AI Visibility Field Guide: Citation Labs Worked Example

We ran Citation Labs through our own Field Guide System as a worked example.

We started with a question: why are more prospects coming to us for link building than AI visibility, and is our online footprint influencing that?

The Field Guide built a prompt set and tested Citation Labs across ChatGPT, Gemini, and Perplexity. We appeared in 0 of 15 unbranded AI-visibility prompts.

From there, it identified two clear gaps: inconsistent AI-visibility positioning across our own site, and weak representation in the third-party sources AI tools use when recommending providers.

The rest of this document shows the diagnosis, the changes it recommended, and how we’ll measure whether they work.

Once we’ve made the changes and re-run the same tests, we’ll publish the results.

Our problem statement (starting prompt after uploading):

So, I think organic traffic has kind of sort of always been low, but recently we keep getting meetings booked by people who are interested in link building and not AI visibility. And I’m wondering if our overall footprint is pushing people or pushing AI tools to recommend us as link builders and not AI visibility consultants, AIO, GEO, type consulting work. And, like, how much work needs to be done on and off page to help change that.

Prompts the Field Guide Built For Tracking

The Field Guide turned our starting question into a fixed set of prompts to track across ChatGPT, Gemini, and Perplexity.

These prompts established the baseline for measuring whether Citation Labs appeared in relevant AI-visibility searches, how the brand was described, and whether that changed after our interventions.

You can see these prompts here.

Preliminary diagnosis

Observed in this ChatGPT baseline:

• Citation Labs appeared in 0/9 unbranded AI-visibility/GEO/AEO prompts.
• iPullRank appeared in 9/9. Amsive and Directive appeared in 6/9 each.
• When asked directly what Citation Labs does, ChatGPT knows about the AI-visibility offering, but consistently frames it as secondary to the established link-building business:
  • “SEO and link-building agency”
  • “best known for enterprise-scale link building and SEO outreach”
  • AI/LLM visibility described as something it has “more recently” expanded into.
• Citation Labs appeared in only 1/3 generic link-building prompts, so I would not conclude “ChatGPT recommends Citation Labs everywhere for link building.” The stronger conclusion is: ChatGPT understands Citation Labs primarily through a link-building frame, while Citation Labs has essentially no unprompted category presence for AI visibility in this test.
• In the brand-specific answers, ChatGPT relied entirely on Citation Labs’ own site. Several cited pages themselves reinforce the old category, including the homepage surfaced as “Enterprise Link Building Agency & Tools.”

That is a meaningful machine-facing result. It does not establish that this is why your booked meetings skew toward link building. That remains a downstream hypothesis. The guides explicitly require us to keep those two claims separate.

The interesting part

This does not currently look like a simple “we need more off-page GEO mentions” problem.

The evidence suggests at least an owned-footprint issue may exist:

ChatGPT can find Citation Labs when explicitly asked about it. It even finds the AI offering. But the information it retrieves tells a coherent story of:

established enterprise link-building specialist → newer AI-visibility extension

That is different from:

established AI-visibility consultancy with link-building expertise as one capability.

At the same time, every unbranded AI query produced other firms and sources, with no Citation Labs evidence entering the answer environment at all. So there may also be an external/category-authority problem.

That gives us a working hypothesis of mixed owned + off-site, but the guide says not to choose the intervention until we inspect the source environment.

These results are enough. Stop testing for now.

What the baseline says

1. Citation Labs has effectively no unbranded AI-visibility category presence in this test.

Across:

• 9 ChatGPT AI-visibility/problem prompts
• 5 Perplexity prompts
• 5 Gemini prompts

Citation Labs appeared 0/19 times.

That is the strongest result.

Meanwhile, the same names recur repeatedly. In the 10 matched Gemini/Perplexity AI prompts:

• iPullRank: 7/10
• First Page Sage: 5/10
• Omniscient Digital: 5/10
• Directive: 4/10
• Go Fish Digital: 4/10
• Siege Media: 4/10

So this is not just one model behaving oddly.

2. When you name Citation Labs, all three systems understand the AI offering, but they still anchor the company in link building.

ChatGPT calls Citation Labs an “SEO and link-building agency” and describes AI/LLM visibility as a more recent expansion.

Perplexity says “enterprise SEO link building and AI-search visibility.”

Gemini says “enterprise-level link building and AI search/LLM visibility optimization,” then leads its capability list with enterprise link building.

So the footprint does communicate AI visibility. It just does not yet make Citation Labs a default member of the AI-visibility consultancy category.

3. The original hypothesis needs narrowing.

Citation Labs only appeared in 1/5 generic link-building recommendation tests across the three environments.

So I would not conclude:

AI systems are broadly recommending Citation Labs as link builders instead of AI consultants.

The evidence supports something narrower:

When AI systems already know Citation Labs is the subject, they still understand the company through a strong link-building frame. But when buyers ask generically for AI-visibility firms, Citation Labs does not enter the consideration set at all.

That distinction matters.

I also checked the current public footprint

There is a genuine owned-site coherence problem.

The homepage itself has moved substantially toward AI visibility. Its H1 is now “Shape how your brand appears in the AI answer space,” and it explicitly offers AI Advisory alongside link building. (Citation Labs)

But several strong signals still tell a different story:

• The homepage is still surfaced with the title “Enterprise Link Building Agency & Tools.” (Citation Labs)
• The About page literally opens with “We’ve built a…link-building agency” and describes the team as link-building strategists. (Citation Labs)
• The AI service is called “LLM Mentions” in the site navigation, while the homepage calls it “AI Advisory.” (Citation Labs)
• Even the AI service page foregrounds 20+ years of link-building experience, 195,468 links built, and link-building revenue proof. (Citation Labs)
• The case-study library is still visibly dominated by link-building wins, although there is now a strong AI case study. (Citation Labs)

At the same time, Citation Labs has meaningful AI evidence already: current AI research, the ZenBusiness AI case study, Search Engine Land work, LinkedIn positioning, Xofu, and a large amount of recent AI-focused content. (Citation Labs)

Working diagnosis

Mixed problem, but not “we need to publish a bunch more GEO content.”

My current read:

Owned footprint: inconsistent category signals. Citation Labs has enough AI material for models to understand the capability when prompted, but the company-level positioning still repeatedly resolves back to link building.

External footprint: likely the larger issue for unbranded discovery. The systems repeatedly assemble AI-consultancy shortlists from competitors and third-party/category sources where Citation Labs is absent. The fact that the models can retrieve Citation Labs’ AI work when you name the brand, yet never surface it in 19 unbranded prompts, points strongly toward a category/source-environment gap.

That is still an inference, but it is now a well-supported one.

The field guide says this is exactly where we stop saying “get links” or “write more

Source-gap map

Proceed with a mixed intervention, weighted toward off-site/category evidence.

My working allocation for the first cycle would be roughly 35% owned-site cleanup / 65% external footprint work. That is a prioritization, not a measured causal split.

The reason is specific: Citation Labs already has substantial AI expertise and proof. The problem is that it is not packaged consistently enough as an AI-visibility consultancy on owned properties, and it is largely absent from the external sources that define the consultancy category.

1. Source-gap diagnosis

Your 19 unbranded AI prompts show a strong pattern in the source layer:

Citation Labs: 0/19.

More interestingly, ChatGPT and Gemini leaned heavily on provider-owned pages, while Perplexity leaned much more heavily on third-party roundups, comparisons, case studies and category pages.

That tells us we have to solve two different things.

Owned: competitors unmistakably declare the category

iPullRank’s navigation contains Generative Engine Optimization, AI Search Strategy and Generative AI. Its AI Search Strategy page explicitly calls the company a pioneering AI Search agency, explains its audit/measurement/omnimedia methodology, and publishes AI-search result proof. (iPullRank)

Go Fish has gone even further. Its homepage now leads:

“Get cited in AI. Get found in Google. Get chosen by buyers.”

It immediately identifies itself as a generative engine optimization agency, before describing its broader services. (Go Fish Digital)

Directive has a page literally titled “Generative Engine Optimization (GEO) Agency for B2B”, says what the service does and follows it with an AI-search case study. (Directive)

Seer packages a diagnostic GEO offer, source/citation analysis, measurement, testing and years of published AI research that has been referenced by major publications. (Seer Interactive)

This is the competitive pattern: category declaration + methodology + proof + repeated category language.

Citation Labs has the substance, but the entity story conflicts

Your current homepage is materially better than the AI tools’ branded descriptions imply. It now leads with:

“SHAPE HOW YOUR BRAND APPEARS IN THE AI ANSWER SPACE”

and offers AI Advisory alongside link building. But the page is still surfaced publicly with the title “Enterprise Link Building Agency & Tools.” (Citation Labs)

The bigger contradiction is the About page. It still calls Citation Labs a “link-building agency,” describes the team as link-building strategists, and states a mission centered on delivering hundreds of links every month. It does contain the newer Citation Optimization story, but the entity-level description remains link-building-first. (Citation Labs)

The case-study index reinforces that. There is now a legitimate AI visibility case study, but the visible case library is still overwhelmingly link-building work. (Citation Labs)

This explains why all three models can understand the AI capability when asked about Citation Labs while still answering:

Citation Labs = link-building company that also does AI visibility.

You don’t have an AI-content shortage. You have a category coherence problem.

2. You already have enough AI proof

I would not commission another batch of GEO articles.

You already have:

• a 4,579-prompt AI visibility study with a concrete measurement methodology; (Citation Labs)
• the ZenBusiness case showing movement in AI Mode and Gemini after an off-domain intervention; (Citation Labs)
• the 48-participant AI Mode buyer-shortlisting study; (Citation Labs)
• a developed Citation Optimization methodology/framework; (Citation Labs)
• an active body of AI-search research and commentary. (Citation Labs)

That is enough raw material to establish expertise.

The problem is that the provider entity isn’t absorbing that evidence cleanly.

3. First owned-site intervention

I would make five concentrated changes, not redesign the website.

A. Fix the homepage title/category metadata

The indexed title cannot continue to lead with:

Enterprise Link Building Agency & Tools

while you’re trying to enter AI-consultancy recommendation sets.

Something in this territory:

AI Visibility Consulting & GEO Services | Citation Labs

The H1 can remain more differentiated. The metadata/entity descriptors need to carry the explicit market category.

B. Establish one canonical AI service

Right now the public footprint contains language including:

• AI Advisory
• LLM Mentions
• AI visibility
• Citation Optimization

Those concepts can coexist, but one must be the commercial category anchor.

I’d make it:

AI Visibility Consulting

with GEO/AEO/AI Search Optimization explicitly incorporated into the page copy and metadata.

Then Citation Optimization can be your methodology, rather than asking the model to infer that Citation Optimization means “firm I should hire for GEO/AEO.”

This is important. Buyers and models don’t have to adopt your proprietary terminology before they can categorize you.

C. Replace /llm-mentions/ as the primary service identity

The existing page is still publicly indexed as “LLM Mentions: Increase Visibility in AI Search Tools” and leads heavily with the company’s link-building history. (Citation Labs)

Either rebuild it or migrate its equity into a canonical /ai-visibility-consulting/ type page.

That page should answer, plainly:

• What problem do you diagnose?
• What platforms do you investigate?
• What evidence do you inspect?
• What can the intervention include?
• How do you measure before/after?
• What have you demonstrably changed?
• Who is this for?

Most of this copy already exists elsewhere. This is consolidation, not new intellectual work.

D. Fix the About/company entity description

This is probably the highest-priority individual copy change.

It should no longer begin by defining Citation Labs as a link-building agency.

Something structurally like:

Citation Labs is an AI visibility and search consultancy that helps enterprise SEO and growth teams understand why brands appear, disappear, get cited, or get recommended across AI answers and search. We diagnose the source and information gaps behind those outcomes, then use content, outreach, link building and other interventions to change them.

Not final copy, but that hierarchy.

Link building becomes an execution capability rather than the corporate category.

E. Reorder the proof architecture

The AI service page/homepage should immediately surface:

1. ZenBusiness AI case
2. 4,579-prompt measurement research
3. AI Mode behavioral research
4. Xofu
5. Citation Optimization methodology

The main case-study page should also make AI Visibility a visible category and put the AI case materially higher.

You don’t need to hide 20 years of link-building proof. You need to stop making models reconstruct the AI proposition from underneath it.

4. The external gap is larger

I checked a representative group of current “best GEO/AEO/AI visibility agency” pages that appeared in or closely match the source environment from your tests.

Citation Labs was absent from First Page Sage, Strategy Story, PikaSEO, GTM8020, AEO Vision, Built In, Similarweb’s agency content and TeamAI. (First Page Sage)

That’s a substantial gap.

And these aren’t irrelevant vanity links. Your Perplexity test repeatedly used exactly this class of page to construct its shortlist.

First Page Sage is particularly instructive. It has made GEO/AEO a top-level service and publishes its own “Top GEO Agencies” study. Its methodology even incorporates AI visibility and media references. (First Page Sage)

I would not try to persuade First Page Sage or Directive to add a competitor to their own commercial lists. That’s wasted effort.

The target is the independent/editorial/provider-selection layer.

First external target set

I’d build a list of roughly 15–20 actual source nodes, split into:

Priority 1: Existing category-selection pages

• independent AI/GEO agency comparisons
• marketing-industry roundups
• B2B agency selection articles
• credible directories/category resources

Your pitch isn’t “please add Citation Labs.”

It is:

You’re comparing AI visibility agencies. Here is a documented client intervention, engine-specific before/after data, a 4,579-prompt methodology study, and our approach to diagnosing source gaps. If that meets your inclusion criteria, here’s the provider information.

You have materially better proof than many of the firms currently appearing.

Priority 2: Existing authoritative descriptions of Citation Labs

Search Engine Land currently describes Garrett as founder of “a research and link-building agency”, despite categorizing his recent work under GEO/AI SEO and noting his current AI focus. (Search Engine Land)

That’s exactly the kind of entity description I would try to update where editorial controls permit it.

LinkedIn is already much closer to the desired positioning, describing Citation Labs around visibility across organic search and AI answers. (LinkedIn)

Audit conference bios, author bios, partner pages, podcast profiles and similar authoritative descriptions the same way.

Priority 3: Earn external citations to your research

Seer has an advantage that has nothing to do with publishing another “GEO guide.” Its AI research has been cited by Forbes, Inc., eMarketer, Fast Company, Search Engine Land and Ahrefs. (Seer Interactive)

Citation Labs has research that is capable of earning similar references. The goal should be to turn your studies into external evidence that Citation Labs is an AI-search research/consulting authority, not merely drive links back to the articles.

5. What I would explicitly not do

• Don’t delete or bury the link-building history.
• Don’t produce ten generic “What is GEO?” articles.
• Don’t start generic link building to an AI service page.
• Don’t chase every “best GEO agency” list. Many are self-serving commercial pages.
• Don’t obsess over schema before the human-readable category contradiction is fixed.
• Don’t try to make “Citation Optimization” replace GEO/AEO/AI visibility as the market category right now.

The Field Guide’s intervention rule applies almost perfectly: owned problems get owned-information changes; external-source problems get publisher/third-party interventions. You choose the surfaces the diagnosis identifies rather than distributing work everywhere.

6. Actual scope

I’d run this as one bounded cycle.

Owned: ~1 week

• homepage metadata/category language
• canonical AI Visibility Consulting page
• About page
• case-study architecture
• internal links/boilerplate/entity descriptions

Probably 5–8 meaningful page changes, not a site rewrite.

External: next 4–8 weeks

• audit 15–20 provider-selection/source nodes
• update controllable external bios/descriptions
• pitch inclusion where the source is legitimate
• distribute the strongest research/case proof to relevant editorial sources
• record every change

Before those changes ship, put the seven test prompts into Xofu as a fixed cohort and establish repeated baseline runs. Your first manual test is good enough to identify the problem, but it’s not a statistically stable visibility benchmark. Your own research makes the same distinction: repeated cohorts are materially more useful than single outputs. (Citation Labs)

Then re-test against the exact same cohort after the owned changes and as external sources begin moving.

So the immediate next action is the owned cleanup, starting with the homepage metadata, canonical AI service positioning and About page. The off-site campaign should start immediately after those are coherent, because every publisher you approach needs a clean source of truth to reference.

Yes. The source-environment diagnosis is now clear enough to act on.

Across your 19 unbranded AI-visibility prompts, the most recurrent source domains were:

There are two distinct gaps.

Owned gap: Citation Labs has strong AI work, but its company-level category signals are inconsistent.

External gap: Citation Labs is missing from the provider-selection/source ecosystem that the models repeatedly use to construct AI-visibility shortlists.

One useful platform difference reinforces that: ChatGPT and Gemini relied heavily on agencies’ own pages, while Perplexity drew much more from roundup/comparison pages. So both surfaces matter.

Source-gap map

The competitor-owned source pattern matters too

The firms showing up aren’t relying on subtle inference.

iPullRank now calls itself an enterprise AI Search agency, puts GEO services directly on the homepage, and sells a dedicated AI Search Strategy program. (iPullRank)

Go Fish opens its homepage with “Get cited in AI” and explicitly defines itself as a generative engine optimization agency. (Go Fish Digital)

Directive has a page literally titled “Generative Engine Optimization (GEO) Agency for B2B”, with methodology and an AI-search case result immediately underneath. (Directive)

Amsive’s service navigation says SEO + AEO, while its service page explicitly promises AI discovery, citations, share-of-voice analysis and source monitoring. (Amsive)

Omniscient has Generative Engine Optimization as a first-class service in its navigation, backed by a case result showing 81% LLM visibility growth and 140% citation growth. (Omniscient Digital)

These aren’t pages we can influence. They’re useful because they show the structural difference: the competitor’s entity and service category are explicit before you even get to the thought leadership.

Diagnosis I’d now put in the working record

Observed: Citation Labs appeared in 0/19 unbranded AI-visibility prompt tests, while several competitors and their owned service pages appeared repeatedly. When Citation Labs is queried by name, the systems recognize its AI-visibility work but continue to characterize the company primarily through link building.

Observed: Citation Labs’ owned environment contains contradictory category signals: current AI-first homepage messaging alongside legacy link-building page titles, company descriptions, service naming and proof architecture.

Observed: Citation Labs is absent from several third-party AI/GEO provider-selection pages that appeared in the test source environment.

Working diagnosis: Citation Labs has credible AI-visibility expertise and proof, but that evidence is not yet organized into a sufficiently consistent company/category signal on owned surfaces or represented strongly enough in external provider-selection sources to enter unbranded AI-visibility recommendation sets.

That is specific enough to choose an intervention.

Here is the measurement contract. One small cleanup is required before any site changes: three of the ChatGPT prompts in the first run were worded differently from the later Gemini/Perplexity cohort. The guide explicitly says exact prompt text and setup need to be preserved for a valid re-test.

So: 3 more ChatGPT runs, then freeze everything.

1. Complete baseline V1

Run these three exact prompts once in fresh ChatGPT Temporary Chats:

P3

Our company wants to get cited and recommended more often when prospective customers use ChatGPT, Gemini, or Perplexity to research vendors. Which agencies or consultancies specialize in helping with this?

P4

Competitors are appearing more often than our company in AI-generated answers used by prospective buyers. Which SEO, AEO, GEO, or AI-search consultancies are best equipped to diagnose why?

P5

Which agencies specialize in identifying the third-party sources that influence how brands are cited and recommended in AI search, then improving a company’s presence in that source environment?

Same conditions as before: fresh Temporary Chat, no added instructions, no forced search, no follow-up.

Right now we have a clean 0/12 Citation Labs appearances across the exact matched unbranded prompts that were run across the three environments. These three runs complete the 5-prompt × 3-platform matrix and give us a proper denominator of 15.

2. Freeze this seven-prompt cohort

Call it CL AI Category Baseline V1, 9 Sep 2026.

Do not edit these prompts during this intervention cycle. If we later decide they’re wrong, we create V2 rather than quietly changing V1. The guide specifically requires exact prompt text, platform/model, sampling approach and material changes to be preserved.

3. Decision this experiment needs to support

Do the owned-site and external-source interventions materially improve Citation Labs’ representation as an AI visibility/GEO/AEO consultancy in the measured AI answer environments, enough to justify continuing or scaling the work?

We are not measuring whether this generates leads yet. That’s downstream.

4. Primary measure

Unbranded AI-category inclusion rate

For P1-P5:

Number of responses that include Citation Labs as a relevant provider ÷ 15 total responses.

Three platforms × five prompts = 15.

Current clean matched baseline: 0/12. After your three final ChatGPT runs, we lock the actual V1 baseline.

This is the main outcome because the diagnosed problem is that Citation Labs does not enter the provider consideration set.

5. Close-in measures

Because the guide says to measure as close as possible to the intervention rather than relying on one broad visibility score, we also need separate measures for the two workstreams.

Owned-site intervention

Primary close-in measure: P6 brand framing.

Code each platform as:

• AI-primary: AI visibility/GEO/AEO is presented as the principal current specialization.
• Co-primary: AI visibility and link building are presented as equal current specializations.
• AI-secondary: AI visibility is described as an extension/newer capability beneath link building.
• AI-absent: AI capability is not meaningfully represented.

Current baseline from your test:

• ChatGPT: AI-secondary
• Perplexity: Co-primary
• Gemini: Co-primary

We also record which Citation Labs pages each system uses to support that answer.

External-source intervention

Primary close-in measure: source-environment movement.

For each third-party target we pursue, record:

• Citation Labs absent/present
• described as what?
• AI visibility/GEO/AEO explicitly associated?
• date changed/published
• source URL
• whether that source later appears in P1-P5 outputs

This tells us whether the external evidence actually changed before expecting an AI recommendation to move.

6. Secondary measures

For every P1-P5 response, record:

• Citation Labs present: yes/no
• recommended/shortlisted: yes/no
• position if ordered
• category/framing
• Citation Labs-owned source used: yes/no
• independent Citation Labs source used: yes/no
• recurring competitors
• source domains
• obvious platform differences

P7 stays primarily a control/context prompt. We aren’t trying to destroy Citation Labs’ association with link building. We want to establish AI visibility strongly enough that one category doesn’t obscure the other.

Comparisons like competitor movement, untreated categories and platform/model differences are specifically recommended because they help distinguish intervention movement from broader environmental change.

7. Expected result patterns, defined now

This is important because we don’t get to decide afterward that whatever happened was success. The guide requires these states to be written before deployment.

Supportive after owned changes

• P6 shifts toward AI-primary/co-primary, especially if ChatGPT stops describing AI visibility as merely a recent expansion.
• Citation Labs-owned AI pages begin contributing more clearly to the answer.
• Ideally, Citation Labs starts appearing on at least some P1-P5 prompts.

Null after owned changes

• P6 framing remains materially unchanged.
• Citation Labs remains absent from P1-P5.
• Source sets remain essentially unchanged.

That would weaken the idea that owned category coherence is enough.

Complicating but useful

• P6 improves materially, but P1-P5 remain at zero.

That would strongly suggest the owned cleanup fixed brand understanding, while the external/category-authority gap remains the binding constraint.

Supportive after external work

• Relevant third-party sources begin explicitly associating Citation Labs with AI visibility/GEO/AEO.
• Those sources begin appearing in the measured answer environment.
• Citation Labs begins entering P1-P5 provider sets.

Adverse

• AI framing becomes less accurate.
• Models start attributing capabilities/results Citation Labs cannot substantiate.
• Changes make the company harder rather than easier to understand.

Inconclusive

• Major model/platform changes happen during the window.
• Prompt/setup changes.
• Search behavior changes materially.
• Results are too unstable to distinguish intervention movement from system noise.

8. Observation windows

Don’t re-test daily.

T0: complete the three missing ChatGPT runs and freeze baseline.

Owned intervention ships: log the exact date each material page change goes live.

T+7 days: optional diagnostic check. Do not treat it as the formal result.

T+21 days: formal owned-intervention re-test. Run all seven prompts once in all three environments under the original conditions.

External work: record each third-party publication/update individually.

Formal external re-test: approximately 6 weeks after external work begins, or once at least three material external source changes are live, whichever is later.

AI retrieval/indexing lag is not sufficiently predictable to justify inventing a precise expected magnitude, so the contract should leave that unknown rather than fabricate one. The guide explicitly allows expected magnitude only where there is a real basis.

9. Change log

From T0 onward, record:

• homepage/title/meta changes
• About page changes
• AI service page changes/redirects
• case-study architecture changes
• internal linking changes
• new AI research/content
• external mentions/inclusions
• PR/editorial activity
• relevant competitor changes you notice
• platform/model updates
• measurement/setup changes

The guide specifically calls out site releases, PR, third-party updates, model/platform changes and competitor activity because any of them can complicate interpretation.

Baseline V1 is now locked. The three added ChatGPT prompts all returned no Citation Labs inclusion, so the matched cross-platform baseline is complete.

CL AI Category Baseline V1

Date: 9 Sep 2026
Platforms: ChatGPT Temporary Chat, Gemini, Perplexity
Cohort: P1-P7 exactly as saved in the workbook

The new P3-P5 ChatGPT answers also correctly recognized the problem as GEO/AEO/AI-search work. So we no longer have the earlier concern that overly broad prompt interpretation is contaminating this matched cohort.

That gives us a cleaner diagnosis:

Citation Labs is understood as having an AI-visibility capability when the brand is named, but it does not currently enter the consideration set when buyers ask these three AI environments for providers in the category.

And the link-building control is useful: Citation Labs isn’t simply being substituted into every link-building shortlist either. The problem we’re addressing is primarily failure to enter the AI-visibility provider set, alongside inconsistent branded category framing.

Measurement contract is now frozen

Primary measure after intervention:

P1-P5 Citation Labs inclusion rate: 0/15 → ?

Close-in owned-site measure:

P6 framing

• ChatGPT: AI-secondary
• Perplexity: co-primary
• Gemini: co-primary

Control:

P7 link-building inclusion: 1/3. We don’t need or want this to disappear.

Do not change the seven prompts or testing conditions during this intervention cycle. The guide requires the exact prompt set and measurement conditions to remain identifiable so the re-test is comparable.

We’re done with measurement setup

No more testing right now.

Next is implementation of the P0 owned-footprint changes. I would do those before external outreach because we want third parties and AI systems landing on a coherent company-level source of truth.

Order:

1. Homepage title/meta/category signals
2. AI service naming and canonical positioning
3. About page company definition
4. AI service page hierarchy/copy
5. Case-study/proof architecture
6. Relevant internal links/boilerplate

Then log the dates, allow an observation window, and re-run this exact V1 cohort. The measurement contract exists specifically so we don’t move the goalposts after seeing the results.

Next action: audit and specify the exact homepage changes.

Directionally correct. I would not rewrite the homepage. The current hero is already AI-first and much better than the old category signal suggests. The homepage problem is consistency, not strategy.

Homepage changes I’d make now

1. Change the title tag. Required.

Current indexed title:

Enterprise Link Building Agency & Tools

That directly conflicts with the current homepage H1, which is already AI-first: “Shape how your brand appears in the AI answer space.” (Citation Labs)

Use:

AI Visibility Consulting & GEO Services | Citation Labs

Meta description:

Citation Labs helps enterprise SEO and growth teams improve brand visibility, citations, and recommendations across ChatGPT, Gemini, Perplexity, and AI search.

This is the single clearest homepage fix.

2. Standardize the service name. Required.

The navigation still says LLM Mentions, while the homepage service card says AI Advisory. (Citation Labs)

That is unnecessary category fragmentation.

Change:

• Nav: LLM Mentions → AI Visibility
• Homepage card: AI Advisory → AI Visibility Consulting

Keep the current card description broadly intact:

Identify your AI visibility gaps, prioritize opportunities, and track the business impact of campaigns.

I would not use “AI Advisory” as the category anchor. It is too broad.

3. Make the hero explicitly name the category once. Required.

Keep the H1:

SHAPE HOW YOUR BRAND APPEARS IN THE AI ANSWER SPACE

It is differentiated and clear.

Change the subhead from:

We help enterprise SEO and growth teams increase brand recommendations when buyers compare options across search and AI.

to:

We help enterprise SEO and growth teams diagnose and improve AI visibility, increasing brand recommendations when buyers compare options across search and AI.

That gives the page an explicit AI visibility category statement without turning the hero into generic GEO copy. The rest of the homepage already supports that positioning. (Citation Labs)

4. Rebalance the proof strip directly under the hero. Required.

Right now, the first three proof signals after an AI-first hero are:

• 20+ years link-building + SEO experience
• 195,468 links built
• 35% average revenue increase

That immediately pulls the entity back toward link building. (Citation Labs)

I would not remove the heritage. I would replace one of those three with explicit AI proof.

Recommended:

• 20+ years search + link-building experience
• 3 weeks to move a key enterprise product differentiator into AI answers
• 6.3% → 33.3% Google AI Mode citation presence in the ZenBusiness case

Those AI numbers are documented in the current case study. (Citation Labs)

I would move “195,468 links built” into the Link Building service section rather than making it a company-level proof point immediately below the AI hero.

5. Move the AI case study to the first proof position. Required.

The homepage already contains the ZenBusiness AI result, but it sits inside a mixed proof section with link-building cases. (Citation Labs)

Make ZenBusiness the first and visually dominant case under “Our Success Stories.”

Then keep the link-building cases beneath it.

The hierarchy should say:

This is what we do now.
This is also what we have historically been very good at.

Not the reverse.

Two smaller fixes

6. Change the footer/company descriptor.

The homepage footer currently says:

Citation Labs helps enterprise teams improve visibility across search and AI.

That is good. Keep it. (Citation Labs)

But this exact or near-exact descriptor should become the standard company boilerplate everywhere, because other pages still end with “Citation Labs builds 100s of relevant, trust-worthy links at scale for enterprise every month.” That contradiction shows up even on the AI service page and ZenBusiness case study. (Citation Labs)

7. Fix the CTA typo.

Current:

See where your band shows up…

Should be:

See where your brand shows up…

(Citation Labs)

What I would leave alone

Do not change:

• the current H1
• “Expertise built for the new era of search”
• “Make AI visibility legible to leadership”
• the FUQ section
• the book/link-building heritage section
• the link-building service itself

Those are not the problem.

The homepage already contains enough AI positioning to support the category. The material issue is that the title, service labels, proof hierarchy, and legacy signals still disagree about what Citation Labs primarily is.

Homepage implementation list

If you want the smallest viable intervention, make these five:

1. Replace title/meta.
2. Rename LLM Mentions/AI Advisory to AI Visibility/AI Visibility Consulting.
3. Add “diagnose and improve AI visibility” to the hero subhead.
4. Replace one or two hero proof stats with ZenBusiness AI proof.
5. Make ZenBusiness the first case study.

Then log the publish date.

After that, move to the AI service page, which currently has a much larger positioning problem than the homepage.

Not directionally correct. This page needs a substantive rewrite, not just relabeling.

The current /llm-mentions/ page still sells an older version of the offer: “increase conversions from ChatGPT,” “LLM rank reporting,” finding where LLMs “train,” and targeting brand mentions/content/links. It also opens with link-building proof and closes by defining Citation Labs as a company that builds hundreds of links every month. (Citation Labs)

That conflicts with the stronger methodology you now have.

1. Reposition the page around diagnosis, intervention and measurement

Current H1

AI TOOLS STEALING YOUR TRAFFIC AND CUSTOMERS?

Replace it.

Recommended direction:

AI Visibility Consulting for Search and Answer Engines

Subhead:

Find where your brand disappears from AI answers, diagnose why, and make targeted changes to the information and sources shaping those results.

This explicitly establishes the category and mirrors the actual method.

The Field Guide says the work is: establish a prompt set, baseline the answer environment, diagnose the gap, choose the intervention, define measurement, then re-test.

That should become the service page spine.

2. Replace the current capability blocks

The current blocks are things like:

• Use our learning
• Get 1st party data
• Strategy design
• LLM rank reporting
• BOFU conversion forensics
• Track rank in LLMs

That makes the service look like a collection of tactics/tools. (Citation Labs)

I would collapse them into four commercial stages:

Diagnose

Establish the prompts and buying situations that matter. Measure whether you are retrieved, cited, mentioned, recommended and represented accurately.

Find the gap

Inspect the sources, competitors, evidence and answer patterns behind weak visibility.

Intervene

Change the surface the diagnosis points to: owned content, third-party evidence, comparison assets, publisher sources, information architecture or another relevant surface.

Measure what changed

Re-run the same prompt cohort and show what moved, what did not and what the evidence supports.

That is much stronger because the service becomes a repeatable consulting method, not “we know some GEO tactics.”

3. Kill or rewrite the claims that are now weaker than your own research

These sections should not survive as written:

“Learn where LLMs train…”

“LLM responses come from training data, citations, and linked content.”

“LLM optimization depends on when AI models update their data sets.”

Those are too broad and increasingly inaccurate descriptions of current answer environments. (Citation Labs)

Your current research is substantially more sophisticated: source sets, retrieval, citations, recommendations, platform differences and repeated prompt cohorts. Your 4,579-prompt study explicitly shows that different surfaces behave differently and that repeated measurement is required. (Citation Labs)

Use that language instead.

4. Replace link-building proof with AI-specific proof at the top

Current top proof:

• 20+ years link-building + SEO
• 195,468 links built
• 35% average revenue increase

Wrong proof for this page. (Citation Labs)

Use something like:

• 4,579 prompts tracked across AI/search environments
• 6.3% → 33.3% citation presence in Google AI Mode
• 3 weeks to close a priority AI visibility gap

Those are directly supported by Citation Labs’ current research/case work. (Citation Labs)

You can retain “20+ years in search/link building” lower down as credibility/history.

5. Make ZenBusiness the page’s central proof

The page already has the testimonial and case study, but it is surrounded by unrelated link-building cases. (Citation Labs)

For this service page:

Primary case: ZenBusiness.

Then use the research as methodological proof:

• 4,579-prompt measurement study
• off-domain comparison experiment
• third-party influence research
• AI Mode buyer behavior study

Those collectively establish that Citation Labs does more than execute content. It studies the mechanics, diagnoses the environment and tests interventions. (Citation Labs)

Move generic link-building cases off this page.

6. Standardize the commercial name

Use:

AI Visibility Consulting

Then use these terms naturally underneath it:

AI visibility / GEO / AEO / AI search optimization

I would treat Citation Optimization as the proprietary methodology.

So structurally:

AI Visibility Consulting
Powered by Citation Optimization

Not:

LLM Mentions

“LLM Mentions” undersells the service. Your own Field Guide explicitly separates being mentioned from being recommended, cited, retrieved and accurately represented.

URL

I would not change /llm-mentions/ yet solely for the sake of having an exact-match slug.

Rewrite the page, title, H1, navigation and internal anchor text first. A URL migration adds complexity and isn’t necessary to test our diagnosis.

If you eventually consolidate around /ai-visibility-consulting/, do it deliberately with a 301 and internal-link migration. It doesn’t need to be part of this first intervention.

7. Fix the company-level signals on this page

Required:

• Footer: remove “Citation Labs builds 100s of…links at scale…” (Citation Labs)
• Services nav: LLM Mentions → AI Visibility
• Newsletter: change “latest link-building experiments” to search/AI visibility research
• CTA typo: “band” → “brand”
• Page title: replace LLM Mentions: Increase Visibility in AI Search Tools

Recommended title:

AI Visibility Consulting & GEO Services | Citation Labs

What I would leave

Keep:

• ZenBusiness testimonial
• Xofu as the measurement platform
• enterprise SEO/growth-team audience
• business-impact/reporting orientation
• the idea that interventions can include off-domain sources/content/links

The intervention itself can still be link building. The page should establish that Citation Labs diagnoses the visibility problem first and uses the intervention the evidence calls for.

Required rewrite scope

This is not a 5-line edit. I would rewrite:

1. title/meta
2. hero
3. proof strip
4. service methodology section
5. case/proof section
6. FUQs
7. footer/service labels

I would not touch the URL in this cycle.

Next owned asset after this is the About page, because that is currently the clearest company-level source still telling AI systems “Citation Labs = link-building agency.”

The About page needs a company-definition rewrite, but not a full rebuild.

The biggest issue is that it explicitly tells both humans and machines that Citation Labs is a link-building agency.

1. Rewrite the opening company definition. Required.

Current:

We’ve built a family-first (and pet-first) link-building agency…

And immediately below:

We’re a remote team of link-building strategists, developers, and executors…

Those are very strong entity signals. (Citation Labs)

Replace the first with something closer to:

Citation Labs is an AI visibility and search consultancy helping enterprise teams understand and improve how their brands are found, cited, compared, and recommended across search and AI.

Then preserve the culture point:

We’re a remote, family-first team of strategists, researchers, developers, editors, and outreach specialists who build and test new ways to improve visibility in competitive information environments.

That alone would materially change what the page says Citation Labs is.

2. Keep the history. Don’t sanitize the link-building origin.

The timeline is useful because it demonstrates how Citation Labs got here:

• SEO/community
• enterprise link building
• co-citation
• purchase committees
• sales-page link building
• AI search
• Citation Optimization
• Xofu

The 2025–Present section already makes that evolution fairly well. It says AI changed how buyers find and shortlist companies and defines Citation Optimization as improving the sources, context, and proof AI systems use. (Citation Labs)

Keep that.

I would just add a short transition above the timeline:

Citation Labs started in link building. Over more than a decade, that work expanded into a broader question: how do brands earn visibility and credibility wherever buyers and retrieval systems look for evidence? Today, that includes organic search, AI answer environments, third-party sources, and the information buyers use to compare options.

That makes the history evidence for the current positioning instead of allowing the history to define the current category.

3. Rewrite the mission. Required.

Current mission:

To help in-house SEOs justify their role by getting them 100s of relevant, high-quality links to sales pages every month…

This directly contradicts the category you are trying to establish. (Citation Labs)

I would replace it with:

To help search and growth teams understand and improve how their brands are represented across search and AI, using evidence, interventions, and reporting they can defend internally.

If you want to retain more of the customer-language flavor:

To help in-house search teams earn the visibility their expertise deserves across search and AI, and show leadership what changed and why it matters.

The first is cleaner.

4. Keep the vision, with a minor update.

Current:

Enable buyer advocacy by sharing new data and expertise through unique approaches to citations that make the search experience (across all tools and platforms) more useful, accurate, and relevant for customers.

This is broadly compatible with the new positioning. (Citation Labs)

I’d tighten it to:

Enable better buyer decisions by improving the evidence, sources, and information that search and AI systems use to represent brands.

Not mandatory for the first intervention, but cleaner.

5. Rewrite “Where We’re Heading.” Required.

Current:

We’re an absolute madhouse of perfection, finding new ways to bring visibility and prospective buyers to key sales pages…

Then:

…drive links, show up in search results, get mentioned in LLMs…

This still sounds like a tactical SEO lab experimenting with several channels. (Citation Labs)

Replace with something closer to:

We’re building better ways to diagnose, influence, and measure brand visibility across search and AI.

Through The Lab, we test how sources, citations, content, links, buyer evidence, and information architecture affect what search and AI systems retrieve, represent, and recommend.

That connects directly to the service methodology.

6. Update the team intro. Small but worthwhile.

Current:

We’ve built a tight-knit team from all walks of life…

Fine culturally, but it says nothing about capability.

Add one sentence:

The team spans search strategy, research, content, outreach, software, and measurement, giving us the ability to investigate visibility problems and build the interventions they require.

That helps reinforce why Citation Labs can legitimately sell consulting rather than simply link execution.

7. Do not rewrite every team bio now.

Several bios remain heavily link-building oriented. Kristina, Lindsay, Rachel and others explicitly talk about links/link-building. (Citation Labs)

I would leave those alone in this cycle.

They describe people’s actual history/work. Trying to purge “link building” everywhere would be artificial and unnecessary.

The company-level signals matter far more:

• opening
• Our Story introduction
• mission
• vision
• current-era timeline
• Where We’re Heading

Fix those first.

8. Navigation terminology must change here too.

The About page still exposes:

Services → Link Building at Scale / LLM Mentions / Citation-Worthy Content

(Citation Labs)

As already specified:

LLM Mentions → AI Visibility

This needs to be global, not page-by-page.

Minimum viable About-page intervention

Make these four changes:

1. Replace “link-building agency” hero definition.
2. Replace “team of link-building strategists…” company description.
3. Rewrite the mission.
4. Rewrite “Where We’re Heading.”

Keep the timeline and culture almost entirely intact.

That turns the page from:

We are a link-building agency that evolved into AI work

into:

We are a search/AI visibility consultancy whose expertise was built through years of link-building, source, citation, and buyer-information work.

That is the category relationship we want the public footprint to communicate.

Next: case-study/proof architecture.

The case-study architecture needs a reorder and recategorization, not a rewrite of every case.

Right now the Results page tells a very clear story: the first three visible cases are link-building cases, ZenBusiness AI is fourth, and the page closes by defining Citation Labs as a company that “builds 100s of relevant, trust-worthy links at scale for enterprise every month.” (Citation Labs)

That is materially working against the category change.

1. Make ZenBusiness the featured case. Required.

It should be the first case a visitor/model encounters.

The proof is strong enough:

• AI visibility gap identified
• visible change in under three weeks
• Google AI Mode citation presence 6.3% → 33.3%
• Google AI Mode rank-one visibility 74.2% → 88.9%
• Gemini rank-one visibility 45.8% → 62.2% (Citation Labs)

I would give it a larger featured card above the rest of the grid.

2. Add explicit case-study categories. Required.

Use three categories:

Don’t pretend you have five AI cases when you have one strong one.

That is fine. One documented case is better than relabeling unrelated work.

Default view can remain All, but ZenBusiness should be featured first and the category label should be visible on every card.

3. Change the page-level category language. Required.

Current H1:

CASE STUDIES

Fine, but too neutral by itself.

Add a subhead:

See how Citation Labs helps enterprise teams improve visibility across AI, search, content, and the sources buyers rely on.

SEO title:

AI Visibility, SEO & Link Building Case Studies | Citation Labs

Meta:

See how Citation Labs helps enterprise teams improve AI visibility, search performance, citations, and authority through research-led interventions.

This lets the Results page itself become another strong company-category source.

4. Do not rewrite the link-building cases to make them sound like GEO.

Leave cases such as the 6,496-link program and the $1M/month SEO case as link-building/search proof. They demonstrate real capability and history. (Citation Labs)

Trying to retrofit “AI visibility” into them would weaken the evidence.

The architecture should communicate:

Citation Labs has deep search/link-building capability AND now has demonstrated AI-visibility capability.

5. Separate research proof from client cases.

The 4,579-prompt research is extremely useful, but don’t turn it into a fake case study. It establishes measurement expertise: fixed cohorts, repeated measurement, platform differences and recommendation tracking. (Citation Labs)

I would add a small section after the case-study grid:

Research Behind the Work

Cards for:

• 4,579-prompt AI visibility study
• Measuring Third-Party Influence on AI Answers
• Decision Architecture

That gives an AI-consultancy evaluator both client proof and methodological proof without conflating them. (Citation Labs)

6. Fix the ZenBusiness case’s service path.

The case should clearly link back to the canonical AI Visibility Consulting service page.

Likewise:

• AI Visibility service → ZenBusiness case
• ZenBusiness case → AI Visibility service
• 4,579-prompt study → AI Visibility service
• third-party influence research → AI Visibility service
• homepage AI service card → AI Visibility service

This is where the owned footprint starts behaving like one coherent evidence network rather than isolated AI articles.

One caveat: the ZenBusiness case currently describes Citation Labs as the “execution partner for content strategy and production.” (Citation Labs)

Do not casually upgrade that to “Citation Labs diagnosed and led the entire AI strategy” unless that is factually what happened. If Citation Labs had a broader role, update it accurately. If not, leave the role bounded.

7. Replace the Results-page boilerplate. Required.

Remove:

Citation Labs builds 100s of relevant, trust-worthy links at scale for enterprise every month.

(Citation Labs)

Use the same standardized company description we establish globally. For example:

Citation Labs helps enterprise teams understand and improve how their brands appear across search and AI.

Also change the newsletter line from:

Join our newsletter to see our latest link-building experiments…

to something like:

Join our newsletter for research and experiments on AI visibility, search, citations, and link building.

The service navigation should also carry the global AI Visibility label rather than LLM Mentions. (Citation Labs)

Minimum intervention

I would ship:

1. ZenBusiness featured first.
2. AI Visibility / Link Building / Content & Research category labels.
3. New page subhead/title/meta.
4. “Research Behind the Work” section.
5. Cross-links between AI service, ZenBusiness and core research.
6. Replace legacy link-building-only boilerplate.

No need to rewrite the existing case studies themselves.

Next is the final owned-footprint layer: internal links, global boilerplate, navigation, and terminology consistency across the site.

This is the final owned-footprint cleanup layer. It is mostly consistency work, but it matters because the site currently repeats the legacy link-building identity across global elements and older pages.

1. Change the global navigation label. Required.

Current:

LLM Mentions

Change globally to:

AI Visibility

Keep:

Link Building at Scale
Citation-Worthy Content

Do not use different service names on different pages.

Recommended commercial hierarchy:

• AI Visibility = navigation/service category
• AI Visibility Consulting = service-page H1/category
• GEO/AEO/AI search optimization = supporting market terminology in copy
• Citation Optimization = Citation Labs’ methodology/framework
• LLM Mentions = retire as a service name

The current homepage still exposes “LLM Mentions,” which is materially narrower than the service you are selling. (Citation Labs)

2. Replace the global footer boilerplate. Required.

This sentence appears across About, FUQ, case studies, process pages, contact and other indexed pages:

Citation Labs builds 100s of relevant, trust-worthy links at scale for enterprise every month.

(Citation Labs)

That is a strong repeated company-definition signal.

Replace it sitewide with:

Citation Labs helps enterprise teams understand and improve how their brands appear across search and AI.

If you want one slightly more descriptive version:

Citation Labs helps enterprise SEO and growth teams improve how their brands are found, cited, compared, and recommended across search and AI.

I prefer the second if space permits.

3. Standardize the newsletter boilerplate. Required.

Where pages still say things like:

latest link-building experiments

change to:

research and experiments on AI visibility, search, citations, and link building

Do not remove link building. Broaden the descriptor so every global component stops reasserting the old category.

4. Build the canonical internal-link network. Required.

The important AI pages should explicitly reinforce one another.

AI Visibility service page should link to:

• ZenBusiness AI case
• 4,579-prompt measurement research
• Citation Optimization framework
• third-party/source influence research
• Xofu where relevant

ZenBusiness case should link to:

• AI Visibility Consulting
• relevant methodology/research

Core AI research should link to:

• AI Visibility Consulting where there is a legitimate commercial next step
• Citation Optimization framework where methodology is discussed
• ZenBusiness where client proof is relevant

Homepage, About and Results should all link to the canonical AI Visibility service page.

The objective is not “more internal links.” It is to make the relationship between:

company → service → methodology → research → client proof

explicit.

5. Use consistent anchor language.

Prefer descriptive anchors such as:

• AI Visibility Consulting
• AI visibility
• AI visibility research
• Citation Optimization
• AI visibility case study

Avoid making every link:

• Learn More
• Read More
• LLM Mentions

You do not need exact-match anchors everywhere. You need the semantic relationship to be obvious.

6. Do not rewrite link-building pages into AI pages.

The current Link Building at Scale page should remain a strong link-building service page. It is allowed to say Citation Labs has 20+ years of link-building experience, has built 195,468 links, etc. (Citation Labs)

What I would remove from that page is the global footer-level definition that implies link building is the whole company’s identity.

In other words:

Page-specific signal:

This is our link-building service.

Good.

Company-wide signal:

Citation Labs is a company that builds hundreds of links every month.

Needs to change.

7. Audit older indexed pages for global company descriptions.

Do not rewrite the actual historical article/case content.

But update shared components on pages such as:

• FUQs
• process pages
• case studies
• category/archive pages
• contact
• About
• blog posts

The old footer appears broadly across those surfaces today. (Citation Labs)

That makes this potentially a one-template fix with a large footprint, which is high leverage.

8. Contact page needs one small category change.

The current Contact page is already relatively well aligned. It says Citation Labs can fit teams investing in:

strategic link building, citation optimization, AI answer visibility, or local and community source coverage

(Citation Labs)

I would change that to:

AI visibility consulting, strategic link building, citation optimization, or local and community source coverage

This puts the explicit commercial category first.

Otherwise, leave the page alone for now.

9. Structured data: check after the copy changes.

I would not make schema the first intervention.

Once the human-readable pages are consistent, inspect:

• Organization schema
• Service schema
• sameAs/entity references
• service names/descriptions

Then make sure they reflect the same category hierarchy.

Do not use schema to say something the visible site does not.

Owned intervention is now specified

The complete P0 owned workstream is:

1. Homepage metadata/category cleanup
2. AI service page rewrite
3. About company-definition rewrite
4. Results/case-study architecture
5. Global nav terminology
6. Global footer/company boilerplate
7. Newsletter boilerplate
8. Internal link network
9. Minor Contact-page category alignment
10. Structured-data check after the visible changes

I would implement those as one owned-site intervention wave, log the publish date, then begin the external-source work.

The next step is the external target list and outreach strategy.

Do this in two waves. First fix the external sources you can control or credibly influence. Then pitch provider-selection pages.

Wave 1: highest leverage

I would not prioritize MentionBench, First Page Sage, VOCTOS, Position Digital, Stratabeat, Schmitdy or other agencies ranking themselves. They may still matter to models, but you are asking a direct competitor to add you to a commercial comparison page. Low expected return. MentionBench, for example, now sells essentially the same AEO/GEO consulting category directly. (MentionBench)

The evidence pack

Use one consistent pack for every external target. Do not custom-build a deck.

Company category

Citation Labs is an AI visibility and search consultancy helping enterprise teams diagnose and improve how their brands are cited, represented, and recommended across AI answer environments.

Proof 1: client movement

ZenBusiness: Google AI Mode citation presence moved from 6.3% to 33.3% after the intervention, with additional rank-one visibility movement in AI Mode and Gemini.

Proof 2: measurement

Citation Labs published research based on 4,579 buyer-prompt runs, using fixed cohorts and repeated measurement rather than single-response screenshots. (Citation Labs)

Proof 3: methodology

Citation Labs’ work covers prompt baselining, source/citation analysis, recommendation behavior, off-domain interventions and re-testing.

Supporting assets

• canonical AI Visibility Consulting page
• ZenBusiness case
• 4,579-prompt research
• Citation Optimization methodology
• Xofu if relevant

Do not send six links by default. Lead with the two strongest and offer the rest.

Outreach angle

The ask should be evidence-based inclusion, not backlink outreach.

For roundup/directory targets:

I came across your 2026 comparison of AI visibility/GEO providers. Citation Labs isn’t currently included, and I think we may now meet the bar you’re using.

We have a documented client case showing engine-specific before/after AI visibility movement, plus published methodology based on 4,579 buyer-prompt runs. We also diagnose third-party source/citation gaps rather than treating GEO as an on-page content program.

If you’re reviewing the list again, happy to send the evidence against your inclusion criteria.

Then customize one sentence against their stated methodology.

For AEO Vision, explicitly map to:

• active delivery
• client improvement
• public results
• published methodology

For PikaSEO, map to:

• technical depth
• documented AI-citation result
• specialization

For Built In, I would be more restrained because its editorial policy says editorial coverage is independent and it does not accept third-party line edits. Treat it as an editorial tip, not “please add us.” (Built In)

Something like:

You included providers helping brands improve AI-search visibility through content, third-party coverage and source analysis. Citation Labs has published a recent enterprise case and measurement research that may be relevant if you revisit the piece. Happy to send the underlying evidence.

Search Engine Land is different

Don’t pitch an article inclusion. Ask them to update Garrett’s contributor bio because the current company description is stale relative to Citation Labs’ current work.

Suggested replacement:

Garrett French is the founder of Citation Labs, an AI visibility and search consultancy that helps brands understand and improve how they are cited, represented, and recommended across AI answers and search. He also leads Xofu, a platform for measuring AI visibility, and ZipSprout.

Their separate Citation Labs author page is already much closer to the desired positioning, describing the company as helping brands improve visibility in search results and AI answers through citation optimization, content and link building. (Search Engine Land)

So this is a clean consistency request.

Execution order

Don’t outreach yet if the owned pages still say “LLM Mentions” and define the business as link building. You would be asking editors to adopt positioning your own site contradicts.

Once the owned wave is live:

Week 1

• Search Engine Land bio
• Clutch
• Built In
• AEO Vision
• TeamAI
• PikaSEO

Week 2

• Follow up once where appropriate
• Ranktracker
• identify 3–5 additional credible provider-selection sources that have appeared in Xofu/test results since baseline

Then stop. Don’t build a 100-site outreach campaign.

The external intervention should initially be 6–10 high-relevance category sources, not generic digital PR. We then log which ones change and see whether those source nodes start appearing in the fixed prompt cohort. That keeps the work tied directly to the diagnosis rather than turning into a link-building campaign.

Use this as the implementation spec. I’d ship the owned changes as one release so the intervention date is clean.

The copy below keeps the current positioning that is already working, removes the strongest link-building-only entity signals, and makes the service architecture reflect the method in the AI Visibility guide: baseline, diagnose, intervene, re-test, interpret.

1. Global/template changes

Navigation

Replace:

LLM Mentions

With:

AI Visibility

Services should read:

• Link Building at Scale
• AI Visibility
• Citation-Worthy Content

The live site currently uses “LLM Mentions” in global navigation while the homepage calls the service “AI Advisory.” (Citation Labs)

Global company descriptor

Use this everywhere a short Citation Labs definition appears:

Citation Labs helps enterprise SEO and growth teams improve how their brands are found, cited, compared, and recommended across search and AI.

Replace the legacy:

Citation Labs builds 100s of relevant, trust-worthy links at scale for enterprise every month.

That old line still appears on the Results page and even the ZenBusiness AI case. (Citation Labs)

Newsletter

Use:

Get our latest research, experiments, and insights on AI visibility, search, citations, and link building.

The homepage footer is already close to this, but other pages still use link-building-only newsletter language. (Citation Labs)

2. Homepage

The current H1 is good. Keep it. The homepage already leads with AI but still carries link-building metadata, mixed service naming, and a link-heavy proof strip. (Citation Labs)

SEO title

AI Visibility Consulting & GEO Services | Citation Labs

Meta description

Citation Labs helps enterprise SEO and growth teams diagnose and improve brand visibility, citations, and recommendations across ChatGPT, Gemini, Perplexity, and AI search.

Hero

Keep H1:

SHAPE HOW YOUR BRAND APPEARS IN THE AI ANSWER SPACE

Replace subhead with:

We help enterprise SEO and growth teams diagnose and improve AI visibility, increasing brand recommendations when buyers compare options across search and AI.

Keep the current reporting line if you want the personality.

Proof strip

Replace the current three:

+20 years link-building + SEO experience
195,468 links built
35% average revenue increase

With:

20+ years in search and off-site visibility
3 weeks to close a priority AI visibility gap
6.3% → 33.3% Google AI Mode citation presence

The last two are directly documented in the ZenBusiness case. (Citation Labs)

Move 195,468 links built into the Link Building service card/page rather than using it as company-level hero proof.

Services

Change:

AI Advisory

To:

AI Visibility Consulting

Replace description with:

Diagnose where your brand disappears from AI answers, identify the sources and evidence shaping those results, and measure whether the intervention changes them.

Leave Link Building and Citation-Worthy Content as distinct services.

Success stories

Keep the current section, but make ZenBusiness the first featured case.

Suggested section intro:

See how enterprise teams have used Citation Labs to improve AI visibility, search performance, authority, and the information buyers rely on.

CTA

Keep:

Book an AI Strategy Call With Our Senior Team

Fix:

See where your brand shows up, where it disappears, and what to fix first.

The live copy currently says “band.” (Citation Labs)

3. AI Visibility service page

Keep /llm-mentions/ for this intervention. Change the visible category and rewrite the page.

The current page still leads with “AI tools stealing your traffic,” link-building proof, “LLM rank reporting,” training-data language, and tactical sections that predate the stronger diagnostic method. (Citation Labs)

SEO title

AI Visibility Consulting & GEO Services | Citation Labs

Meta description

Diagnose why your brand is missing from AI answers, identify the sources shaping citations and recommendations, intervene where the gap occurs, and measure what changed.

H1

AI VISIBILITY CONSULTING FOR SEARCH AND ANSWER ENGINES

Hero copy

Find where your brand disappears from AI answers, diagnose why, and make targeted changes to the information and sources shaping those results.

Citation Labs helps enterprise SEO and growth teams measure visibility across AI environments, identify citation and recommendation gaps, and choose the intervention the evidence supports.

CTA:

Book an AI Visibility Strategy Call

Proof strip

4,579 buyer prompts tracked across repeated runs
3 weeks to close a priority AI visibility gap
6.3% → 33.3% Google AI Mode citation presence

The 4,579-prompt study used repeated cohorts across multiple AI/search surfaces, while the ZenBusiness case documents the intervention result. (Citation Labs)

Section: Start with the answer environment

H2

DON’T GUESS WHAT “GEO” TACTIC YOU NEED

Copy:

Being absent from an AI answer can mean very different things.

Your evidence may not be discoverable. Another source may be cited instead. Your brand may be mentioned but not recommended. Important facts may be represented incorrectly. Or competitors may simply have stronger evidence for the criteria driving the answer.

We establish what is happening first, then decide what needs to change.

This directly reflects the guide’s distinction between retrieval, citation, mention, recommendation and framing rather than collapsing them into a single visibility score.

Section: The method

H2

FROM VISIBILITY CONCERN TO TESTABLE INTERVENTION

1. BASELINE

Define what is worth measuring

Build a tracked prompt set around the buyer questions, decisions, competitive situations, and business priorities that matter. Preserve the setup so later results are comparable.

2. DIAGNOSE

Find where the gap occurs

Inspect whether your brand and evidence are being found, cited, mentioned, recommended, and represented accurately. Where available, analyze source sets, query branches, competitors, and the evidence AI systems rely on.

3. INTERVENE

Change the surface the diagnosis points to

The answer may be owned content, third-party evidence, comparison coverage, publisher outreach, citation-worthy research, information architecture, links, or another targeted intervention.

We do not start with a predetermined tactic.

4. RE-TEST

Measure the thing you intended to change

Re-run the same prompt cohort and compare the new answer environment with the baseline. Record what moved, what did not, and where results differ by platform.

5. INTERPRET

Report only what the evidence supports

AI visibility movement is useful evidence. It is not automatically proof of better buyer outcomes or business impact. When the next question is downstream, we investigate that separately.

That last point is important to retain because the guide explicitly separates AI-answer results from human and organizational outcomes.

Section: What we can investigate

H2

WHERE IS YOUR AI VISIBILITY BREAKING?

Use six cards:

Discoverability
Is the relevant evidence public, current, technically accessible, and easy to find?

Retrieval
Where observable, are AI systems selecting the evidence you expected them to use?

Citations
Which sources are AI answers relying on, and why are competitors or third parties cited instead?

Brand presence
Does your company appear for the situations and criteria that matter?

Recommendations
Are you merely mentioned, or actually presented as a qualified option?

Accuracy and framing
Are your capabilities, conditions, differentiators, and limitations represented correctly?

Section: Interventions

H2

THE INTERVENTION FOLLOWS THE GAP

Copy:

Different visibility problems require different changes.

We may improve owned evidence, consolidate fragmented information, build a more usable comparison resource, earn relevant third-party coverage, supply evidence to publishers, create citation-worthy content, or use targeted outreach and link building.

The goal is not to distribute work everywhere. It is to change the surface that appears to be constraining the result.

That matches the guide’s surface-selection rule.

Section: ZenBusiness proof

H2

FROM INVISIBLE TO RECOMMENDED IN UNDER 3 WEEKS

Copy:

ZenBusiness had relevant information on its own site, but AI systems were not connecting the brand to a priority competitive differentiator.

Xofu identified an off-domain evidence gap. Citation Labs developed and deployed comparison content designed to strengthen that connection in the source environment.

The result:

Google AI Mode citation presence: 6.3% → 33.3%
Google AI Mode rank-one visibility: 74.2% → 88.9%
Google Gemini rank-one visibility: 45.8% → 62.2%

CTA:

Read the ZenBusiness Case Study

Keep Citation Labs’ role bounded to what the case actually establishes. The published case describes Citation Labs as the execution partner for content strategy and production. (Citation Labs)

Section: Measurement/research

H2

MEASURE AI VISIBILITY WITHOUT PRETENDING ONE SCREENSHOT IS A KPI

Copy:

AI answers vary. That does not make measurement useless.

In one client study, Citation Labs tracked 4,579 prompts across 458 buying-intent cohorts over 12 weeks. The work showed why repeated measurement, fixed prompt cohorts, Presence Rate, Recommendation Rank, and source tracking are more defensible than single-run rankings.

We use the same principle in client work: preserve the cohort, repeat the measurement, save the evidence, and log the changes.

CTA:

Read the 4,579-Prompt Study

The published research explicitly argues for locked cohorts and repeated measurement rather than single-output rank claims. (Citation Labs)

Service page FUQs

What is AI visibility consulting?

AI visibility consulting investigates how your brand appears across AI search and answer environments, why those systems cite or recommend particular sources and companies, and what intervention could improve the measured result.

Is this the same as GEO or AEO?

GEO, AEO, AI SEO, and AI visibility are overlapping market terms. We use AI visibility as the broader category because the work can involve retrieval, citations, mentions, recommendations, accuracy, source environments, and measurement, not only optimizing content for an answer engine.

Do you just create more content?

No. Sometimes content is the right intervention. Sometimes the useful evidence already exists but is fragmented, difficult to verify, or missing from the third-party sources shaping the answer. We diagnose the gap before choosing the tactic.

Do links still matter?

They can. Links, publisher coverage, brand mentions, comparison pages, research, documentation, and other sources can all participate in the information environment. We use link building when the diagnosis gives us a reason to believe it can affect the targeted result.

How do you measure AI visibility?

We define a tracked prompt cohort, preserve the measurement conditions, record how the brand and relevant sources appear, and repeat the test over time. Where the surface supports it, we track citations and source changes alongside brand presence and recommendation behavior.

Can you guarantee that ChatGPT or Gemini will recommend us?

No. AI environments are probabilistic and change over time. We can diagnose the current environment, make evidence-based interventions, and measure whether the targeted conditions change.

How is Citation Labs different from a traditional GEO agency?

We do not begin with a fixed GEO playbook. We identify where visibility breaks, choose the intervention that fits that diagnosis, and define what should change before the work ships.

4. About page

Do not rebuild the page or erase the history. Replace the company-level definitions that currently say “link-building agency” and “team of link-building strategists.” (Citation Labs)

Hero copy

Keep:

ABOUT CITATION LABS

Replace the current subhead with:

Citation Labs is an AI visibility and search consultancy helping enterprise teams improve how their brands are found, cited, compared, and recommended across search and AI.

Optional culture line immediately afterward:

We’re family-first, pet-first, remote, curious, and unusually willing to build the tool when the tool we need doesn’t exist yet.

Our Story opening

Replace current opening paragraphs with:

Citation Labs started in search and link building. Over more than a decade, that work expanded into a broader question: how do brands earn visibility and credibility wherever buyers and retrieval systems look for evidence?

Today, our team spans search strategy, research, content, outreach, software, and measurement. We investigate visibility problems across search and AI, then build the intervention the evidence calls for.

We still believe in the principle that shaped the company from the beginning: useful expertise deserves to be easier to find.

Then leave the historical timeline largely intact.

Mission

Replace the current link-volume mission. The live mission currently centers on getting in-house SEOs hundreds of links every month. (Citation Labs)

Use:

OUR MISSION

Help search and growth teams understand and improve how their brands are represented across search and AI, using evidence, interventions, and reporting they can defend internally.

Vision

Use:

OUR VISION

Make the information environments people and AI systems rely on more useful, accurate, and capable of supporting better decisions.

Where We’re Heading

Replace the current “drive links / show up in search / get mentioned in LLMs” copy. (Citation Labs)

Use:

WE’RE BUILDING BETTER WAYS TO DIAGNOSE, INFLUENCE, AND MEASURE VISIBILITY ACROSS SEARCH AND AI.

Through The Lab, we test how sources, citations, content, links, buyer evidence, and information architecture affect what search and AI systems find, represent, and recommend.

The work becomes new research, new services, and, when necessary, new software.

CTA:

EXPLORE THE LAB

Do not rewrite individual staff bios in this wave.

5. Results/case-study page

The live page currently leads with three link-building cases before ZenBusiness and ends with the link-building-only company descriptor. (Citation Labs)

SEO title

AI Visibility, SEO & Link Building Case Studies | Citation Labs

Meta description

See how Citation Labs helps enterprise teams improve AI visibility, search performance, citations, authority, and the information buyers rely on.

H1

Keep:

CASE STUDIES

Add subhead

See how enterprise teams use Citation Labs to improve visibility across AI, search, content, and the sources buyers rely on.

Featured case

Place ZenBusiness above the main grid.

Label:

AI VISIBILITY

Title:

ZenBusiness Wins AI Visibility for a Key Competitive Differentiator in 3 Weeks

Proof line:

Google AI Mode citation presence increased from 6.3% to 33.3%, with additional gains in AI Mode and Gemini recommendation visibility.

CTA:

Read the Case Study

Categorize the remaining cases

Use visible labels:

• AI Visibility
• Link Building
• Content & Research

Do not retrofit AI labels onto old link-building work.

Add section after case grid

RESEARCH BEHIND THE WORK

Our client work is informed by ongoing research into how buyers, search engines, and AI answer environments find and use information.

Feature:

4,579 Buyer Prompts
How repeated prompt cohorts make AI visibility measurable.

Decision Architecture
How AI citations and source sets intersect with buying decisions.

AI Mode Shortlisting Study
How 48 participants used Google AI Mode versus traditional search when building consideration sets.

The current behavioral study documents the 48-participant work and reports that 74% of AI Mode shortlists came directly from AI output. (Citation Labs)

6. Internal-link map

Implement this deliberately rather than sprinkling links everywhere.

This should communicate:

Citation Labs → AI Visibility service → methodology/research → proof → measurement

Do not force commercial links into articles where they do not belong.

7. Contact page

Where the site describes possible engagements, use this order:

AI visibility consulting, strategic link building, citation optimization, or citation-worthy content and source coverage

No broader rewrite required.

8. Structured-data check

Do this after visible copy ships.

Check:

Organization

• name = Citation Labs
• description aligns with global company descriptor
• sameAs references accurate/current
• logo/url correct

Service

• AI service named AI Visibility Consulting
• description reflects diagnosis/intervention/measurement
• do not create separate fake services for GEO/AEO merely to stuff synonyms

Breadcrumbs

• Services → AI Visibility

Do not use schema to introduce claims/category language that the visible page does not support.

9. Release/change log

Ship the owned work as Wave 1 and record one primary release date.

Log individually:

• homepage metadata
• homepage copy/proof
• global navigation
• global boilerplate
• AI service rewrite
• About rewrite
• Results architecture
• internal links
• Contact update
• structured data

Also record any AI articles, PR, external profile changes, or major site releases during the observation window because those could complicate the re-test. The Field Guide specifically calls for logging site releases, PR, third-party updates, model/platform changes, and competitor activity.

Ship order

Developer/template first

1. Navigation
2. Footer
3. Newsletter
4. global descriptor

Page edits
5. Homepage
6. AI Visibility service
7. About
8. Results
9. Contact

Then
10. Internal links
11. Structured data
12. QA/indexability check
13. Log release date

After that release, start the external-source work already scoped. Do not run another formal prompt test immediately. The frozen V1 cohort is for the post-intervention re-test.

Next is the external implementation pack. Do this only after the owned changes are live.

Wave 1 targets

Start with five:

1. Search Engine Land author bio
   • Current bio still calls Citation Labs a “research and link-building agency,” while also saying Garrett’s current focus is AI-generated-answer visibility. (Search Engine Land)
   • Ask for a factual bio update.
2. Clutch GEO category
   • Clutch now has a dedicated GEO category built around firms helping brands get cited/recommended in AI answers. (Clutch)
   • Make sure Citation Labs is listed/classified there with the AI service and ZenBusiness proof.
3. Built In AI-search provider article
   • Exact category fit. Their article covers agencies/platforms helping brands appear in AI-generated results and explicitly references content, third-party coverage, citations and external signals. Citation Labs is absent. (Built In)
   • Treat this as an editorial tip, not a backlink request.
4. AEO Vision directory
   • Strong fit. Their inclusion criteria explicitly include active AEO/GEO delivery, measurable client AI visibility improvement and public methodology/results. (AEO Vision)
   • Citation Labs has evidence against those criteria.
5. PikaSEO AI SEO agency list
   • They evaluate technical depth, documented AI-citation results, measurement and specialization. (PikaSEO)
   • Again, good evidence fit.

Evidence pack

Use the same core evidence everywhere:

Company

Citation Labs is an AI visibility and search consultancy helping enterprise teams diagnose and improve how their brands are cited, represented, and recommended across AI answer environments.

Client proof

ZenBusiness: Google AI Mode citation presence increased from 6.3% to 33.3% after the intervention.

Method proof

Citation Labs tracked 4,579 buyer prompts using fixed cohorts and repeated measurement across AI/search environments.

Differentiator

The work diagnoses where visibility breaks first, including source/citation/recommendation gaps, then selects the intervention instead of starting with a predetermined GEO tactic.

Send two links initially:

• AI Visibility Consulting page
• ZenBusiness case

Offer the methodology/research if useful.

Outreach copy

Search Engine Land bio update

Hi [Name],

Small factual update request for Garrett French’s contributor bio. Citation Labs’ current positioning has moved beyond the older “research and link-building agency” description into AI visibility and search consulting.

Suggested update:

“Garrett French is the founder of Citation Labs, an AI visibility and search consultancy that helps brands understand and improve how they are cited, represented, and recommended across AI answers and search. He also leads Xofu and ZipSprout.”

Happy to send the updated company page if useful.

Editorial/provider-list outreach

Hi [Name],

I came across your comparison of AI visibility/GEO providers. Citation Labs isn’t currently included, and I think our recent work may now meet the bar you’re using.

We have a documented enterprise case showing engine-specific before/after AI visibility movement, plus published methodology based on 4,579 buyer-prompt runs. Our work also focuses heavily on third-party source and citation gaps rather than treating GEO as an on-page content program.

If you revisit the list, happy to send the evidence against your inclusion criteria.

Then add one target-specific sentence.

For AEO Vision:

Your criteria around measurable client movement and public methodology are why I thought it was worth flagging.

For PikaSEO:

Your emphasis on documented AI-citation results and measurement is why I thought Citation Labs might now fit the methodology.

For Built In:

Your framing around third-party coverage and external signals closely matches the work we’ve been testing.

Tracking sheet

Track only:

One follow-up max.

Stop condition

Do not turn this into 50-site outreach.

First wave = 5 targets.

If 2–3 meaningful sources change, that is enough to justify observing whether the answer environment starts moving before expanding the campaign.

After these five are underway, the next step is launch QA and change-log setup for the owned intervention, then we wait for the T+21 re-test.

Next is launch QA + change-log setup for the owned intervention.

1. QA the owned changes before publish

Check these as one release.

Global

• Nav says AI Visibility
• No remaining global LLM Mentions service labels
• Footer/company descriptor updated
• Newsletter boilerplate updated
• Shared CTAs/boilerplate no longer define Citation Labs only through link building

Homepage

• New SEO title/meta live
• Hero subhead updated
• AI-specific proof strip live
• ZenBusiness featured first
• AI Visibility Consulting service label consistent
• “band” typo fixed

AI Visibility service

• New H1/title/meta
• Methodology follows baseline → diagnose → intervene → re-test
• Legacy “training data”/“LLM rank” language removed where specified
• AI-specific proof used
• ZenBusiness linked
• 4,579-prompt research linked
• CTA consistent
• URL remains /llm-mentions/ for this wave

About

• Opening company definition updated
• “link-building agency” removed as the current company category
• Mission updated
• “Where We’re Heading” updated
• Historical link-building timeline retained

Results

• ZenBusiness first/featured
• Case categories visible
• AI research section added
• Old link-building-only company boilerplate removed
• AI service linked from relevant proof

Internal links

Verify the intended chain works:

Company → AI Visibility service → methodology/research → client proof → measurement

No broken links. No accidental redirect chains.

2. Technical QA

Before calling the intervention live:

• page titles actually render in source
• meta descriptions updated
• canonicals unchanged/correct
• no accidental noindex
• robots rules unchanged
• sitemap contains the edited pages
• structured data still validates
• mobile/desktop rendering looks normal
• new internal links are crawlable HTML links
• any shared template changes propagated across indexed pages

Do not change unrelated technical SEO during this release. It would muddy the intervention.

3. Create the change log

Use one simple sheet.

Then add external changes later in the same log:

• Search Engine Land bio
• Clutch
• AEO Vision
• Built In
• PikaSEO
• any other material source update

The guide specifically says to log site releases, PR activity, third-party updates, platform/model changes, and competitor activity so later movement is interpretable.

4. Establish the intervention date

Use the date when the major owned changes are all live, not the date the first page gets edited.

Call that:

Owned Intervention Wave 1: [DATE]

If the rollout takes several days, record the individual dates, but use the final material publish date as the primary intervention date.

5. Do not re-test immediately

Once live:

• T+7: optional diagnostic spot-check only
• T+21: formal re-test of the frozen seven-prompt cohort
• same prompts
• same platforms
• same testing conditions

The measurement setup needs to remain comparable to the baseline.

6. Start external work after owned QA passes

You do not need to wait 21 days to start external outreach.

Sequence:

1. Publish owned wave
2. QA it
3. Record intervention date
4. Start the five external targets
5. Log every external change
6. Re-test at T+21
7. Interpret owned movement separately from any external changes already live

That is the next execution step.

Next is the post-launch re-test protocol and interpretation framework. Prepare it now, but don’t run it until the owned wave has been live for ~21 days.

Formal re-test

Run the frozen P1–P7 cohort exactly as baseline:

• ChatGPT Temporary Chat
• Gemini clean/non-personalized session
• Perplexity clean session
• one fresh conversation per prompt
• prompt only
• no forced search
• no added citation request
• no follow-up

Do not change wording.

Score P1–P5

For each response, record:

Primary metric remains:

Citation Labs inclusion across P1–P5: baseline 0/15

Do not invent a percentage target now. Any credible movement from 0 is evidence worth inspecting.

Score P6 separately

Prompt:

What does Citation Labs specialize in?

Use the frozen categories:

• AI-primary
• Co-primary
• AI-secondary
• AI-absent

Baseline:

• ChatGPT: AI-secondary
• Gemini: Co-primary
• Perplexity: Co-primary

This is the cleanest measure of whether the owned category cleanup changed brand understanding.

Keep P7 as control

Prompt:

What are the best link-building agencies or consultancies?

Baseline:

Citation Labs: 1/3

We are not trying to make this go to zero.

If AI-category inclusion improves while link-building recognition remains, that is a good result. It means Citation Labs expanded the category association rather than destroying existing equity.

How we interpret the first re-test

Result A: P6 improves, P1–P5 stay at 0/15

Strong signal that the owned-site work fixed entity/category understanding, but did not solve unbranded consideration.

That would increase confidence that external category evidence is the binding constraint.

Result B: P6 improves and P1–P5 start including Citation Labs

Best initial result.

Then inspect why:

• Citation Labs-owned pages?
• external sources?
• which platforms?
• which prompts?
• recommended or merely mentioned?

Do not reduce this to “visibility increased.”

Result C: P6 unchanged, P1–P5 stay at zero

Owned intervention produced no detectable movement under these conditions.

Then inspect:

• whether systems are using the updated pages
• indexing/retrieval lag
• whether stronger old signals still dominate
• whether the diagnosis needs revision

Do not immediately publish more content.

Result D: P1–P5 improve but P6 does not

Interesting and plausible if external sources moved quickly.

That would suggest the provider-selection ecosystem can influence category inclusion even while branded entity framing remains messy.

Result E: competitors all move substantially too

Be careful.

If Citation Labs and multiple competitors move together, a platform/model change may explain part of the result. The guide specifically says comparison reduces ambiguity but does not turn the work into a causal experiment.

After the first re-test

We make one of four decisions:

• Maintain if the result is already strong enough.
• Continue external work if owned framing improved but category inclusion remains weak.
• Refine the intervention if movement is partial or points to a narrower source gap.
• Re-diagnose if nothing moves and the updated evidence is clearly being retrieved.

Only after we establish a meaningful AI-answer result do we return to your original business question:

Are better AI-category signals associated with a change in the kinds of inbound meetings Citation Labs receives?

That becomes the next guide/business-impact leg. We should not treat AI-answer movement as proof of lead impact.

At this point, the investigation design is complete. The next actual work is implementing the owned-site changes. There isn’t another useful analysis step before that.

Citation Labs

Citation Labs helps brands improve visibility in search results and AI answers through citation optimization, citation-worthy content, and contextually relevant link building at scale.

We help brands show up more clearly in recommendation and comparison environments, earn citations that influence shortlists, and reduce friction for buyers deciding which brand to trust and choose.