I Benchmarked My Own AI Visibility and Scored Zero

Q: What does zero AI citations actually mean for a new site?

It means AI engines did not cite that domain for the specific queries tested, not that the person or brand lacks authority. For a domain that is only a few weeks old, zero is the expected baseline. AI engines tend to lean on sources with accumulated entity signals, third-party mentions, and crawl history, none of which a new site has yet. The score is a starting line, not a verdict.

Q: How do you benchmark your AI search visibility?

Define a set of lane-specific buyer queries, run them across the engines your buyers use such as Perplexity and ChatGPT, and record which domains get cited and how often. I tested 32 queries across my category on two engines and logged every cited domain. The output is a citation count by domain, which shows who AI currently treats as the answer in your space.

Q: Who does AI cite in the fractional CMO and GEO category?

In my benchmark, AI cited a mix of rivals and adjacent players: Growtal, Profound, Scrunch AI, beOmniscient, rankedcmo, Directive Consulting, Semrush, and Ahrefs. These are domains with established entity signals and third-party coverage. That mix tells you which sources AI currently trusts in the category, which is exactly the map you need to close the gap.

I ran an AI visibility benchmark on my own website, andrewbyzov.com, across 32 lane-specific buyer queries on Perplexity and ChatGPT. The result was zero citations. Not one. AI named my competitors instead. I am publishing that score on purpose, because it is the most honest demonstration I can offer of the exact problem I help clients fix, and it is the realistic “before” of a case study I intend to finish in public.

I’m Andrii Byzov, a fractional CMO for B2B tech who runs generative engine optimization for B2B SaaS. The benchmark above comes from my AI search visibility benchmark, where I tested 148 queries across the category and logged 1,055 citations spread over 508 domains. My own domain landed on zero of them. Here is what that means, what it does not mean, and the work it points to.

Key takeaways

Across 32 buyer queries in my category on Perplexity and ChatGPT, andrewbyzov.com received zero citations.
AI cited rivals and adjacent players instead: Growtal, Profound, Scrunch AI, beOmniscient, rankedcmo, Directive Consulting, Semrush, and Ahrefs.
Zero means zero in this benchmark, not zero authority. I am a working CMO; the site is only weeks old.
For a new domain, zero is the expected baseline, not a failure.
I am running the GEO playbook on my own site and will re-run the benchmark to show whether the number moves.

What I tested and what came back

I built 32 queries the way a buyer in my lane would actually ask them: fractional CMO, GEO, AI-SEO, and marketing consultant searches, phrased as questions and comparisons rather than keywords. I ran each one on Perplexity and ChatGPT and recorded every domain the engines cited. The point was not to find myself. The point was to see who AI treats as the answer when someone with my buyer profile goes looking.

My domain was cited zero times. Instead, the engines surfaced a consistent cast: Growtal, Profound (tryprofound), Scrunch AI, beOmniscient, rankedcmo, Directive Consulting, Semrush, and Ahrefs. Some are direct rivals, some are adjacent tooling and agency players, but all of them share one thing I do not yet have: accumulated entity signals, third-party coverage, and a crawl history the engines have already learned to trust.

That cast is not a wall. It is a map. It tells me exactly which sources AI currently reaches for in my category, and therefore exactly where the gap is.

What zero does and does not mean

This is the part I want to be careful about, because it would be easy to overplay this either direction.

Zero does not mean I have no authority. I am a working fractional CMO with real clients and real outcomes. It does not mean the engines have judged my expertise and found it lacking. They have not assessed me at all, because there is almost nothing for them to assess yet.

What zero actually means is narrow and specific: my domain was not cited for these 32 queries, on these two engines, at this moment in time. My website is only weeks old. It has thin crawl history, few inbound mentions, and limited structured entity data. In that situation, zero citations is not a surprising failure. It is the expected baseline for a new domain. AI engines lean on sources that have already earned signals over time, and a site that has not existed long enough to earn them starts at the bottom by default.

So I am reading this as a starting line, not a verdict. The honest framing matters here, because the whole discipline lives or dies on intellectual honesty. If I dressed up a new-domain baseline as a crisis, or hid it because it looks bad, I would be doing the thing I tell clients never to do.

Why I am publishing the bad number

Two reasons.

First, proof. Anyone can write that AI search visibility matters. Fewer will show you their own score when it is zero. Running this on my own site, with my name on the result, is the most credible way I can demonstrate that I actually do the work, that the benchmark is real, and that I read the data straight. If you want to see the full category picture behind my zero, it is all in the AI search visibility benchmark.

Second, this is the genuine “before.” A zero-to-something case study is only worth anything if the zero was real and documented before the work started. I would rather show you the honest baseline now and let you watch the number move, or not, than show you a tidy “after” with no verifiable starting point. The way to track that movement is the same loop I use with clients: see how to track AI visibility and what is share of model.

What I am doing about it

The plan is the GEO playbook, run on myself, in the order that tends to move the needle fastest.

Fix the foundation. Entity clarity and structured data first, so the engines can resolve who I am, what I do, and what I am known for. This is the highest-leverage, fastest-moving layer for a new domain.
Ship answer-shaped content. Pages that directly answer the buyer queries I tested, written so an engine can lift a clean, extractable answer. The post you are reading is part of that.
Build source authority. Earn the third-party mentions and category coverage that the cited domains already have. This is the slowest layer, and there is no shortcut.
Re-run the benchmark. Same 32 queries, same engines, on a recurring cadence, so the citation count becomes a metric I track over time rather than a one-time snapshot.

If you want help mapping your own category and running this loop, that is what I do as a GEO consultant for B2B SaaS and through my GEO consulting work. You can also run a quick read on your own domain with my free AI visibility score tool before you commit to anything.

The honest caveat

A benchmark is a measurement, not a promise. No one controls how Perplexity or ChatGPT choose their sources, so the realistic goal is improved odds of being the cited answer over time, not a guaranteed placement. I expect my own number to move slowly, and I expect some of these queries to stay stubborn. That is fine. The point of publishing the zero is to commit to showing the real trajectory, up or flat, and to be the proof that I read my own data the same way I would read yours. I will report back when I re-run it. If you want to talk through where your category stands, I’m on LinkedIn.

FAQ

What does zero AI citations actually mean for a new site? It means AI did not cite that domain for the queries tested, not that the brand lacks authority. For a site only weeks old, zero is the expected baseline, because engines lean on accumulated entity signals, third-party mentions, and crawl history a new domain has not earned yet.

How do you benchmark your AI search visibility? Define lane-specific buyer queries, run them across the engines your buyers use such as Perplexity and ChatGPT, and record which domains get cited and how often. I tested 32 queries on two engines and logged every cited domain.

Who does AI cite in the fractional CMO and GEO category? In my benchmark, AI cited Growtal, Profound, Scrunch AI, beOmniscient, rankedcmo, Directive Consulting, Semrush, and Ahrefs. That mix shows which sources AI currently trusts in the category, which is the map for closing the gap.

Can you guarantee you will get cited after doing GEO work? No. No one controls these systems, so the honest promise is improved odds over time, not a guarantee. The work is to fix entity clarity, ship answer-shaped content, and build source authority, then re-run the benchmark to see whether the count moves.