AI Citations vs Mentions: The Difference and Why Both Count

An AI mention is when an answer names your brand in its text. An AI citation is when the answer links to a page as a source, whether that's your own page or a third-party page about you. They're different signals. Mentions shape what the buyer remembers and shortlists. Citations bring clicks and show which pages the engine relied on. The strongest outcome is both at once: named, recommended and linked.

On this page

• What is the difference between a mention and a citation?
• What do the four combinations look like?
• From MentionX data: how often mentions come without citations
• Why do both count?
• How do you log mentions and citations correctly?
• How do you score a mention properly?
• How do you turn citations into mentions?
• How do you turn mentions into citations?
• Worked example: Ledgerline's two gaps
• How do engines differ in how they cite?
• What should a monthly report show?
• Common mistakes
• Checklist
• Related
• Sources

Updated 31 October 2026. Anything marked "Our view" is MentionX's opinion, not documented engine behaviour. The worked example uses a fictional company.

What is the difference between a mention and a citation?

A third outcome sits above both: a recommendation, where the engine suggests your brand for the user's specific need. Treat it as the highest-value form of mention.

What do the four combinations look like?

Imagine the prompt "how do I reduce Shopify checkout abandonment".

1. Mentioned and cited. "Tools like YourBrand can recover abandoned carts [1]", where [1] links to your guide. Best case.
2. Mentioned, not cited. Your brand is named, but the sources are a review site and a Reddit thread. Your reputation is in others' hands, so check what those sources say.
3. Cited, not mentioned. The answer uses your blog's checklist, cites it, and names three competitors as tools. You did the teaching; they got the shortlist.
4. Neither. You are absent. Look at which sources were cited and why.

From MentionX data: how often mentions come without citations

Method. We ran read-only aggregate queries over completed MentionX runs from 22 August to 29 October 2026 (UTC): 2,154 runs across 18 projects, mostly B2B software. Engines were called through their APIs with web search or grounding switched on where available. "Mentioned" means our extraction found the tracked brand in the answer text. "Cited about the brand" means at least one returned citation was linked by our extraction to the tracked brand, either its own site or a third-party page about it. Brands are anonymised and no individual answers are shown.

Two things stand out. First, most mentions came without a source about the brand. Second, the "cited, not mentioned" case was rare in this data, because our citation measure only counts pages linked to the brand. A how-to page on your blog that the engine cited without naming you would count, but it didn't happen often.

Engines differ a lot in whether they cite anything:

When an engine cites nothing, every mention is by definition "named, not cited". So a low citation rate on DeepSeek tells you about the engine, not about your content.

Limits. This is a convenience sample of MentionX projects, not a market survey. Perplexity's sample is small. API results can differ from what a signed-in user sees in the consumer app.

Our view: report citation rate per engine, and alongside it the share of runs that cited anything. Without that second number, a drop in citations can look like your problem when the engine simply stopped searching.

Why do both count?

Mentions matter because answers are often the end of the journey. Many users read the answer and act on it without clicking. If your name isn't there, you're not on the list.

Citations matter for three reasons:

• They are the only part of the answer that sends measurable traffic, which you can track in GA4 (guide).
• They show that the engine can reach and use your pages, which confirms your technical setup works.
• They hint at what the engine trusts for that topic, which helps you plan content.

Google says AI Overviews and AI Mode may use "query fan-out", running several related searches, which lets them display a wider and more diverse set of supporting links than a classic search (Google Search Central). How often those links are clicked varies widely by query, so measure your own.

How do you log mentions and citations correctly?

Use one row per prompt per engine per run, with these columns:

Two rules keep the data honest:

1. Record all cited URLs, not just yours. The list of other sources is your off-page target list.
2. Run each prompt more than once, ideally three times, because answers vary. Count the rate, not the best result.

How do you score a mention properly?

Not every mention is worth the same. A yes/no "mentioned" column is a start, but three more details tell you whether the mention helps.

Position. Where answers rank brands in a list, being first matters. In MentionX data, when the tracked brand appeared in a list in an unbranded answer (n=102), it was first in 18% of cases, second or third in 14%, fourth or fifth in 13% and sixth or lower in 56%. Most list appearances were near the bottom.

Framing. "Ledgerline is the best fit for teams over 100 people" is a recommendation. "Other options include Ledgerline" is a mention. "Ledgerline is pricier and harder to set up" is a mention you'd rather not have. Log sentiment and whether the brand was suggested for the user's stated need.

Accuracy. Write down every wrong fact: pricing, features, integrations, company size. These are your most actionable findings, because each one points to a page somewhere that's either wrong or missing.

Our view: a mention with the wrong price, low in a list of eight, is worth less than no mention for some buyers. Score the quality, not just the count.

How do you turn citations into mentions?

If you're often cited but rarely named:

• Put your brand in the content the engine quotes. A checklist that says "Step 4: set up an exit-intent offer (in YourBrand this is under Campaigns > Triggers)" ties the instruction to your product without spam.
• Add a short, factual "Tools for this" section with your product and named alternatives. Engines like balanced comparisons; a fair list is more likely to be reused than a sales pitch.
• Use clear, consistent naming. If your product name is a common word, pair it with the category ("YourBrand email platform").

How do you turn mentions into citations?

If you're named but other pages get the links:

• Publish the page the engine is missing. If the cited source is a comparison table, create a better and more honest one on your domain.
• Answer first. Put a direct 40-60 word answer at the top so it's easy to extract.
• Check access. Confirm OAI-SearchBot and PerplexityBot are allowed in robots.txt (OpenAI, Perplexity).
• Add dated facts such as pricing, limits and benchmarks with a "last updated" note. Fresh, specific facts make a page worth citing.

Worked example: Ledgerline's two gaps

Ledgerline is a fictional accounts-payable tool for mid-sized companies. Its team runs 40 unbranded prompts across three engines, three times each, and logs mentions and citations separately.

The results split cleanly by prompt type:

• "How to" prompts ("how to automate invoice approval"): Ledgerline's guide is cited in about a third of runs, but Ledgerline is named in very few. The answer reuses its steps and lists two competitors as tools. This is the "cited, not mentioned" gap.
• "Best tool" prompts ("best AP automation for 200-person company"): Ledgerline is named in around half the runs, almost never with a link to its site. The sources are two review sites and an analyst roundup. This is the "named, not cited" gap.

The team fixes them differently. For the how-to guide, it adds a short "Doing this in Ledgerline" box after each step and a fair "Tools for this" section naming itself and two alternatives. For the best-tool prompts, it checks what the review sites say (one lists an old price), asks for a correction, and publishes an honest comparison page with a dated pricing table.

Six weeks later, the how-to prompts name Ledgerline more often. The best-tool prompts still mostly cite third parties, but now with the correct price. Illustrative: we haven't measured these specific effects, and your results will depend on the engines and prompts involved.

How do engines differ in how they cite?

The same brand can look very different per engine, so never average them blindly:

Keep the engine and mode in every row so you can tell a real change from a change in how the engine answered.

What should a monthly report show?

Report mention rate, recommendation rate and citation rate side by side, per engine, with one competitor line each. Add the top 10 cited domains in your category and flag the ones that don't mention you. That's your next quarter's off-page plan.

Common mistakes

• Counting a citation as a mention. If your page is a footnote but the answer recommends a competitor, you've lost the shortlist. Score the two separately.
• Only counting your own domain as a citation. A favourable review page about you is still a citation about you. Track both "own site cited" and "any source about us cited".
• Comparing citation rates across engines. An engine that rarely cites will always look worse. Compare each engine with itself over time.
• Treating "named, not cited" as a failure. Many buyers read the answer and never click. Being named for the right need is often the outcome that matters most.
• Checking once and calling it done. Answers vary between runs. A single run can show you named and cited one day and absent the next. Base every rate on repeated runs.

• Ignoring which competitors share the answer. In our data, unbranded answers named 1.11 brands on average when they had citations (n=1,002) and 1.44 when they had none (n=682). Uncited answers tended to list more brands, so you share the attention with more rivals.

Checklist

• Mentions and citations logged as separate columns
• "Own site cited" and "any source about us cited" tracked separately
• Share of runs with any citation recorded per engine
• Cited URLs saved for every "named, not cited" result
• Brand named in the parts of your content that engines quote
• Third-party pages about you checked for accuracy each quarter

Related

• How AI search builds an answer
• Earned media vs owned content
• Track AI traffic in GA4

Sources

• Google Search Central, AI features and your website
• OpenAI, Overview of OpenAI crawlers
• Perplexity, Perplexity crawlers