or I had to have Zelle, or I had to have, you know, a Cash App. I had very few choices.
But now I think that with agentic commerce, you now have an infinitely larger palette of possibilities, and you don’t have to figure it out. I can just say what I want to have happen, and it can be really complicated, and it can just be. And so that, I think, does raise the bar. And do I have the answer for that? No, but I think that between the two of us, we’re starting to see this really take shape, and not just in commerce, but I do see it changing the nature of the product set. And again, to me, that’s a big push for the continued push we’ve been on for 10 years of why financial institutions need to continue to innovate, because this space is moving really quickly. And if you want to be part of that answer, you have to be part of that conversation.
Alex: Yeah, absolutely. Let’s switch it up a little bit and let’s do a lightning round. Just give me a quick answer that comes to mind, and, you know, don’t think too much about it. So what’s your favorite non-fintech podcast?
Steven: Most of the ones I like are all fintech. I’m gonna go with Lenny’s podcast, which is around product management, right? So technology product management, not exactly fintech, so I think I slide in there.
Alex: What’s a challenge you’re currently excited about tackling?
Steven: Well, obviously boosting AI visibility for banks and fintechs.
Alex: Right. Great. Great. That was kind of a layup. Outside of the Bay Area, what’s your favorite spot?
Steven: Oh, New York City, easily.
Alex: Nice. Great answer. As a Queens, New York native.
Steven: I don’t know, you know, I was born in New York, I left as a child. I don’t know if I could live there as an adult, but I certainly love visiting New York as an adult.
Alex: Yeah, I totally agree with that.
Steven: Assuming I’m an adult. Let’s assume that for a moment.
Alex: Yeah, let’s assume. What is overhyped in fintech that is not AI?
Steven: I think stablecoin is overhyped. Tremendous value, but still tremendous hype.
Alex: Yeah, yeah, absolutely. I totally agree with you on that one. All right, so one final question. What’s one step a marketing leader could take this week to start improving how their bank shows up in AI answers?
Steven: Take a close look at your current content. Is it easy to extract? Are there real answer blocks? Are you answering questions that people are likely to have? I think you can begin to get a sense of this, Alex, in 20 minutes. Start to read through. And if those blocks, those extractable sort of chunks of content, are not immediately obvious to you, then you’ve got an opportunity.
Alex: Great. Well, thank you, Steven, for the time and all the great ideas. As you said earlier, if you have any more questions about AI visibility, Steven is the person to look to, and just go to beyondthearc.com and you’ll have some answers, and you can connect with Steven there.
Steven: Great, Alex, thank you very much. I’ve really enjoyed this conversation.
Alex: Yeah, same here. And I’m sure we can have other conversations in the future about other topics.
Steven: Excellent, I look forward to it.
Alex: All right, thanks.
