GEO, AEO, AI Search, AI Optimization. It goes by many names, but the idea is the same: getting your dealership to show up in answers from LLMs like ChatGPT, Gemini, Claude and others. You’re hearing about it in 20 groups, OEM meetings and vendor pitches, and you’re trying to separate buzz from substance and figure out how to measure your position in this new landscape.
Let’s explore the emerging world of GEO and, more specifically, AI visibility tracking. You’ll see what tools exist, how they attempt to measure visibility and the limitations you need to understand before trusting the numbers they produce. Reacting to flawed visibility data can carry real costs, and right now the industry feels a bit like the Wild West: plenty of activity, but very few reliable signs pointing the way.
How GEO Tracking Works
Where there’s demand, supply will always fill the void. Every business in every industry is wondering how they show up in AI search, so it’s no surprise that there are already dozens of tools offering some version of tracking.
Names like Profound, Semrush and Ahrefs are some of the big ones in the space as of this writing. Each offers a window into how your brand shows up in responses to AI prompts. They offer pre-configured scans or fine-tuning options for a range of user skill levels. They can compare how you show up to how other dealers or entities show up for the same prompts. They even offer a score for an easy way to watch your progress (more on the dangers of this later).
Each tool essentially takes prompts from a precompiled master catalog — or allows you to make your own — and runs those prompts through the chosen AI/LLM. They track whether you appear in the results and how high up in the answer your citation appears. In theory, it’s best to show up above other brands (presumably competitors) for maximum impact.
Each tool will default to a generic geo-center, like the United States, giving you broad and generalized results, while also having the ability to select a specific search location if you use more advanced features.
While costs vary a bit between platforms, you can expect to pay about $300/mo or so to track one website and just 100 specific prompts. Some of these platforms are a bit less expensive if you use their precompiled prompt list, but we’ll get into the problems with that below.
One key thing to understand here is that, if you track this yourself, you should expect to spend some money and need to know what you’re doing with configurations. If you have a trusted agency do this for you, their AI visibility upcharge is not just “because they can.” They have a hard cost between the platform fee and their labor, and the economy of scale is currently minimal, if any. The alternative is that the vendor adds this for “free,” but all costs are ultimately passed to the end user (you). Free AI visibility tracking will usually mean something else was reduced, and most SEO solutions in automotive are already very light on real deliverables.
At Wikimotive, we wrestle with the realities of AI visibility tracking for our dealer partners on our SEO program. On the one hand, this is a hot-button issue, and everyone wants to know how they show up. On the other hand, the value of dealers showing up on LLMs is much less than people realize as of early 2026, and the insights from AI visibility tracking are not significantly actionable. That means the tracking cost is wasted money unless it’s worth it to you, simply to have the historical data as the landscape evolves — and for many dealers, it is indeed worth it.
Let’s explore the pitfalls of AI visibility measures, so you can temper your expectations and the pivots you make with this information.
Navigating AI Visibility Challenges
Let’s start with an example from traditional SEO/organic rankings to ground this in something more familiar before we expand the example to AI visibility.
Say you’re a Chevy dealer and you have rank tracking set up to see your Google organic rankings. One of your keywords is “Silverado Towing Capacity,” and you rank #1 for that keyword. Is that good? Well, Google identifies that keyword as an informational keyword, which means they know that people who search that keyword are seeking information, as opposed to being a transactional keyword that would signal an immediate shopping intent. This means your #1 ranking for “Silverado Towing Capacity” is probably getting you traffic to a blog, but it’s not producing any leads. It’s a vanity ranking only.
Let’s change the example and say the keyword is “Chevy Dealers,” and you rank #1. Is that good? You bet! It’s a low-funnel, transactional-intent keyword, and it’s also one of the most frequently searched keywords.
In the Chicago metro, for example, Google reports that the keyword is searched 5,100 times every month. But there’s a snag: WHERE is your rank tracking set? If it’s your own zipcode, that’s good, but what about five miles away? Typically, a dealer will drop between two and six positions just five miles away. Bear with me, as this is going somewhere!
Now, what if the keyword is “Best Chevy Dealers” and you’re #1? This is a bit closer to something a person might enter as a prompt in an LLM. Well, according to Google, just 10 people per month search that keyword in the Chicago metro. That means your #1 ranking is again a vanity metric because there’s not enough juice to meaningfully squeeze.
According to a 2026 study by Conductor, just 0.5% of website traffic in the automotive vertical is driven by LLMs. This tells us that, while tons of people use ChatGPT, Gemini, Claude, etc., only about one in 200 low-funnel searches is being performed on LLMs, and none of them actually report monthly prompt-volume statistics like Google does for search.
So what am I saying here?
The first problem with AI visibility tracking is that you have no idea how often the prompts you’re tracking are being used.
The second problem is that the low-funnel automotive prompts — just like in Google search — are going to be location-dependent. That means you need to track in many locations at the same time if you want to use this as more than a leading indicator. Each location multiplies the number of prompts you have to pay to track, so the cost balloons very fast.
The third problem is that AI results are probabilistic, not deterministic. In short, a Google ranking is a fixed position until it changes. Search the same keyword twice in a row, and you’ll get the same ranking result. Search the same prompt in ChatGPT twice, and you’ll get variations in the answer and the cited sources.
The fourth and most major problem is that nearly all of the default AI visibility tracking is absolutely filled with informational prompts. These are the ones where it simply does not matter if you show up because no one is going to click the cited source. Trying to push your visibility here is doing Tier 1’s job on your dollars.
The result of all this is that default tracking leads to alarmingly low AI visibility “scores” due to prompts being informational and your site being compared to the national brands and authoritative resources that LLMs will always prioritize for those prompt types. The alternative is to configure your tracking manually: Select your location, chosen LLM, and prompts, and the information will still be subject to probabilistic aberrations and unknown “search” volume.
What To Do
If you don’t mind spending some money for a window into what’s happening in the LLMs, there’s nothing wrong with setting up tracking. This can help keep you aware and give you a starting point if the landscape evolves significantly. What you don’t want to do is treat the results as actionable insights, especially if you used the out-of-the-box prompt set and national geography.
Most vendors are using the out-of-the-box setup specifically for their sales audits. They’re set up in a way that will make you look bad every time, and, whether or not they even realize the flaws in their approach, they’re effectively preying on your fear of missing out to sell you something.
At Wikimotive, we’re taking a steady and measured approach to this. As we do with everything in the search environment, we’re testing carefully with our dealers to understand the pros and cons of this space. This leads to informed and contextual decision-making, which avoids knee-jerk reactions to “scores” in a cookie-cutter tool configuration.
I implore you to do the same and to ask critical questions of your agencies — or those who are pitching you — when they bring AI visibility tracking up as the latest, greatest shiny thing.