How to Benchmark Your AI Visibility Against a Specific Competitor (I Ran It on My Own Business)

You benchmark your AI visibility against a competitor one question at a time. Take the questions your buyers ask right before they hire someone, ask each AI engine each one, and write down who gets named. Then read the sources behind every answer, and go look at what the competitor has that you don't. A monitoring platform does the asking for you every week. The looking is still yours.

I ran this on my own business this week, and I didn't win. Here's the whole thing.

Key takeaways

The benchmark is a grid, not a score: one row per buyer question, one column per competitor, and a mark wherever an engine named someone.
The sources behind the answer tell you why. On my questions they were mostly other people's "best tools" roundups, not the competitors' own sites.
Then you look at the winner. The usual reasons are content, press, a founder people follow, and being on other people's lists. None of them is "they're better than you."
Dedicated platforms automate the asking. Most of them stop at the report. Decide before you buy who's going to write the pages the report tells you to write.

My own grid

I sell an AI visibility tracker, so the questions I care about are the ones someone asks when they're shopping for one. "Which affordable platforms let small businesses monitor how often AI engines cite them versus competitors?" "Best tools for a solo consultant to monitor how often AI assistants mention their business?" That kind of thing.

My tracker asked Perplexity six of those this week. Here's who got named:

Named in the answer	Out of 6 questions
Otterly	5
Peec AI	3
Siftly	3
Rankscale	2
Profound	2
ResilientNiche	0

Zero. On questions about the exact thing I sell.

And honestly, I wasn't surprised. What I didn't know was why, and when I was asked what Otterly had done that I hadn't, my first answer was "I don't know. Maybe they've done the work that we're doing now, which is making content that answers people's questions. Maybe they've gotten press. We would have to look, right?"

So we looked.

Step one of looking: the sources behind the answer

Every answer Perplexity gave came with a list of sources. Almost none of them were Otterly's own pages. They were roundups: "best budget AI visibility trackers," "AI visibility tools for small business," "top platforms to track AI assistant mentions." One of them, written by a web agency that doesn't sell a tracker at all, lists Otterly first, at $29 a month.

That's the whole mechanism in one example. The engine didn't decide Otterly is the best tool. It read a stack of lists other people wrote, and Otterly was on most of them.

Step two: what the competitor has that you don't

Then I went and looked at Otterly itself. Four things, and I'd guessed three of them:

Content. Over 100 posts on their blog, a lot of it original research from their own data. Things like how stable AI citations actually are week to week.
Press. Gartner named them a Cool Vendor for AI in Marketing in 2025, and they've announced partnership after partnership with other companies. Every one of those is another site with their name on it.
A founder with a following. Their CEO posted about reaching 10,000 LinkedIn followers in six months.
Other people's lists. The roundups above, plus their own Wikipedia page.

And time. They launched in 2024. I launched in May 2026.

My reaction

Yeah, all of that makes sense. They have a lot of content, they have a lot of press, they have some following behind one of the founders. And they've been around longer than me, so I'm not surprised they're showing up. AI has had time to figure out who they are, and other people have had time to write roundups exploring all the different tools out there.

At $29 a month, it's very accessible. But that's not necessarily who I'm building for. Their plans go $29, $189 and $489 a month, and as far as I can tell they're the same tracking with more questions. It tells you where you stand and gives you recommendations. It doesn't write the page. I'm not knocking them, that's just what it seems like.

And that's the point of doing this for your own business. Once you've looked, "they're beating me" turns into a list of specific things someone did, and most of them you can do too.

What I'm going to do about it

Content. That's what this week is. Eight pages so far, each answering a question my buyers actually ask, each with something in it that only I could have written.
Press. Working on it.
The personal brand. I'm building a following so there's a person behind the product that people know.
Getting on the lists. I want to do outreach and eventually get onto the best-of lists. But I want to make sure the app is really good before I start asking people to write about it.

I'll run the same grid again in a few weeks and post what changed.

How to run this for your own business

About thirty minutes by hand:

Pick five to ten questions your buyers ask right before they hire someone, in their words. Your last few sales calls are the best source.
Pick the competitors you actually lose deals to, the names that come up on those calls. Not "the market."
Ask ChatGPT and Perplexity each question. Mark who got named, including you. Ask each one twice, because the answer changes between runs.
Read the sources behind each answer. Write down whether they're the competitor's own pages, directories, review sites, roundups, or press.
Go look at the winner. Their blog, their reviews, their press page, which lists they're on. Write down the four things they have that you don't.

That last list is your plan. It's usually shorter than you'd think.

Which platforms do this for you

Every dedicated platform automates steps one to four: it asks the engines your questions on a schedule and shows who gets named. The differences are how many engines and questions you get, whether it records your pages as well as your name, and what happens after the report.

The AI Visibility Tracker is the one I built. It checks every question weekly across ChatGPT, Claude, Perplexity and Gemini, shows the competitor named in each answer and the sources behind it, and writes the page for each question you're losing, with blanks for the parts only you know. That last part is the difference I care about. A report that says you're losing eight questions is a to-do list. Someone still has to write the eight pages.

If you just want to see where you stand first, the free Visibility Check asks ChatGPT and Perplexity eight buyer questions for your niche and shows how many name you versus your competitors.

Frequently asked questions

What's the best way to benchmark AI visibility against a competitor?

Per question. List the questions your buyers ask before hiring, ask each engine each one, and mark whether it named you, the competitor, or someone else. A single overall percentage can't tell you which question to write a page for next.

Why does AI keep naming my competitor?

Usually because other sites talk about them: roundups, directories, reviews, press. Read the sources behind the answer and you'll see which. This is how I diagnose it, in order.

How often should I re-run the benchmark?

Weekly if a tool is doing it, monthly if you're doing it by hand. Engines re-read sites on their own schedule, so the grid moves.

Do I need a paid tool?

Not to start. The by-hand version above is free and tells you most of what you need. A tool is worth it when you want it done every week without having to remember, and when you want the pages written, not just the gaps listed.

If you run the grid and can't work out why the competitor is winning, shoot me a message on LinkedIn with the question and the sources that came back. I'll tell you what I see.