GEO Checklist: AI Visibility in GPT‑5, Gemini & Co.
GEO checklist for AI visibility: 20-point audit with llms.txt, Schema, FAQs & GEO Score. Tracking for GPT-5, Gemini, Claude & Perplexity.
This 20-point checklist shows in 15 minutes whether your website is ready for AI recommendations — or whether you remain invisible. Print it out, work through it step by step, and know exactly where you stand afterward. You will also receive concrete priorities for quick wins. This reduces uncertainty and helps you avoid costly detours.
Short answer: For greater visibility in ChatGPT, Gemini, Claude, and Perplexity, you need three things: clear, citable answers (Answer-First + FAQ Schema), verifiable company data (Organization/FAQ/Article Markup, consistent NAP), and ongoing GEO monitoring across all models (e.g., with ai-geotracking.com).
Generative Engine Optimization (GEO) increasingly determines whether your business is recommended by AI models. However, many teams do not know whether their website meets the basic requirements. That is exactly what this GEO audit is for: it translates AI requirements into 20 clear, verifiable measures — including scoring, prioritization, and concrete action steps. You will also learn how strategies for AI search differ from traditional SEO and how to effectively combine both.
Why is a GEO audit critical right now?
Research is visibly shifting into AI dialogues and AI search. According to Gartner, Generative AI is significantly accelerating the transformation of search and application usage (source). Google confirms that structured data supports presentation in search and AI features (source). OpenAI demonstrates with GPTBot that websites can actively control AI crawler access (source). Therefore, a GEO audit is the fastest path to clarity about status, gaps, and priorities.
Concretely: when someone asks ChatGPT "Which marketing agency in Hamburg is good for B2B?", your GEO optimization determines whether you are mentioned — or your competitors. A GEO audit creates measurable foundations, sets the right priorities, and prevents you from getting lost in pure SEO fine-tuning that has little impact on AI search.
Which independent sources confirm GEO relevance?
For strategic decisions, you need reliable facts. The following data points are verifiable and show why GEO is a priority right now. They come from official product pages, developer documentation, and established industry sources:
- Google & structured data: Google documents that structured data makes content more understandable for search and AI features and improves presentation (source).
- FAQ/HowTo display changed: Since August 2023, Google only shows FAQ rich results in limited cases, primarily for authoritative sites (source). For GEO, this means: quality and relevance beat pure markup volume.
- OpenAI GPTBot: OpenAI provides GPTBot, a crawler that can be controlled via robots.txt. Websites can thus allow or restrict AI access (source).
- AI dialogue usage: Gartner describes Generative AI as a catalyst for new search and work patterns — content must be curated to be AI-compatible (source).
- LinkedIn reach: LinkedIn reports over 1 billion members worldwide, meaning actively maintained profiles send strong authority signals (source).
These sources provide the strategic foundation: structured, citable content plus controlled AI access and strong off-site signals significantly increase the chance of being mentioned in AI responses. Consequently, you should consistently implement the following audit points.
Which 20 points belong in your GEO checklist?
Work through the following 20 points in four categories. Rate each point from 0 to 5 points (scoring details at the end). Maximum score: 100. For faster navigation, here is an overview with anchor links. In addition, several points include concrete verification methods and examples so you can take action more quickly.
- llms.txt in root
- Organization Schema
- FAQ Schema
- Article/BlogPosting
- URL + internal links
- H1/H2/H3 hierarchy
- Answer-First
- FAQ sections
- Definitive language
- Regular content
- Consistent NAP
- Directories
- Google Business
- Reviews
- Measure GEO Score
- All 4 models
- GEO keyword set
- Content refresh
- Competitor monitoring
Which technical foundations are mandatory (points 1–5)?
The technical foundation determines whether AI models can properly read and understand your website at all. Without these basics, all further measures are ineffective. Therefore, focus first on crawl-friendliness, markup quality, and clean information architecture — only then refine the content.
- [ ] 1. llms.txt created and placed in root
The llms.txt file tells AI models which information about your company is relevant. It functions like a curated table of contents for LLMs, pointing for example to product pages, pricing, terms, press, team, and references.
How to check it: Navigate to yourdomain.com/llms.txt. If a 404 page appears, the file is missing. Also add a last-updated date so models can recognize freshness.
Why it matters: AI models like Claude and Perplexity actively read llms.txt. Without this file, you lose control over how AI categorizes your company. By clearly prioritizing sources, you reduce hallucinations and promote accurate citations.
Structure, templates, and implementation details can be found in our llms.txt guide.
- [ ] 2. Organization Schema Markup implemented
Organization Schema tells search engines and AI models exactly who you are: name, address, industry, logo, contact details, and social media profiles. Use sameAs whenever possible to link official profiles.
How to check it: Open the Google Rich Results Test and enter your URL. Is "Organization" displayed as a recognized type? Also check for valid @id references.
Why it matters: AI models use Schema Markup to verify facts about your company. Without Organization Schema, they fall back on unreliable sources. This lowers your AI trustworthiness — especially in AI search.
Implementation guides for all Schema types can be found in our Structured Data Guide.
- [ ] 3. FAQ Schema on the most important pages
FAQ Schema marks question-and-answer pairs so that AI models can recognize them directly as citable facts. One of the strongest levers for AI Visibility, provided the content is precise and verifiable.
How to check it: Test your top 5 pages for FAQ markup with the Rich Results Test. Ideally: each important page has 3–5 FAQ entries with clear, definitive answers and unambiguous source references where possible.
Why it matters: Pages with FAQ Schema are cited significantly more often by AI models as a source. The FAQ format directly matches the way language models extract information — which is why it increases the probability of being cited.
JSON-LD templates and best practices can be found in the Structured Data Guide.
- [ ] 4. Article/BlogPosting Schema for blog posts
Every blog post needs the appropriate Schema Markup: Article or BlogPosting. This tells AI models the author, publication date, headline, and description in a structured way. Also add dateModified for freshness signals.
How to check it: Open one of your blog posts and check the source code for @type: BlogPosting or @type: Article. Alternatively, use the Rich Results Test.
Why it matters: AI models rate content with structured metadata as more trustworthy. Blog posts without Schema are considered less authoritative; this lowers your chances of being cited in AI responses.
- [ ] 5. Clean URL structure and internal linking
Clear, descriptive URLs and logical internal linking help AI models understand the thematic structure of your website. Each important page should be reachable in at most 3 clicks from the homepage, as deeper pages are crawled less frequently.
How to check it: Are your URLs readable (e.g., /blog/geo-checklist instead of /p?id=4738)? Do your blog posts link to each other? Is there a clear site architecture with hub-&-spoke clusters?
Why it matters: AI models weight pages higher that are well embedded in the overall structure. Isolated pages without internal links are used less often as sources, which limits your AI visibility.
How do you increase content quality for GEO (points 6–10)?
Content is the core of every GEO strategy. AI models prefer content that is clearly structured, definitively worded, and regularly updated. The most important answers should also appear in the first sentences — this minimizes losses through summarization.
- [ ] 6. Clear H1/H2/H3 hierarchy on all pages
Every page needs exactly one H1 heading and a logical hierarchy of H2 and H3 subheadings. AI models use this structure to extract core statements and place them in context.
How to check it: Install a browser extension like "HeadingsMap" and check your top 10 pages. Are there duplicate H1s? Does the hierarchy jump from H2 directly to H4? Correct any jumps consistently.
Why it matters: A clean heading hierarchy helps AI models correctly organize your content and cite it as a source. This increases the consistency of your mentions in AI responses.
- [ ] 7. Answer-First structure (key message first)
Place the most important statement at the beginning of each section. Do not introduce, do not paraphrase — deliver the answer directly. AI models preferentially extract the first 1–2 sentences of a section, which is why concise wording is critical.
How to check it: Read the first sentence of each section on your top pages. Does it immediately answer the question posed by the heading? Or does it start with "In today's world..."? Therefore, systematically revise your opening sentences.
Why it matters: If there is no clear statement at the beginning of a section, your content is skipped. Write short, complete sentences that work as standalone citations — for example with concrete numbers, definitions, or instructions.
- [ ] 8. FAQ section on the top 10 pages
Each of your most important pages should contain a FAQ section with 3–5 relevant questions. Phrase the questions exactly as users would type them into ChatGPT or Perplexity, as this increases semantic matching.
How to check it: Do your product, service, and landing pages each have a FAQ section? Are the questions phrased in natural language style (e.g., "What does GEO Tracking cost?" instead of "Price overview")? Also ensure answers are precise and free of contradictions.
Why it matters: FAQs in natural language style directly match the prompts users send to AI models. This massively increases the probability of citation and strengthens your AI visibility in GPT-5 and Gemini.
- [ ] 9. Definitive language instead of vague phrasing
Replace "could", "possibly", "under certain circumstances" with clear statements. AI models prefer definitive phrasing because it allows them to generate more precise answers and leaves less room for interpretation.
How to check it: Search your content for words like "perhaps", "possibly", "could", "to some extent". Can you replace these with clear facts? Define terms and name responsibilities.
Why it matters: Content with definitive language is used more often as a source. Instead of "GEO could improve your visibility", write "GEO improves your visibility" — as long as you back it up with examples or sources.
- [ ] 10. Regular new content (at least once per week)
AI models with real-time access (Perplexity, Gemini) prefer current content. Those who produce no new content gradually become irrelevant — especially for transactional prompts.
How to check it: When was your last blog post published? Is there a fixed editorial schedule? Minimum standard: one new, substantive post per week, supplemented by monthly updates of top pages.
Why it matters: Websites with regularly new content are rated as more current and authoritative by AI models. Tips for content planning can be found in our Content Strategy Guide.
Which authority signals count for AI (points 11–15)?
AI models recommend businesses that are rated as trustworthy and authoritative. These signals come not only from your website, but from the entire digital ecosystem. Therefore consistent NAP data, maintained directories, and active social profiles all count.
- [ ] 11. Consistent NAP data across all platforms
NAP stands for Name, Address, Phone. This data must be exactly identical on your website, Google Business, LinkedIn, industry directories, and all other platforms. The smallest discrepancies weaken your entity coherence.
How to check it: Google your company name and compare the contact details in the first 10 results. Do name, address, and phone number match everywhere? Correct any discrepancies immediately.
Why it matters: Inconsistent NAP data confuses AI models. If Google Business shows "Sample Street 5" and your website shows "Sample St. 5a", your trustworthiness in AI evaluation drops.
- [ ] 12. Industry directory listings up to date
Entries in relevant industry directories (e.g., ProvenExpert, Kununu, Trustpilot, WLW) are important authority signals. AI models use these sources to verify companies and identify sentiment.
How to check it: Search your company name in the top 5 industry directories for your sector. Are the listings current, complete, and with correct data? Also maintain categories and descriptions.
Why it matters: Industry directories are among the frequently used sources in training and real-time data. A current listing significantly increases your probability of being mentioned.
- [ ] 13. Google Business Profile optimized
Your Google Business Profile (GBP) is one of the most important sources for AI models — especially for local recommendations. It should be fully filled out, furnished with current photos, and regularly updated with posts.
How to check it: Log into your Google Business Profile. Is every field filled out? Are there current reviews? Have you published a GBP post in the last 30 days? Update opening hours seasonally.
Why it matters: For local queries like "Best tax advisor in Cologne", AI models — especially Gemini — draw heavily from Google Business Profiles and combine them with web signals.
- [ ] 14. LinkedIn profile built as thought leader
LinkedIn is one of the most heavily weighted sources in AI data. An active company profile and personal thought-leader profiles send strong authority signals and increase the chance of being mentioned by name.
How to check it: Do you or your management team post regularly on LinkedIn? Are there expert articles linked to your expertise? Do the sameAs links in the Organization Markup match your LinkedIn URLs?
Why it matters: AI models link people to companies and areas of expertise. An active LinkedIn profile with specialist content strengthens the association between your name and your field — and leverages the platform's enormous reach.
- [ ] 15. Reviews on relevant portals
Reviews on Google, ProvenExpert, Trustpilot, and industry-specific portals are strong trust signals. AI models actively use reviews to generate recommendations and compare alternatives.
How to check it: Do you have at least 10 reviews on Google? Is your average rating 4.0 or above? Are there recent reviews from the last 3 months? Respond promptly to feedback.
Why it matters: Businesses with many positive reviews are mentioned more often and more positively in AI responses than businesses without a review profile. Additionally, recent reviews increase perceived freshness.
How do you implement monitoring & optimization (points 16–20)?
GEO is not a one-time project. Your AI visibility is constantly changing — through model updates, competitor activities, and content changes. Systematic monitoring is therefore mandatory in order to react quickly to fluctuations.
- [ ] 16. GEO Score is measured regularly
The GEO Score measures how visible your company is in AI responses — combining mention rate, position, and sentiment. Without regular measurement, you optimize blindly and recognize trend breaks too late.
How to check it: Do you have a tool or process that measures your GEO Score at least weekly? Do you know your current score? Document changes and derive actions from them.
Why it matters: What you don't measure, you can't improve. A practical starting point is the GEO analysis at ai-geotracking.com.
What exactly goes into the GEO Score and how it is calculated is explained in the article GEO Score explained.
- [ ] 17. All 4 models are tracked (GPT-5, Gemini, Claude, Perplexity)
Each AI model makes recommendations differently. A company can have high visibility on Perplexity and significantly lower visibility on GPT-5. Anyone who tracks only one model has a distorted picture of their actual AI visibility.
How to check it: Do you measure your visibility across all four major models? Or only in ChatGPT? The difference can exceed 40 percentage points, especially for information-intensive prompts.
Why it matters: Perplexity uses real-time web search; GPT-5 relies more heavily on training data. A strategy that works for all models needs data from all models. How the individual models differ is explained in the article How AI models evaluate your brand.
Model tracking: Recommended designations
Use consistent model tags in your dashboards so that filters, alerts, and reports are unambiguous:
- GPT-5 — OpenAI GPT‑5 (ChatGPT)
- Gemini — Google Gemini (AI search, AI Overviews)
- Claude — Anthropic Claude
- Perplexity — Perplexity (with web citations)
- [ ] 18. Keyword set for GEO defined
Just as with SEO, you need a defined set of keywords and prompts for GEO that you want to be visible for. The difference: GEO keywords are often longer, conversational questions with a clear search intent.
How to check it: Have you defined a list of 20–50 prompts that potential customers send to AI models? Do you track your visibility for these prompts systematically — separately for GPT-5, Gemini, Claude, and Perplexity?
Why it matters: Without a defined keyword set, you measure randomly rather than strategically. Define prompts by purchase intent: "Which GEO tracking tools are there?" has more business value than "What is GEO?".
- [ ] 19. Monthly content refresh of top pages
Your most important pages need regular updates: new figures, current examples, fresh phrasing. Maintain modification dates (dateModified) in Schema so that models can recognize and correctly communicate freshness.
How to check it: When were your top 10 pages last updated with new content? Are all figures and facts still current? Are there outdated screenshots or examples? Therefore plan a fixed monthly update.
Why it matters: Outdated content is cited less and less by AI models over time. A monthly refresh keeps your AI Visibility stable and protects against loss of relevance.
- [ ] 20. Competitor monitoring for AI visibility
You need to know how your competitors perform in AI responses. If a competitor suddenly gets recommended at position 1 in ChatGPT, you need to react — ideally within days, not weeks.
How to check it: Do you track the AI visibility of your top 5 competitors? Do you know which prompts recommend your competition instead of you? Document countermeasures systematically.
Why it matters: AI visibility is often a zero-sum game. If an AI model recommends three companies and you are not among them, your competitors win. More on this in the article Why your competitors are being recommended.
How do GPT‑5, Gemini, Claude, and Perplexity differ concretely?
The models set different priorities — which is why GEO needs a cross-model strategy. Additionally, updates continuously change weightings, which is why continuous tracking is indispensable.
- GPT‑5 (GPT-5): Primarily uses training data and optionally plugins/browsing. Precise, strong at structured answers; benefits greatly from clear citations, consistent Schema, and "Answer-First" sections.
- Gemini (Gemini): Closely linked to the Google ecosystem and AI search. Draws data from the web, Google Business Profile, and structured data; prefers current, entity-clear content.
- Claude: Strength in safety and consistency checks; responds positively to clear source references and verifiable company data in Organization/FAQ/Article Markup.
- Perplexity: Relies on real-time web search with visible sources. Pages with clear answers, concise headings, FAQ sections, and clean permalinks are cited more frequently.
Action takeaway: Build a core strategy (Answer-First, Schema, FAQs, NAP) and extend it per model with specific tactics — for example GBP fine-tuning for Gemini or citation optimization for Perplexity.
How do you rate your GEO status?
Rate each of the 20 points on a scale of 0 to 5. The resulting scoring makes your GEO status measurable and prioritizable. This allows you to track progress transparently and identify bottlenecks early.
| Points | Meaning |
|---|---|
| 0 | Not present, not implemented |
| 1 | Rudimentarily present, but inadequate |
| 2 | Partially implemented, clear need for improvement |
| 3 | Solid foundation, but optimization potential exists |
| 4 | Well implemented, minor improvements possible |
| 5 | Exemplarily implemented, best practice |
Overall rating (maximum 100 points)
| Score | Rating | Recommended action |
|---|---|---|
| 0–25 | Critical | Your website is practically invisible to AI models. Start immediately with the quick wins. |
| 26–50 | Needs improvement | The fundamentals are partially missing. Focus on categories 1 and 2. |
| 51–70 | Solid | Good foundation in place. Optimize authority signals and begin systematic monitoring. |
| 71–85 | Strong | You are well positioned. Fine-tuning and competitor monitoring will deliver the next boost. |
| 86–100 | Excellent | Top performance. Maintain the level and extend your lead. |
Which GEO measures deliver immediate results?
You cannot implement everything at once. These five quick wins deliver the greatest ROI with the least effort. Start with them to build momentum, and then anchor processes for continuous improvement.
1. Create llms.txt (point 1) — 30 minutes of effort, immediate effect. A simple text file that tells AI models who you are. No developer needed. Guide here.
2. FAQ sections on top pages (point 8) — 2 hours of effort for 5 pages. Formulate 3–5 questions per page that customers would ask ChatGPT. This significantly increases your citability.
3. Implement FAQ Schema (point 3) — 1 hour with a WordPress plugin or 30 minutes manually. Makes your FAQs machine-readable. Implementation guide here.
4. Implement Answer-First structure (point 7) — Revise the first sentences of your top 10 pages. Key message first; examples and evidence follow immediately.
5. Optimize Google Business Profile (point 13) — 1 hour of effort. Fill in all fields, upload current photos, publish a first post. Repeat maintenance monthly.
More immediate measures can be found in our article 5 Quick Wins for AI Visibility.
How do you work through the audit step by step?
- Inventory: Check llms.txt, Schema, heading structure, internal links. Also document status and due dates.
- Prioritize: Select 5 top pages with business relevance. Start there for maximum impact.
- Answer-First: Rewrite the key messages for each page. Add unambiguous definitions and examples.
- FAQ build: Add 3–5 questions in users' own wording. Keep answers short, precise, and citable.
- Schema: Implement Organization/FAQ/Article cleanly. Validate with the Rich Results Test.
- Internal links: Connect top pages + thematic clusters. Use descriptive anchor texts.
- Authority: Update GBP, directories, LinkedIn. Consolidate NAP across all profiles.
- Monitoring: Set up GEO tracking across 4 models. Tag prompts by funnel stage.
- Refresh routine: Monthly update of top pages. Maintain
dateModifiedin markup. - Competition: Quarterly benchmarking and gap fixes. React specifically to visibility losses.
Which 12 questions should you track in your prompt set?
Use this template as a starting point for your GEO keyword set (point 18). Adapt the questions to your industry and target audience. Important: measure responses separately for GPT-5, Gemini, Claude, and Perplexity.
- Which GEO tracking tools are recommended in 2026?
- What is Generative Engine Optimization (GEO) and how do I get started?
- Best agency for GEO in [city]?
- How do I optimize llms.txt for ChatGPT and Gemini?
- GEO vs SEO: What is the difference?
- Which KPIs meaningfully measure AI Visibility?
- How do I create FAQ Schema for products?
- How do I improve my GEO Score in 30 days?
- What role does Google Business play for AI visibility?
- How do I improve citations in Perplexity?
- How do I test visibility in GPT-5?
- How does optimization differ for Gemini vs Claude?
Further reading:
Frequently asked questions (FAQ)
What is a GEO audit?
A GEO audit is a systematic review of your website for factors that influence your visibility in AI-generated responses. The 20 points in this checklist cover all relevant areas: technology, content, authority, and monitoring. This creates a prioritizable, measurable improvement plan.
How long does a complete GEO audit take?
With this 20-point checklist, you need 15–30 minutes for an initial inventory. For an in-depth analysis including competitor comparison, plan 2–3 hours. Implementation takes between one week (quick wins) and three months (full optimization) depending on your starting point.
What GEO Score is good?
A score of 51–70 points indicates a solid foundation. Over 70 points is above average, over 85 is excellent. More important than the absolute score is the trend: are you improving week by week? Details on score calculation can be found in the article GEO Score explained.
Can I carry out the GEO audit on my own?
The technical points (category 1) require basic web knowledge or a developer. Content and authority signals you can assess yourself. For automated monitoring, a tool like ai-geotracking.com is recommended, which automates measurement across all four AI models.
How does a GEO audit differ from an SEO audit?
An SEO audit checks Google ranking factors (load time, backlinks, keyword density). A GEO audit focuses on AI-specific factors such as llms.txt, Answer-First structure, and Schema Markup. Both complement each other. The detailed comparison can be found in the article Why your SEO tool is not enough.
How do I combine GEO and SEO without conflicting goals?
Prioritize information architecture, Schema, and clear answers. These help both traditional SERPs and AI search. You also prevent duplicate content issues by maintaining canonical URLs cleanly.
What role do GPT-5 and Gemini play in tracking?
With consistent model tags, you segment your visibility granularly. This allows you to recognize whether optimizations in GPT‑5 or Gemini take effect more quickly and where adjustments need to be made.
How do you measure your GEO Score with ai-geotracking.com?
The platform ai-geotracking.com monitors mentions, positions, and sentiment in ChatGPT, Gemini, Claude, and Perplexity. You can also compare prompts and competitor data. This creates end-to-end monitoring — from keyword definition to action plan.
- Dashboards: Overview of GEO Score trend, model splits, and top prompts.
- Alerting: Notifications on visibility losses by model or prompt.
- Guides: Step-by-step instructions for llms.txt, Schema, and FAQ strategies.
Ready for your GEO audit and monitoring?
Do you know what ChatGPT, Claude, and Gemini say about your company?
Most companies know their Google ranking — but not their AI Visibility. With ai-geotracking.com, you measure your GEO Score across all relevant AI models and receive concrete recommendations for action. This allows you to prioritize effectively and save resources.
Request your free GEO audit now and find out within 48 hours where your company stands — and what the next steps are.
Last updated: March 2026
Internal links:
- llms.txt: The new robots.txt for AI models
- Structured Data for Generative AI
- GEO Score explained: How to measure your AI visibility
- How AI models evaluate your brand
- Why your SEO tool is not enough
- Why your competitors are being recommended
- Homepage
About the author
GEO Tracking AI Team
The team behind GEO Tracking AI builds tools that help businesses measure and optimize their visibility across AI models like ChatGPT, Claude, and Gemini.
Related Articles
5 Quick Wins for Immediately Better AI Visibility
Discover 5 practical measures to measurably improve your AI visibility in ChatGPT, Perplexity, and Gemini within one week — from FAQ pages and Schema Markup to llms.txt.
GEO Score Doubled in 30 Days – AI Visibility Case Study
GEO Case Study: GEO Score from 24% to 48% in 30 days. With llms.txt, Q&A optimization and structured data – incl. timeline, data and checklists.
AI Visibility 2026: GEO, llms.txt & the Hidden Costs
Boost AI visibility with GEO, llms.txt, and structured data. Learn how to win GPT-5/Gemini citations, grow AI share of voice, and improve ChatGPT rankings.