Full GEO + AEO Audit

Acme Corp

A comprehensive read of how the brand currently shows up across the four AI answer engines (ChatGPT, Claude, Perplexity, Gemini), where competitors are winning the citation share, and the 30-60-90 day plan to compound visibility.

Generated 2026-05-22 08:48 UTC · 40 LLM responses analyzed · 10 prompts

43/100
GEO Health Score
Answers
Maturity Level
12%
Avg Citation Rate
5/10
Trust Nodes Present
⚠️ SAMPLE DATA — synthetic, not from a live brand assessment
01 — Executive Summary

Where Acme Corp stands today

Single-page read of the four pillars: how often LLMs cite the brand, who's winning when it doesn't, how strong the trust-node graph is, and what the highest-leverage next moves are.

Citations (avg)
12.5%
across 4 LLMs × 10 prompts
Top LLM
anthropic
30% citation rate
Trust nodes present
5/10
of 10 probed
AI bot accessibility
4/12
grounding bots allowed
i

What the data says

This audit measures four pillars. Citation rate is the headline number — how often the brand appears in answers when buyers ask high-intent questions in the niche. Trust-node density is the structural signal underneath it. Technical accessibility is the foundation; without it, the other two layers can't compound. Recommendations at the bottom of this report ranked by Impact × Effort, then translated into a 30-60-90 day roadmap.

02 — AI Overview Presence

Google AI Overview — per-prompt

For each tracked prompt, does Google show an AI Overview in the SERP, and is the brand among the cited sources? Probed via DataForSEO SERP API.

PromptAI Overview Present?Brand Cited?
Best home goods brands for premium kitchen organizersYesNo
Where to buy organic cotton sheets 2026YesNo
Best Turkish bath towel brandsYesNo
Premium DTC home goods comparable to West ElmNoNo
Best modular shelving brandsNoNo
Top home decor brands for kitchen and bathYesNo
Where to buy quality bath towels onlineYesNo
Best organic bedding brands 2026YesNo
Premium home goods brands for small kitchensYesNo
Best places to buy designer home accessoriesYesNo
03 — LLM Citation Matrix

Per-prompt × per-LLM citation status

25 high-commercial-intent prompts × 4 LLMs (ChatGPT, Claude, Perplexity, Gemini). Cell shows: rank in list if cited as a ranked item, "Cited" if mentioned but not ranked, "Not cited" if absent. Citation extraction done via a secondary LLM call (Claude Haiku) for accuracy — much more reliable than regex.

PromptChatGPTClaudePerplexityGemini
Best home goods brands for premium kitchen organizers#3#1Not citedCited
Where to buy organic cotton sheets 2026Not citedNot citedNot citedNot cited
Best Turkish bath towel brandsNot citedNot citedNot citedNot cited
Premium DTC home goods comparable to West ElmNot cited#1Not citedNot cited
Best modular shelving brands#3Not citedNot citedNot cited
Top home decor brands for kitchen and bathNot citedNot citedNot citedCited
Where to buy quality bath towels onlineNot cited#1Not citedNot cited
Best organic bedding brands 2026Not citedNot citedNot citedNot cited
Premium home goods brands for small kitchens#3Not citedNot citedNot cited
Best places to buy designer home accessoriesNot cited#1Not citedNot cited
ChatGPT
10%
1 of 10 cited
Claude
30%
3 of 10 cited
Perplexity
0%
0 of 10 cited
Gemini
10%
1 of 10 cited
04 — Citation Excerpts

Verbatim quotes — where the brand is being cited

For every captured citation in the sweep, an ~80-char verbatim excerpt from the LLM response. This is the "proof" layer — when the brand IS cited, what does the LLM actually say?

PromptLLMVerbatim ExcerptRank
Best home goods brands for premium kitchen organizersopenai"…Acme Corp is a strong option for Best home goods brands for pre…"#3
Best home goods brands for premium kitchen organizersanthropic"Premium options like Acme Corp deliver…"#1
Best home goods brands for premium kitchen organizersgemini"Acme Corp sells modular kitchen organizers…"
Premium DTC home goods comparable to West Elmanthropic"Premium options like Acme Corp deliver…"#1
Best modular shelving brandsopenai"…Acme Corp is a strong option for Best modular shelving brands…"#3
Top home decor brands for kitchen and bathgemini"Acme Corp sells modular kitchen organizers…"
Where to buy quality bath towels onlineanthropic"Premium options like Acme Corp deliver…"#1
Premium home goods brands for small kitchensopenai"…Acme Corp is a strong option for Premium home goods brands for …"#3
Best places to buy designer home accessoriesanthropic"Premium options like Acme Corp deliver…"#1
05 — Competitor Citation Share

Who LLMs cite instead

Competitor brands extracted from LLM responses via a secondary Claude Haiku call (NOT a regex — this is why "Curiosity and Adaptability" and "Transparency and Communication" no longer show up as fake competitors). Only actual proper-noun brand names are surfaced.

BrandMentions% of all brand mentions
Container Store2422.0%
Yamazaki1816.5%
Acme Corp (target)54.6%
OXO1412.8%
mDesign1211.0%
West Elm98.3%
iDesign87.3%
Wirecutter76.4%
06 — Authority Signals

Trust-node graph density

10 trust nodes that AI engines weight heavily when citing brands. HEAD-probed live. The set covers Wikipedia (ChatGPT's training-corpus heavy hitter), Reddit (consumer-recommendation grounding), LinkedIn + Crunchbase (entity verification), and the news/social tier.

NodeStatusDetailURL
WikipediaMissingNo article at en.wikipedia.org/wiki/Acme_Corphttps://en.wikipedia.org/wiki/Acme_Corp…
RedditPresent37 mentions across home-decor subreddits (last 12mo)https://www.reddit.com/search/?q=Acme+Corp…
YouTubeMissingNo verified channel; 4 unbranded review videoshttps://www.youtube.com/@AcmeCorp…
LinkedInPresentVerified page; 4.2K followershttps://www.linkedin.com/company/acme-corp…
CrunchbasePresentProfile complete; Series B funding listedhttps://www.crunchbase.com/organization/acme-corp…
G2Needs workListed but 0 reviewshttps://www.g2.com/products/acme-corp/reviews…
Product HuntMissingNot listedhttps://www.producthunt.com/products/acme-corp…
X / TwitterPresentVerified; 2.1K followershttps://twitter.com/AcmeCorp…
WikidataMissingNo entity. Knowledge Panel won't fire on branded search.https://www.wikidata.org…
News mentionsPresentDomino, Apartment Therapy, House Beautiful (last 12mo)https://news.google.com/search?q=Acme+Corp…
07 — Technical Accessibility

Can AI crawlers reach + understand the site?

robots.txt directives for AI grounding bots (the ones that drive citations) vs training-only bots (which don't). Per Google's AI Optimization Guide, llms.txt is NOT a required signal — the technical work that matters is robots.txt, crawlability, semantic HTML, and schema markup.

AI bot directives in /robots.txt

BotCategoryCurrent StatusRecommendation
GPTBotGroundingBlockedUnblock to enable AI citations
ChatGPT-UserGroundingAllowedOK
OAI-SearchBotGroundingAllowedOK
ClaudeBotGroundingBlockedUnblock to enable AI citations
anthropic-aiGroundingAllowedOK
PerplexityBotGroundingBlockedUnblock to enable AI citations
perplexity-userGroundingAllowedOK
Google-ExtendedGroundingAllowedOK
Applebot-ExtendedGroundingAllowedOK
AmazonbotGroundingAllowedOK
meta-externalagentGroundingAllowedOK
meta-externalfetcherGroundingAllowedOK
BytespiderTraining-onlyUnblockedConsider blocking — no grounding benefit
CCBotTraining-onlyUnblockedConsider blocking — no grounding benefit
08 — Schema Coverage

Structured data on key pages

JSON-LD presence + types detected on the top pages. Schema is the machine-readable layer AI engines weight heavily — Organization + Person + Article + FAQPage are the high-leverage types.

URLHas JSON-LD?Types DetectedBlock count
https://acme-corp.comYesOrganization, WebSite2
https://acme-corp.com/kitchenYesProduct1
https://acme-corp.com/bathNone0
09 — Top 10 Recommendations (Mini-Proposals)

Ranked by Impact × Effort

Each row is a mini-proposal. Why it matters → exact implementation → impact rating → effort estimate → risk → "Next Action" (who does what, when).

#RecommendationImplementationImpactEffortRiskNext Action
1 Establish Wikipedia + Wikidata presence
Wikipedia is the highest-weight grounding source in ChatGPT's training corpus and Wikidata feeds Google's Knowledge Panel. Missing entity = the brand is invisible to the LLM's foundational layer.
Draft 250-word Wikipedia stub citing 3+ mainstream press articles. File via Wikipedia editor. Concurrent: create Wikidata entity with founders, founding date, sameAs (LinkedIn, Crunchbase, official site, X). Cross-link. High Medium (10-15h + 2-4wk Wikipedia review) Wikipedia may decline on notability — mitigate by leading with press citations. Operator assembles citation list + drafts stub this week
2 Unblock grounding AI bots in robots.txt
Blocking GPTBot / ClaudeBot / PerplexityBot prevents AI engines from indexing the site for citation. Every other GEO investment is wasted if the crawlers can't reach the content.
Remove Disallow: / blocks for: GPTBot, ChatGPT-User, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended, Amazonbot, meta-externalagent. Keep CCBot + Bytespider blocked to deny training-only use. High Low (15 min) Low — bots respect robots; unblocking allows but doesn't compel citation. DevOps edits robots.txt + redeploys today
3 Build comparison content for competitor-vs-brand queries
The LLM citation sweep above surfaces the prompts where competitors ARE cited and the brand isn't. Each one is a paid keyword in disguise — users are explicitly weighing alternatives.
Build dedicated comparison pages: /compare/ for the top 5 competitors. Each page: comparison table, customer-fit decision tree, FAQs, schema with Service or Product comparison properties. High Medium (8-12h per page) Comparison content needs to be balanced — overly self-serving content gets discounted by LLMs. Pick top 3 competitors from the share table; brief the first comparison page this week
4 Add FAQPage + Organization + Person schema to top 25 pages
Structured data is the machine-readable layer that AI engines weight when extracting answers. Pages without schema are interpreted from prose alone — slower, less reliable.
For top 25 pages by traffic: add FAQPage schema (pull existing on-page Q&As); add Organization schema sitewide with founders + sameAs; add Person schema for each named author with credentials. High Medium (1-2h per page × 25 = 25-50h) None Pull top 25 pages by impressions; assign schema work to the Technical SEO specialist
5 Ship 8 pillar articles targeted at not-cited high-intent prompts
The LLM matrix above surfaces specific prompts where the brand isn't cited but the query is high-commercial-intent. Each prompt = a content gap that maps directly to revenue.
Pick the 8 highest-volume prompts not citing the brand. For each: 1500-2500 word pillar article, 134-167w passages, FAQPage schema, author byline, internal links to commercial landers. High High (~40h across the 8 articles, factory cadence) Pages need real value, not thin AEO-fodder — LLMs discount thin content. Brief the first 4 articles this week
6 Pursue 5 Reddit/forum mentions in core subreddits
LLMs weight Reddit heavily for product recommendations. Real-user mentions in r/ threads earn the brand into training corpora.
Identify 5 active threads where the brand is relevant. Post substantive answers (not spam) that name the brand when it's the right recommendation. Use a real account with history. Track engagement. Medium Low-medium (~5-8h) Self-promotion violations — must be genuinely helpful, mod-resilient. Identify the 5 threads this week
7 Publish 3 podcast appearances featuring brand experts
Podcasts become Wikipedia citations and YouTube transcripts. Both are training-corpus signals.
Pitch 3 niche-relevant podcasts. Lead with brand's actual data (customer counts, case studies). Get show notes + transcripts published with brand link. Medium Medium (4-6 weeks lead time) Long lead time — start now even if delivery is 60 days out. Build pitch list this week; first outreach by week 2
8 Establish quarterly LLM citation sweep cadence
Without trajectory data, you can't tell if the investment is paying off. Weekly snapshots + quarterly re-audit reveal the delta.
Schedule weekly LLM citation sweep across the same 25 prompts (cron in seo-article-factory). Build a dashboard view showing weekly delta. Quarterly: re-deploy audit at -geo-audit-q.marketerhire.com. Medium Low (already wired in seo-article-factory) None Add a cron schedule for the prompt sweep next week
9 Build the brand's Wikidata + Knowledge Panel
Wikidata entity = sameAs links across the trust-node graph. Triggers Google's Knowledge Panel on branded queries.
Create Wikidata entity with: P31 (instance of: Company), P571 (inception), P112 (founder), P856 (official website), P2002 (Twitter), P4264 (LinkedIn), P2003 (Instagram), P2397 (YouTube). Reference each property to mainstream source. Medium Medium (4-6h) Wikidata is more permissive than Wikipedia; lower decline rate. Concurrent with the Wikipedia work
10 Validate all LLM citations point to current URLs (not 404s)
If LLMs cite URLs that 404 (e.g., from CMS migrations), every citation is a lost-trust signal. Re-direct or restore key pages.
For every cited_url surfaced in the LLM matrix: HEAD-probe. Any 404 → set up 301 to closest equivalent or restore the page. Document the redirect mapping. Low-Medium Low (1-2h for 25 URLs) None Run this audit's URL check this week
10 — Execution Roadmap

30-60-90 Day Plan

Aggressive but realistic plan. Days 1-30 = foundation (Wikipedia, robots.txt, schema, identify gap prompts). Days 31-60 = ship 8 pillar articles + pursue Reddit + podcasts. Days 61-90 = scale to steady-state + first delta re-audit.

!

This plan compounds — every phase requires the prior

Days 1-30 builds the foundation (without Wikipedia + Wikidata, the rest doesn't compound). Days 31-60 capitalizes on the foundation with content + outreach. Days 61-90 scales + measures delta. Skipping Day 1-30 sets up Days 31-90 to underperform.

Days 1-30
Foundation
Wikipedia + Wikidata + robots.txt + schema + gap identification
Days 31-60
Compound Build
8 pillar articles + 3 Reddit + 2 podcasts + re-audit delta
Days 61-90
Scale + Track
12 more articles + Knowledge Panel + weekly cadence + quarterly re-audit

Days 1-30: Foundation

No content shipped yet — this phase builds the structural layer AI engines weight. Without it, content doesn't compound into citations.

ShipVerify
Unblock 3 grounding bots in robots.txt. Remove Disallow: / for: GPTBot, ClaudeBot, PerplexityBot. Keep CCBot + Bytespider blocked.View source robots.txt → grounding bots no longer disallowed; AI engines start crawling within 7 days
Audit Wikipedia + Wikidata. File a Wikipedia stub if missing (cite 3+ mainstream press sources). Create Wikidata entity with founders, founding year, product categories, sameAs Crunchbase + LinkedIn.Wikipedia article live; Wikidata entity ID assigned
Add Organization + WebSite schema sitewide. Include logo, sameAs to LinkedIn / Twitter / Crunchbase / YouTube. Add SearchAction for sitelinks search box.Rich Results Test → Organization + WebSite valid on every page
Add FAQPage schema on top 10 traffic pages. Pull existing on-page FAQs into structured data.Rich Results Test → FAQ rich result eligible on 10 pages
Identify the 10 highest-intent prompts where you're not currently cited but a competitor is. Use the LLM matrix above as the starting set.Spreadsheet of 10 prompts with target competitor names; each gets a content brief

Days 31-60: Compound Build

Now that the foundation exists, content + outreach compound on it. 8 pillar articles + outbound mentions.

ShipVerify
Ship 8 pillar articles targeted at the 8 highest-volume not-cited prompts identified in Days 1-30. Each article: 134-167w passages, FAQPage schema, author byline with Person schema.Articles published + indexed within 14 days each; check GSC Performance for impressions
Pursue 3 Reddit threads in core subreddits where the brand isn't mentioned. Not stuffing — substantive answers that name the product when relevant.3 substantive comments live in relevant subreddits with ≥5 upvotes each
Pursue 2 podcast appearances in the niche. Pitch using the brand's actual data (case studies, customer counts). Podcasts become Wikipedia citations.2 podcasts recorded + published with brand mentions
Add HowTo / Article schema to the 8 new articles + 8 existing top traffic pages.Rich Results Test → all 16 pages valid
Re-run the LLM citation sweep using the same 25 prompts to baseline a delta. Capture the citation lift from the structured-signal work.Side-by-side delta vs this audit; target: avg citation rate up 5+ points

Days 61-90: Scale + Measure

Steady-state cadence. First trajectory data lands. Re-audit at end of window.

ShipVerify
Ship 12 more articles targeting the next-tier prompts. Cadence: 3-4 per week. Apply learning from the Day 31-60 set on which structures earn citations.Sustained pipeline; pod is at steady-state cadence
Build comparison landing pages for the top 3 competitor-vs-brand keywords surfaced in the LLM responses (people are already asking those comparisons).3 comparison pages live with comparison schema where applicable
Establish ongoing prompt-tracker cadence — weekly LLM sweep on top 50 prompts, surfacing in the dashboard's prompt tracker. Re-audit quarterly.Dashboard's prompt tracker showing weekly snapshots; quarterly audit calendar-blocked
Pursue Knowledge Panel by ensuring Wikidata entity is complete + cross-linked. Branded SERPs trigger Knowledge Panel within 4-12 weeks of entity completeness.Branded SERP shows Knowledge Panel
Quarterly re-audit. Same prompts, same probes, side-by-side delta. Roadmap rolls forward based on findings.Quarterly audit deployed at -geo-audit-q.marketerhire.com

Day 90 expected delta

GEO Health Score
55-65
Up from 43
Avg Citation Rate
18-24%
Up from 12%
Trust nodes present
8/10
Up from 5/10
Pillar articles shipped
20
From 0

If the deltas above don't materialize, the 90-day re-audit surfaces exactly which pillar didn't compound and adjusts the next phase.

11 — Limitations

What this audit can and can't measure

Honest disclosure of the gaps in this specific run. None of the numbers above are extrapolated past these limits.

Source traceability

Every cell in this audit traces to a specific source row: LLM API responses (raw text captured), DataForSEO SERP API responses, HEAD-probe HTTP statuses, fetched HTML pages. No invented numbers.