Search is changing fast, and most marketers are already feeling the impact of these changes (even if they haven’t put a name to it yet). AI chatbots like ChatGPT, Perplexity, and Gemini have become a real source for your audience to do B2B research. Buyers are building shortlists inside these tools before they ever land on a website, which means the old rules of visibility (like ranking on Google page one) only cover half the game now.
That shift is exactly why we started running AI visibility audits, both for our clients and for ourselves. An AI visibility audit is a structured way of testing how AI tools see and describe a brand. By using the kinds of questions a real buyer would ask, you can assess how your brand shows up, how you’re described when you do, and where the AI is sourcing its information. We’ve been running this process for clients for awhile, helping them understand their blind spots before they become real downstream problems. One of the best ways to lead is by example, so we’ve been conducting the same audit on our own brand, documenting our findings, tracking notable changes, flagging errors, and making updates where necessary.
TL;DR:
- We audited how ChatGPT, Perplexity, Gemini & CoPilot describe our brand using real buyer prompts.
- We show up for broad and local searches, but disappear for specific strategic-positioning questions.
- The gap is positioning, not visibility. AI reflects what you’ve published, not everything you do.
- Different platforms pull from different sources. One LLM isn’t the full picture.
- AI visibility isn’t one-and-done; it needs regular re-auditing as citation behavior shifts.
Why AI visibility is worth tracking
For the better part of two decades, “search visibility” meant one thing: you ranked on Google. That metric still matters, but it’s no longer the whole picture. Buyers are now asking ChatGPT, Perplexity, Gemini, and CoPilot to do the shortlisting for them, comparing vendors, summarizing categories, and recommending agencies before they ever open a browser tab to search manually.
If your brand isn’t part of the LLM’s answer, you’re not just losing a “ranking.” You’re missing out on being a part of the conversation entirely (often without knowing it was happening).
That’s the whole premise behind an AI visibility audit. Instead of guessing how AI tools talk about your brand, you find out directly. You run a set of real prompts (the kind your buyers would actually type) across the major LLMs, and look at three things:
- Whether you’re mentioned at all
- Gauging how you’re described when you are
- Determining where the AI is pulling its information from
That last piece matters more than people expect. An LLM’s answer is only as good as its sources, and most platforms still lean heavily on third-party roundups and comparison sites rather than a brand’s own content. Knowing your source mix tells you exactly where to focus.
This isn’t a perfect science (yet)
Before covering the results of the audit, it’s worth mentioning that AI visibility tracking is still a new discipline and it’s constantly changing. You can run the same prompt twice and get two different answers. Citation results that look stable one month can shift dramatically the next as the platforms tweak how they retrieve and weigh sources.
That instability isn’t a reason to skip auditing, but rather, it’s a reason to do it regularly. A CMSWire analysis of AI citation trends found that community-platform citation (i.e Reddit) share climbed steadily for several months in late 2025 and into early 2026, then dropped sharply within a matter of weeks after a licensing dispute changed how one major AI platform was allowed to pull from that content. If a single legal dispute can swing citation behavior that fast, no audit you run today is a permanent scorecard. It’s a snapshot, and you need new snapshots on a regular cadence to actually trust what you’re seeing.
A recent Search Engine Journal recap of an AI visibility webinar made a similar point. Brands earning AI citations tend to be the ones with a steady, ongoing community presence and credible data, not a one-time optimization push. Meaning, this isn’t a project with an end date. It’s a habit.
What we found when we audited Syrup
We ran two prompts across some of the major LLMs like ChatGPT, Perplexity, Gemini, and CoPilot. Here’s what we saw:
- The competitive question: “Who are the top B2B marketing agencies in Atlanta?” We were impressed with the initial results. We showed up on 3 of the 4 platforms, landing in the same conversation as several agencies we genuinely respect. When we were named, the language leaned toward our creative and execution strengths like branding, web work, visual identity. All true and all things we’re proud of. And also not the whole story of what we do.
- The specific buyer problem: “What kind of marketing agency helps B2B companies fix their messaging and brand before scaling?” This prompt showed totally different results. 0 of the 4 platforms named us. And the answers weren’t incorrect, but the AI tools simply didn’t have strong signals connecting Syrup to that specific strategic positioning (even though it’s certainly one of the ways we work with clients).
With those two results side by side, a clear pattern emerges. We show up when AI tools are asked who does good creative work in our city, and we disappear when they’re asked who solves a strategic problem. That’s a positioning gap, not a visibility gap. The fix isn’t “get mentioned more.” It’s “give AI tools more first-party proof that strategy is what we lead with.”
We also noticed something else worth mentioning. One platform didn’t mention us at all in either prompt, and its source list looked nothing like the other three. Different AI tools are pulling from different corners of the internet, which means a strategy built around showing up in a single LLM will leave real gaps in your overall visibility.
What this actually means for marketers
A few things became obvious pretty quickly, and they’re not unique to just the Syrup brand.
First, being visible in branded or local searches doesn’t guarantee visibility in category searches, and category searches are usually the ones with real buying intent behind them. Someone asking “who’s the best agency in Atlanta” already has Atlanta in mind. Someone asking “who fixes B2B messaging before scaling” is shopping for a solution and may not care where you’re located.
Second, the language AI tools use to describe you is a mirror of what they can find, not necessarily what’s true. If your strongest work is strategic but your content footprint reads as executional, that’s what gets reflected back. AI tools can only cite the story they’ve got access to, or that you’ve actually published.
Third, where the citations come from matters as much as if you’re cited at all. Relying on third-party “best of” lists for your visibility means you’re at the mercy of whoever ranked that comparison page that month. Building first-party content, like original points of view, real client outcomes, the strategic reasoning behind your approach, gives AI tools something durable to point to that you actually control.
What to do next
This audit told us we’re known for one thing and underrepresented for another, which is a far more useful (and fixable) problem than starting from zero. We’re already turning the findings into action by publishing more first-party strategic content, tightening the language across our own site so it matches how we actually talk to clients, and re-running this audit at a regular cadence, because as the data above makes clear, a single test is just a snapshot.
If you’ve never run this kind of test on your own brand, it’s worth twenty minutes and a handful of prompts in ChatGPT or Perplexity. You might be surprised by what comes back!
If you want help going deeper than a few manual prompts (like actually mapping where you stand across platforms, identifying the gaps, and building the content roadmap to close them), that’s exactly what our AEO services are designed to do.
