The 12 AI Mention States

Every brand now has an answer to a question it never used to have to answer: what does AI say about us — and when?

Not "do we rank." Not "do we have backlinks." Not even "do we have good content." The question is narrower and more specific than that, and it's the one that determines whether you show up in the moment a buyer is actually forming a decision: at which stage of the buying journey does AI mention your brand, and in what light?

That's the question this framework answers. We call it the 12 AI Mention States — a diagnostic model that classifies every brand into exactly one of twelve positions based on how it performs across four stages of the AI-mediated buying journey: Top of Funnel, Middle of Funnel, Bottom of Funnel, and Evaluation (post-purchase recommendation).

This matters because generative engines (ChatGPT, Perplexity, Gemini, Claude, Copilot, and the AI Overviews sitting on top of Google) don't treat your brand as a single, static entity. They treat you as a shifting probability that changes at every stage of a buyer's research. A brand can be everywhere at the awareness stage and nowhere at the decision stage. A brand can dominate branded queries and disappear the moment a prospect asks an AI to compare it to a competitor. This is normal. It's also fixable — but only once you know exactly where the breakdown is happening.

That's what this article is for. Below, we define all twelve states, show you exactly what each one looks like in the wild, tell you what to look for when diagnosing your own brand (including the specific fan-out queries and cited domains that reveal each state), and lay out the tactical GEO (Generative Engine Optimization) moves required to fix it.

Part 1: The TMBE Framework

Before the twelve states make sense, you need the grid they're built on.

We score every brand across four stages of the buying journey, using a simple traffic-light system: Red (R) = absent or actively harmful, Yellow (Y) = inconsistent or weak, Green (G) = strong and consistent.

Stage	Code	What's happening in the buyer's head	Typical query shape
Top of Funnel (Awareness / Research)	TOFU	The buyer doesn't know your name yet. They're asking AI to educate them on a category, a problem, or a use case.	"What should I look for in a [category]?"
Middle of Funnel (Comparison)	MOFU	The buyer has a shortlist. They're asking AI to compare named options, weigh trade-offs, and narrow the field.	"[Brand A] vs [Brand B] vs [Brand C] — which is better for [use case]?"
Bottom of Funnel (Decision)	BOFU	The buyer is close to committing. They're asking AI to validate a near-final choice, check pricing, or confirm fit.	"Is [Brand] a good choice for [specific scenario]?"
Evaluation (Post-Purchase Recommendation)	EVFU	The buyer already owns your product. AI is now evaluating whether the brand is trusted, recommended, and worth sticking with, shaping renewal, loyalty, and word-of-mouth.	"Do customers tend to recommend [Brand] after using it?" / "Is [Brand] considered trustworthy in [category]?"

A brand's mention state is simply its four-letter combination (TOFU, MOFU, BOFU, EVFU) read left to right. There are 3⁴ = 81 mathematically possible combinations, but in practice, brand visibility clusters into twelve recognizable, recurring patterns. Those are the twelve states below.

Why this beats a single "visibility score"

Most AI-visibility tools give you one number: a percentage of prompts where you were mentioned. That number is close to useless on its own, because it collapses two brands with wildly different problems into the same score. A brand that's mentioned constantly at the research stage but vanishes at the decision stage has a conversion problem. A brand that's invisible at research but wins every branded query has a discovery problem. Same overall "visibility," completely different fix. The TMBE grid exists to stop that collapse — it forces you to diagnose by stage, not by average.

Part 2: What to Look For at Every Stage

Regardless of which state you land in, diagnosing it correctly depends on gathering the same three categories of evidence at each funnel stage. This is the diagnostic core of the whole exercise — and it's exactly what a platform like SurfaceMap.cc is built to surface automatically, stage by stage, rather than through one-off manual prompting.

1. Fan-out queries, segmented by stage

Generative engines don't answer a single prompt — they internally "fan out" a user's question into multiple sub-queries, retrieve sources against each, and synthesize the result. To diagnose your state accurately, you need to test with a representative sample of fan-out queries at each stage, not just your brand name. Examples:

TOFU fan-out: "best [category] for [industry]," "how to choose a [category] provider," "what does a [category] solution actually do," "[category] buying guide [current year]"
MOFU fan-out: "[Brand] vs [Competitor]," "[Brand] alternatives," "top 5 [category] platforms compared," "[Competitor] vs [Brand] pricing"
BOFU fan-out: "is [Brand] worth it," "[Brand] reviews [current year]," "[Brand] for [specific use case] — pros and cons," "should I choose [Brand] or [Competitor] for [scenario]"
EVFU fan-out: "[Brand] add-ons," "how to get more out of [Brand]," "[Brand] vs [Competitor] for scaling teams," "switching from [Brand] to [Competitor]"

2. Cited domains, at every stage

Every AI answer is built from sources it trusts enough to cite or pull structured data from. The single highest-leverage diagnostic action you can take is logging which domains get cited at each stage (the answer is almost never the same domain twice across the funnel). A category-defining publication might dominate TOFU citations; a review aggregator or comparison site might dominate MOFU; G2, Capterra, or Reddit threads might dominate BOFU; a knowledge-base or community forum might dominate EVFU. If you don't know which domains are being cited at the stage where you're weak, you don't have a content plan — you have a guess.

3. Currently ranking / cited competitor content

For every fan-out query where your brand doesn't appear, identify which specific competitor page is winning the citation — not just which competitor, the actual URL. Pull it apart:

What structured data does it carry (schema.org markup, FAQ blocks, comparison tables)?
What format is it (long-form guide, comparison page, review roundup, Reddit thread, YouTube transcript)?
What entities and claims does it explicitly name that your content doesn't?
How recently was it updated?

This is the content you are functionally competing against for the citation slot — and reverse-engineering it is faster and more reliable than guessing what AI "wants."

SurfaceMap.cc automates all three of these (pulling fan-out queries by funnel stage, logging cited domains and their frequency, and flagging the specific competitor URLs winning citations where you're absent), but you can run this manually with a spreadsheet and patience. What matters is that you do it by stage, every time.

Part 3: The Twelve States

#	State	TOFU	MOFU	BOFU	EVFU
1	Invisible Brand	R	R	R	R
2	Customer Secret	R	R	R	G
3	Category Ghost	R	Y	Y	Y
4	Research Casualty	G	R	R	R
5	Traffic Donor	G	Y	R	R
6	Comparison Drop-Off	G	G	R	Y
7	Branded Bubble	R	R	G	Y
8	Decision Weakling	G	G	Y	R
9	Proof Deficit	Y	G	R	R
10	Post-Purchase Silence	G	G	G	R
11	Expansion Blindspot	Y	Y	Y	R
12	Interchangeable Vendor	G	G	G	G

Below, each state gets a full breakdown: what it means, what it looks like in real query testing, what to look for diagnostically, and the fix.

1. Invisible Brand — R R R R

AI rarely mentions you at any buying stage.

This is the starting point for most brands the first time they actually test AI visibility, and it's more common than the marketing world likes to admit. Your website may rank fine on Google. Your sales team may be closing deals. None of that means AI models have any confidence citing you, because generative engines are pulling from a different evidence base than classical SEO rewarded — structured entity data, third-party corroboration, and content that directly answers synthesized sub-queries, not just keyword-matched landing pages.

What it looks like: Run 20 fan-out queries across all four stages and your brand name appears in zero to one of them, and never in a citation — only, at best, as a passing mention buried in a list.

What to look for:

Check whether you have any structured data (Organization, Product, Service schema) that would let a model resolve "who you are" as an entity distinct from your competitors.
Check whether your name appears anywhere in the cited domains for TOFU queries in your category — trade publications, "best of" roundups, Reddit, G2.
Check whether AI models can even correctly describe what you do when prompted directly ("What is [Brand]?"). If the model hallucinates or has no data, that's your root cause.

The fix:

Establish baseline entity clarity first — schema.org markup (Organization, Product/Service, FAQPage) that lets models correctly resolve who you are, what you sell, and who it's for. This is prerequisite infrastructure, not a nice-to-have.
Get corroborated in the TOFU cited domains for your category — contribute to, get reviewed by, or get referenced in the specific publications and forums models are already pulling from (identified via the diagnostic above).
Publish direct-answer content that mirrors the actual fan-out query language buyers use, not brand-first copy.
Re-test monthly. Invisibility is a data-availability problem before it's a content-quality problem — fix the availability first.

2. Customer Secret — R R R G

Loved by clients. Unknown to new buyers.

This state is common among service businesses, B2B vendors with strong renewal rates, and companies with word-of-mouth-driven growth. Your existing customers get reinforced by AI (it recommends your add-ons, defends your value against churn-risk queries) — but nobody outside your existing base ever discovers you through AI-mediated research.

What it looks like: EVFU fan-out queries ("what should I add to [Brand]," "is [Brand] still the right fit as we scale") return confident, positive answers. TOFU, MOFU, and BOFU queries return nothing.

What to look for:

Compare your citation frequency in your own knowledge base / help center (likely well-indexed and cited for EVFU) against your near-zero citation frequency in third-party TOFU/MOFU/BOFU content.
Look for a gap between your G2/Capterra review volume (often strong, since happy customers write reviews) and your visibility in comparison content — reviews alone don't get you cited at the comparison stage without structured aggregation.

The fix:

Mine your own customer proof (case studies, reviews, outcome data) and repurpose it outward — turn EVFU-only proof into TOFU/MOFU/BOFU content: comparison pages, "why customers choose us over X" pages, outcome-based landing pages that a model can cite when a new buyer asks a category question.
Get that proof placed in the third-party domains models cite at TOFU and MOFU — don't just publish it on your own site.
Build comparison content proactively, because right now competitors are the only source models have for how you stack up.

3. Category Ghost — R Y Y Y

AI discusses your category without naming you.

This is a subtler and more dangerous version of invisibility. Models clearly understand your category (they generate confident, detailed answers about it), but you're structurally excluded from the answer even at comparison and decision stages. This usually means the category-defining content (analyst reports, "best X tools" roundups, Wikipedia-adjacent sources) has been built without you, and models are anchoring on that content as ground truth.

What it looks like: Ask "what are the leading options for [category]" and get a confident five-name list — none of them you. Ask "[Brand] vs [Competitor]" directly and you get a thinner but real answer (hence Yellow, not Red, at MOFU/BOFU).

What to look for:

Identify the specific "best of" / "top X" articles being cited for TOFU category queries and check publication date and update cadence — these lists are often stale and represent a closed door you need a specific plan to open (outreach, sponsored inclusion, or displacing them with your own comparison content).
Check whether analyst or industry-body content mentions your category taxonomy at all without naming vendors — that's a sign the category itself needs a knowledge-graph presence before individual brand inclusion is possible.

The fix:

Target inclusion in the specific roundup/list content already winning TOFU citations — this is an outreach and PR play, not just a content play.
Build and structure your own category-definition content (with proper schema markup) so models have an alternative, citable source that includes you by default.
Make sure Wikidata, Wikipedia, and industry-body listings correctly categorize you — these are disproportionately influential as grounding sources for category-level AI answers.

4. Research Casualty — G R R R

You're mentioned early, erased during evaluation.

You show up when buyers are learning about the category (often because you've invested in strong educational content), but the moment the buyer asks AI to actually compare or decide, you vanish. This is one of the most common and most frustrating states, because it means your top-of-funnel investment is generating awareness for a purchase you never get credit for.

What it looks like: TOFU queries return you prominently, often first. MOFU "vs" and comparison queries omit you entirely, even against competitors you objectively compete with on the ground.

What to look for:

Identify who is showing up in the MOFU comparisons your TOFU content should be feeding into — pull the actual competitor comparison pages winning those citations and check what they contain (feature tables, pricing, named use cases) that your site doesn't.
Check whether you have any comparison-format content at all. Educational content and comparison content are structurally different documents to a retrieval system — a blog post that explains a category will almost never surface for a "X vs Y" query.

The fix:

Build direct comparison content — "[Brand] vs [Competitor]" pages for every competitor you're actually up against, with explicit, structured trade-off data (pricing, features, ideal-fit scenarios). Vague positioning does not survive a fan-out comparison query; specifics do.
Add comparison and decision-stage schema (Product, Offer, FAQPage answering "how does X compare to Y") to make this content machine-legible, not just human-readable.
Audit your TOFU content for calls-to-comparison — internally link and reference your own comparison pages so crawlers associate the educational and evaluative content as one entity story.

5. Traffic Donor — G Y R R

You educate. Competitors close the deal.

A more advanced version of the Research Casualty problem — you're contributing meaningfully to buyer education (TOFU strong, MOFU inconsistent) but you're actively losing at the decision stage, meaning your content is doing the unpaid work of warming up buyers who then get closed by a competitor cited more confidently at BOFU.

What it looks like: TOFU strong, MOFU shows up but thin/inconsistent, BOFU ("is [Brand] worth it," "[Brand] reviews") returns weak or negative signal, or surfaces a competitor's proof instead of yours.

What to look for:

Look specifically at what's cited for BOFU validation queries in your category — this is almost always third-party proof (reviews, case studies, independent comparisons), and if a competitor's proof dominates that citation slot, that's your specific leak point.
Check your own review volume and recency on the platforms models trust for BOFU validation (G2, Capterra, TrustRadius, industry-specific review sites) against competitors'.

The fix:

Prioritize review and third-party validation velocity — this is the single highest-leverage fix for a Traffic Donor, because BOFU citation is disproportionately weighted toward independent proof, not owned content.
Publish outcome-specific case studies structured around the exact decision-stage query language ("results after switching to [Brand] for [use case]") rather than generic success stories.
Close the loop with a pricing/fit page that directly answers "is [Brand] worth it" — many brands never publish content that actually answers this exact question in these exact words, leaving the answer to whoever does.

6. Comparison Drop-Off — G G R Y

Shortlisted briefly, replaced by competitors later.

You make it further than a Research Casualty (you're present in both awareness and comparison content), but something breaks specifically at the final decision moment, and you have a partial, inconsistent presence in evaluation-stage queries too. This state often points to a genuine trust or proof gap that surfaces right when a buyer is asking AI to validate a near-final choice.

What it looks like: You appear in "best X" and "X vs Y" content confidently. But "is [Brand] the right choice for [scenario]" or "[Brand] reviews" queries return thin, dated, or absent answers — or worse, surface unresolved complaints.

What to look for:

Pull every BOFU fan-out query and check literally what sources get cited in the answer — often this reveals a small number of negative or outdated review threads with outsized influence because nothing more recent or more positive is competing for that citation slot.
Check content recency specifically at the decision-stage layer — many brands keep their homepage and blog current but let pricing pages, review responses, and validation content go stale, which is exactly the layer models pull from at BOFU.

The fix:

Actively manage and refresh third-party review presence — respond to negative reviews (models will surface resolution, not just complaint, if it's there), and drive fresh review volume on a consistent cadence.
Publish scenario-specific validation content ("[Brand] for [specific team size / use case / industry] — what to expect") that directly answers the fan-out variants of "is this right for me."
Extend structured proof into the expansion layer too, since your EVFU score is only Yellow — make sure post-purchase content (onboarding guides, expansion use cases) is present enough that AI doesn't have to guess at retention questions either.

7. Branded Bubble — R R G Y

You win name searches, lose competitive queries.

You have strong presence when someone already knows your name and asks a direct decision question about you — but if a buyer hasn't heard of you yet, or is comparing you to alternatives, you don't exist. This state is common among brands with strong direct/branded search history whose GEO presence has never been built independently of brand recognition.

What it looks like: "Is [Brand] good for [use case]" returns a solid, confident answer. "Best [category] tools" and "[Brand] vs [Competitor]" return nothing or omit you.

What to look for:

Confirm the gap is real, not a sample-size artifact — test at least 15–20 non-branded TOFU and MOFU queries across query phrasings before concluding you're absent, since fan-out variance is real.
Check what percentage of your existing content is written from a "you already know us" posture (product pages, feature pages) vs. a "here's how we compare / here's the category" posture. Branded Bubble brands are usually 90%+ the former.

The fix:

Build outward-facing, non-branded content deliberately (category education and named comparisons), because your branded-query strength proves the underlying product story works; it's simply never been translated into content a stranger's query would surface.
Get corroborated in third-party comparison and "best of" content, since your own site alone will rarely outrank the aggregators and roundups that already own these citation slots.
Treat this as a content-gap problem, not a product or trust problem — the fastest-fixing of the twelve states, because the underlying proof (you already win BOFU) already exists.

8. Decision Weakling — G G Y R

AI mentions you, but rarely recommends you.

You have broad presence through TOFU and MOFU, and partial presence at BOFU — but the quality of the mention is weak. You're listed, not recommended. You appear as "another option" without a decisive endorsement, and expansion content is absent entirely.

What it looks like: Comparison answers name you alongside competitors but consistently give the confident recommendation to someone else — "for most teams, X is the better choice, though [Brand] is also an option." The pattern repeats across multiple fan-out phrasings.

What to look for:

Look at the language of BOFU mentions, not just presence/absence — search for hedge words ("also," "another option," "worth considering") vs. decisive language ("best for," "ideal if") attached to your name vs. competitors'.
Identify the specific claims or differentiators competitors have that you don't — usually a named use case, a specific metric, or a category leadership claim that's been explicitly and repeatedly corroborated across sources.

The fix:

Sharpen and repeat a specific, ownable differentiation claim across every piece of content and every third-party source you influence — vague positioning produces vague, hedge-language mentions; specific claims ("fastest onboarding for teams under 50," "only platform with X certification") produce decisive ones.
Get that specific claim corroborated in third-party sources, not just stated on your own site — models weight repeated, cross-source claims far more heavily than single-source assertions.
Build out evaluation-stage content since EVFU is fully absent — a Decision Weakling that fixes BOFU language without addressing EVFU risks winning the sale and then losing the renewal to the same hedge-language problem one stage later.

9. Proof Deficit — Y G R R

AI can't justify why you're the better choice.

You show up inconsistently at awareness stage but reasonably well in comparisons — models know enough to include you when a buyer asks "X vs Y." But once the conversation moves to actually justifying a choice (decision) or defending it after purchase (expansion), there's no evidence base for the model to draw on, and you drop out.

What it looks like: Comparison tables include you with feature-level facts (accurate, but generic) — no outcome data, no named customer, no third-party validation. BOFU and EVFU queries surface nothing because there's no proof layer to cite.

What to look for:

Audit what data actually exists about you across the internet that a model could cite as proof rather than description — case studies with numbers, review excerpts, analyst mentions, named-customer outcomes. Most Proof Deficit brands find this layer is nearly empty even when their descriptive content is fine.
Check whether competitors winning BOFU citations have quantified outcomes (X% improvement, $ saved, time reduced) attached to their mentions, and whether you have any equivalent data published anywhere, in any format.

The fix:

Produce and publish quantified proof (specific, numbered outcomes tied to named or anonymized customers), because this is the single asset class missing, not general content volume.
Prioritize placing that proof in third-party, independently-hosted locations (review platforms, case study aggregators, press) since self-published proof is weighted lower than externally corroborated proof.
Layer proof into both BOFU (why choose us) and EVFU (why stay / expand) content — a proof deficit rarely respects funnel boundaries, so fix both simultaneously.

10. Post-Purchase Silence — G G G R

Customers buy. AI never reinforces loyalty.

This is a strong state overall (you're winning awareness, comparison, and decision), undermined by one specific, high-value gap: nothing happens for your customers after they buy. AI never reinforces the relationship, never surfaces upsell paths, and worse, will happily recommend a competitor if an existing customer asks an evaluative question post-purchase.

What it looks like: Everything pre-purchase looks healthy. But "should I switch from [Brand]" or "[Brand] alternatives" queries (asked from an existing-customer framing) return neutral-to-positive answers about competitors with no defense of your own retention case.

What to look for:

Test EVFU fan-out queries specifically from a churn-risk framing ("outgrowing [Brand]," "is there something better than [Brand] for scaling teams") — this is the exact query shape AI-savvy competitors are already targeting to win your existing customers.
Check your knowledge base, help center, and customer-only content for indexability and citation — if it's gated or has no schema, models literally can't retrieve it to defend you in these moments.

The fix:

Build public, indexable, schema-marked customer success and expansion content ("getting the most out of [Brand]," advanced use cases, upgrade paths) so models have something to cite when an existing customer's query starts drifting toward a competitor.
Directly address the switching-cost and stay-vs-leave query shape with content that isn't purely defensive — show growth paths within your product, not just reasons not to leave.
Since this is otherwise a strong-performing brand, treat this as the highest-ROI single fix available: you've already won the sale, this content protects revenue you already have rather than chasing revenue you don't.

11. Expansion Blindspot — Y Y Y R

AI doesn't surface your upsell paths.

A broader, earlier-stage version of Post-Purchase Silence — presence is inconsistent across the entire pre-purchase journey (not fully absent, not fully strong) and completely absent post-purchase. This state often shows up in mid-market or category-diversifying brands where the core offering is known but adjacent products, tiers, or expansion use cases have never been separately established in AI-retrievable content.

What it looks like: Core-product queries return moderate, inconsistent presence throughout TOFU/MOFU/BOFU. Any query about add-ons, upgrades, or "what else does [Brand] offer" returns nothing, even though the product/tier actually exists.

What to look for:

List every product, tier, add-on, and expansion path you actually sell, then individually test whether each one has any independent AI-retrievable footprint — most Expansion Blindspot brands find their flagship product has content but every adjacent offering has none.
Check whether your own site has dedicated, schema-marked pages per product/tier, or whether expansion offerings are buried inside a single generic pricing page a model can't parse into distinct, citable entities.

The fix:

Give every expansion product or tier its own dedicated, schema-marked page (Product/Offer schema per SKU or tier) — an undifferentiated pricing page is functionally invisible to entity-based retrieval even if a human can read it fine.
Fix the inconsistent TOFU/MOFU/BOFU presence in parallel, not sequentially — Expansion Blindspot brands often mistakenly treat the core-product weakness and the expansion gap as two different projects, when the same entity-clarity and structured-data fixes solve both.
Publish direct-answer content for "what else can I do with [Brand]" and "[Brand] add-ons" phrasing specifically — this is a distinct query shape from the core-product queries and needs its own content, not an assumption that broad brand content will cover it.

12. Interchangeable Vendor — G G G G

AI treats you as one of many options.

The final state is the most counterintuitive one, because on paper it looks like success: you're present, consistently, at every single stage of the funnel. The problem isn't visibility anymore — it's differentiation. You show up everywhere, alongside everyone else, with no distinguishing recommendation logic attached to your name. You've won the visibility game and are now competing purely on the margin of specific, ownable claims.

What it looks like: Every fan-out query at every stage returns you, reliably — but always in a list, always described in the same generic terms as competitors, never singled out as the recommended pick for a specific scenario.

What to look for:

Compare the language attached to your mentions against a genuinely differentiated competitor's mentions in the same space — look specifically for whether anyone gets singled out for a named use case, persona, or scenario, and whether it's ever you.
Identify whether you have any claim, certification, feature, or outcome that is genuinely unique to you (not just true of you) — Interchangeable Vendors often have real differentiators that were simply never surfaced distinctly enough for a model to attach them specifically to your name rather than treating them as category-standard.

The fix:

Identify and aggressively narrow to your single most defensible, specific differentiation claim, then build content, schema, and third-party corroboration exclusively around that claim rather than continuing to publish broad, category-standard content that reinforces sameness.
Target specific personas, use cases, or scenarios explicitly rather than trying to remain broadly appealing — "best for [narrow, specific segment]" content is what breaks a model out of treating you as interchangeable, because it gives the model a retrieval reason to select you over an equally-visible competitor for a specific query.
Recognize this is a genuinely different problem than every other state on this list — you don't have a visibility problem to fix, you have a distinctiveness problem, and the tactics that got you full green coverage (broad, consistent, well-structured content) will not be the tactics that fix this. This state requires narrowing, not more of the same.

Part 4: How to Diagnose Your Own State

You don't need twelve separate audits to find your state — you need one disciplined process, run consistently:

Build your fan-out query set. Roughly 5–6 representative queries per funnel stage (20–24 total), written in the actual language buyers use — not brand-first phrasing.
Run them across the major generative engines — at minimum ChatGPT, Perplexity, Gemini, and Google AI Overviews, since citation behavior differs meaningfully across platforms.
Score each stage R/Y/G based on presence, prominence, and sentiment of your mention (not just presence — a hedge-language mention is Yellow even if you technically appear).
Log the cited domains at every stage, independent of whether you appeared, so you know exactly who owns the citation slot you need to win.
Pull the specific competitor content winning citations at your weakest stage(s) and reverse-engineer its structure, claims, and schema.
Match your four-letter result to the twelve states above and run the corresponding fix sequence.
Re-test on a fixed cadence (monthly, at minimum), because citation patterns shift as models retrain, re-index, and as competitors publish new content. Your state is not permanent in either direction.

This is exactly the workflow SurfaceMap.cc is built to run continuously rather than as a one-time audit (tracking fan-out queries by funnel stage, logging cited domains over time, and flagging exactly which competitor content is winning the citation slots you're losing), but the diagnostic logic holds regardless of what tooling you use to execute it.

Closing

The old SEO question was "do we rank." The new question is narrower, more consequential, and stage-specific: at which moment in the buyer's journey does AI actually say your name — and does it say it with confidence?

Every brand reading this is currently living in one of these twelve states, whether or not anyone at the company has measured it yet. The brands that win the next decade of category leadership won't be the ones with the most content. They'll be the ones who diagnosed their actual state honestly, stage by stage, and built the specific, structured, corroborated proof needed to move one state closer to green — at every single stage, not just the one that was easiest to fix.

Find your state. Then go fix it.

See your own brand's mention state

SurfaceMap.cc runs this exact diagnostic continuously — fan-out queries by funnel stage, cited domains over time, and the competitor URLs winning the citations you're losing.

See Where You Stand