CitedOrNot - AI Search Visibility

npx add-mcp 'https://citedornot.com/mcp'

The shortlist is made before anyone visits your site.

• Buyers ask AI which brand to pick “Best sustainable sneakers.” “Eco-friendly shoe brands.” Every question ends in a shortlist of brands. Allbirds makes three of these four and is missing from the fourth.
• Every engine writes its own shortlist Here the brand is named in 86% of checks on four engines and 67% on Google AI Overview. Being on one engine's list says little about the others.
• Whoever isn't named never gets the visit When an answer names a rival instead, the buyer goes to the rival. There's no visit, no referrer and nothing in your analytics to say it happened.

Each of Allbirds' tracked questions with every engine's result: best running shoes for everyday wear is not named on any engine; eco-friendly shoe brands and best wool running shoes are named on every engine that answered

Rivals next to Allbirds: each one's mentions, share of answers and sentiment

From a lost answer to a plan to win it back

The question Allbirds lost, followed through CitedOrNot: who took its place, why, and what to do about it. Every screenshot below is the live product.

01 · Competitor gaps

Know the exact answers you're losing

Add the competitors you care about. Every answer then shows who AI named alongside you, or instead of you, on each engine and for each question.

• Each competitor's share of answers, next to yours
• The questions a rival wins and you don't, with how often
• Changes in your feed, plus a daily or weekly digest by email or Slack

Competitor tracking for Allbirds: each rival's mentions, share of answers and sentiment next to Allbirds', and where rivals beat it

02 · Sources

A visibility score tells you what happened. This tells you why.

AI names a rival because some page told it to: a review, a roundup, a video, the rival's own product page. See the pages each engine leaned on, and what to do about each one.

• Sources to earn: sites that only showed up when a rival was named
• A next step for every site: earn a mention, pitch to be listed, publish your own version
• Which answers each page fed, and who they named

Pro & Agency

Sources AI cites for Allbirds' questions: a source to earn, and each site with what to do about it and which brands the answers citing it named

03 · Action plans

Walk into Monday with a task, not a worry

Every gap becomes something a person can do this week: what to publish or pitch, the buyer question it should win, why the rival got picked, and a checklist. “Improve our AI visibility” becomes a task with an owner.

• Built from the real answers, with the questions you lose most often first
• Refreshed as new answers arrive. Old advice is retired, never silently rewritten
• Tick off the steps, then mark it done

Pro & Agency

An action plan for Allbirds: get listed in Runner's World and The Run Testers roundups for everyday running shoes, why Nike got picked, and a four-step checklist

04 · Proof

See whether it worked

Mark a task done and CitedOrNot keeps asking the question it was meant to win. Once there are checks on both sides of the day you shipped, you see how often AI named you before and after, and how many checks each figure rests on.

• Not enough checks yet? It says so instead of guessing
• A running record of what you shipped and what moved. A one-off audit can't give you that

Pro & Agency

Allbirds' Your results page: nothing shipped yet, and how the before-and-after fills in

Empty on purpose: nobody has shipped Allbirds' plan, so there's no before-and-after to show. We won't make one up.

Set it up once. Then it runs without you.

No script to install, no access to your site, nothing for your developers to do.

1. Add your website

   Your brand name, category and likely competitors are filled in from it. Ran the free test? That part is already done.

2. Pick the questions buyers ask

   Keep the ones suggested from your site, or add your own: “best running shoes for flat feet”, “best accountant for freelancers”, “alternatives to [competitor]”.

3. Get told what changed

   Every engine is asked again on its own schedule. What moved, who's winning and what to do next arrive in your feed, your inbox or Slack.

Your options, honestly

Most people here are choosing between checking it themselves and hoping their rank tracker covers it. Here's what each one actually gives you.

What you need	CitedOrNot	Checking it yourself	A rank tracker
See whether AI names you in its answer
Every engine, not just the one you opened	Pro+	In theory
A trend over time, not a one-off answer			Rankings only
Told when something changes			Rankings only
Which competitor was named instead of you
The pages AI relied on to pick them	Pro+	When it shows sources
A specific thing to publish or pitch next	Pro+
Whether what you shipped worked	Pro+	Only if you log it
Finds buyer questions you aren't tracking	Pro+		Keywords only
What it costs	From $29/mo	Your time, every week	You already pay

Scroll sideways to compare →

Start free. Pay when it's earning its keep.

Every plan starts with the same 7-day free trial. Move up when you need more engines, more brands, or a plan to act on.

The trial covers 1 brand and 2 questions, checked on ChatGPT and Gemini. No card needed.

Starter

$29 _/mo

Billed monthly

For a founder keeping a proper eye on one or two brands.

• 2 brands
• 15 questions per brand
• 2 competitors per brand
• ChatGPT, Gemini and Perplexity
• 750 checks / month
• 31 markets
• Change feed, email & Slack digests
• 2 logins for your team
• AI crawler tracking (30 days)
• 5 first drafts of what to publish / month

Agency

$249 _/mo

Billed monthly

For agencies making AI visibility a client deliverable.

• 50 brands
• 500 questions in total
• 10 competitors per brand
• 18,000 checks / month
• Everything in Pro
• Client management
• White-label reports & API
• Data Studio connector
• 15 logins for your team
• AI crawler tracking (180 days)
• 150 first drafts of what to publish / month

No card to start 7-day free trial. No payment details to sign up.

Nothing to install No script, no tag, no access to your site or analytics.

Cancel in one click From billing settings. No contract, no fee, no call.

Your data stays yours Trial ends? The account pauses and your history is kept.

Full feature comparison →

The questions people actually ask

Still not sure? The free test needs no account. Run it →

Want to know exactly how we measure? Read the methodology →

AI answers change every time you ask. How can tracking them mean anything?

It means something precisely because you are not relying on one answer. A single check is a coin flip; a repeated check is a signal. CitedOrNot re-queries each engine on its own adaptive schedule and records every result, so what you see is your mention rate over time, not a snapshot. When the trend genuinely moves, you get alerted. Every result also stores the full AI response verbatim, so you can read exactly what was said and judge it yourself rather than taking our word for it.

How do you decide whether AI "mentioned" us? Can I check your numbers?

A mention means your name — or a name you told us you go by — appearing in the text the AI actually wrote. Not an estimate, and not a guess at what was implied. Every result keeps the full answer and the date it was given, so you can open any row and read it yourself. Before-and-after results only appear once there are checks on both sides of the day you shipped, and always show how many there were. The full method is written up in plain English on our methodology page.

Can I not just ask ChatGPT myself once a week?

You can, and most people start there. It breaks down for three reasons: one answer is not a trend, you would have to repeat it across 7 engines and every buyer question you care about, and you have no record of what changed or when. CitedOrNot is that same habit run consistently, logged, compared against your competitors and turned into a to-do list — without you remembering to do it.

What do I actually do with the results?

More than look at them. CitedOrNot flags every query where a competitor is recommended and you are not, then builds recommendations from the actual AI answers — the content format to publish, the proof points to include, and the buyer query it should own. You get a prioritized to-do list, not another dashboard to interpret. (AI action plans are on Pro and Agency.)

Which AI engines does CitedOrNot track?

Seven: ChatGPT, Perplexity, Gemini, Claude, Grok, Google AI Overview and Google AI Mode. They are tracked separately, because being strong on one and invisible on another is the normal case, not the exception. Starter covers ChatGPT, Gemini and Perplexity; Pro and Agency cover all 7.

How is this different from SEO or rank tracking?

A rank tracker tells you your position in a list of blue links. CitedOrNot tells you whether AI names your brand inside its generated answer — on engines that publish no rankings at all and send no referrer you could measure in Analytics. Different metric, different playbook. This is GEO (Generative Engine Optimization), the layer that sits on top of SEO rather than replacing it.

Do I need to install anything or give access to my website?

No. No script, no tracking tag, no access to your site or analytics. You add your brand and the questions your buyers ask AI; CitedOrNot does the querying. That also means you can track competitors — or evaluate a brand before you own it.

Can I track competitors, not just my own brand?

Yes. Add the competitors you care about and every result shows who the AI recommended alongside or instead of you, per engine and per query. That competitor gap is usually the fastest path to winning more citations, because it tells you exactly which answers you need to be in.

How does monitoring work, and what counts as a check?

One check is one engine queried for one of your questions, so a question checked across all 7 engines uses 7 checks each time it refreshes. First results appear within minutes of setup; after that each engine re-checks on an adaptive schedule — faster when answers move, slower when they are stable — so your allowance goes where it matters. Monthly allowances: Starter 750, Pro 6,000, Agency 18,000.

What happens when the free trial ends?

Nothing is charged and nothing is deleted. The trial is 7 days and takes no credit card. If you do not upgrade, your account simply pauses with your data intact so you can pick it up later. Paid plans start at $29/mo and cancel from your billing settings in one click — no fees, no contract, no call with anyone.

Every feature, in one place

A tag shows the plan a feature needs. No tag means it's on every plan.

• 7 AI engines Pro+
• The questions buyers ask
• Competitor gaps
• Sources AI cites Pro+
• AI action plans Pro+
• First drafts
• Before-and-after proof Pro+
• Question discovery Pro+
• How AI describes you Pro+
• Alerts and digests
• 31 markets
• AI readiness audit
• Search Console link
• AI crawler tracking Starter+
• Team logins Starter+
• MCP connector Pro+
• Built for agencies Agency