ClerkMinutes Audit Dashboard

GEO + content audit · Sept 29, 2026 · GSC data: Sept 29, 2025 – Sept 29, 2026

clerkminutes.com → Performance dashboard
Total clicks (12mo)
—
Total impressions (12mo)
—
Device split
Desktop 81% · Mobile 17% · Tablet 1%
3,923 / 806 / 28 clicks
US share of clicks
84%
3,925 of 4,757 total
Overall SEO Score
65 /100
Your Grade
C
Good Performance
Solid foundation — a few fixes will boost you further.
29
Passed
All good!
6
Warnings
Need attention
1
Failed
Critical issues
5
Info
Helpful insights
Score Breakdown
100%
Technical SEO
71%
On-Page SEO
50%
Performance
100%
Mobile SEO
88%
Security
67%
User Experience
ClerkMinutes · Q4 2026

Strategic SEO Recommendations

Six prioritized moves for organic and AI search growth. Technical hygiene addressed separately below.

Priority 1
Build intent-led topic clusters

Complete core municipal-workflow pillar pages first, then organize supporting content around them through deliberate internal linking. Focus authority around a small number of commercially relevant subjects rather than disconnected blog production. Content should not just be landing pages or blogs, but also include free lead-gen tools.

Q4 action
  • Identify the core municipal-workflow topics and complete a pillar page for each
  • Map existing blog posts and product pages to a pillar or flag as standalone
  • Add deliberate internal links from cluster pages to each pillar
  • Identify and scope one free lead-gen tool per cluster
Success signal

Pillar pages are live and cluster pages link to them. GSC shows impressions consolidating around pillar URLs. At least one lead-gen tool is scoped or shipped.

Priority 2
Capture existing search demand

Use GSC to prioritize high-impression pages and queries with suppressed CTR or rankings within striking distance of stronger first-page visibility. Exclude anomalous/bot-driven impression patterns from prioritization.

Q4 action
  • Export GSC queries with high impressions and suppressed CTR; exclude the Feb–May 2026 anomaly window
  • Shortlist pages ranking positions 5–20 with genuine buyer-intent queries
  • Update title tags and meta descriptions on shortlisted pages to match query intent
Success signal

CTR improves on targeted pages in GSC. Clicks increase on shortlisted queries without proportional impression growth.

Priority 3
Expand state-level search coverage

Create jurisdiction-specific pages around open-meeting requirements, meeting-minutes workflows and municipal clerk needs. Prioritize states using customer concentration, regulatory specificity and population rather than blindly creating all 50.

Q4 action
  • Score states by customer concentration, regulatory specificity, and population to identify first-priority targets
  • Build a repeatable page template covering open-meeting law, minutes requirements, and ClerkMinutes fit
  • Ship pilot pages for the top 3–5 states before committing to broader build-out
Success signal

Pilot state pages are indexed. GSC shows impressions on jurisdiction-specific queries. At least one page earns first-page ranking within the quarter.

Priority 4
Optimize priority content for answer retrieval

Improve commercially and informationally important pages with concise answer-first passages, descriptive headings, self-contained factual sections, relevant FAQs and appropriate structured data. Do not mechanically turn every heading into a question.

Q4 action
  • Add answer-first passages to the top of priority commercial and informational pages
  • Add FAQs using exact user question language; wrap in FAQPage schema
  • Add Article + Author schema to blog content (currently 1 schema block site-wide)
  • Review heading structure on priority pages — improve descriptiveness without converting all headings to questions
Success signal

Priority pages contain answer-first passages and FAQs. Schema validates cleanly. Manual LLM testing shows ClerkMinutes appearing in answers for target queries where it currently returns nothing.

Priority 5
Strengthen ClerkMinutes entity signals

Standardize product positioning, features, integrations, pricing, HeyGov relationship and target audience across authoritative pages. Reinforce these facts with appropriate structured data, named expertise, citations and proprietary evidence where available.

Q4 action
  • Audit product description consistency across homepage, press page, and authoritative third-party listings
  • Add Organization schema site-wide with sameAs links to LinkedIn, Crunchbase, and HeyGov
  • Add Person schema for any named authors or experts cited in content
  • Identify proprietary data or customer evidence that can support factual claims on key pages
Success signal

Product description is consistent on-site and off-site. Organization schema is live and validates. LLM responses about municipal meeting-minutes software include accurate ClerkMinutes descriptions.

Priority 6
Measure GEO against business outcomes

Use repeatable ChatGPT/Claude/Gemini/Perplexity testing as a diagnostic baseline, not the primary KPI. Measure impact through branded GSC activity, qualified organic visibility, self-reported signup attribution and trial volume.

Q4 action
  • Establish a fixed set of queries for monthly LLM testing across ChatGPT, Claude, Gemini, and Perplexity
  • Add a self-reported "how did you hear about us?" field to trial signup if not already present
  • Use branded GSC query volume as the primary GEO signal, with Sept 2025 as baseline
  • At 90-day review: evaluate trial volume, self-reported organic attribution, and qualified organic clicks — not citation count
Success signal

Branded GSC query volume trends up from Sept 2025 baseline. Self-reported organic attribution is trackable. Trial volume correlates with organic click growth.

Technical hygiene — supporting remediation
Tighten homepage title and meta description
Review page weight and asset compression
Investigate excessive markup/text-to-HTML ratio without treating the automated score itself as a KPI
Resolve remaining accessibility/alt-text issue
Review CSP recommendation with development
Note on the impression spike: Feb–May 2026 shows a massive impression surge (up to 39k/day) driven by automated city-specific queries ("city of X council meeting minutes software applications"). These are likely sales intel bot traffic, not real buyer intent. Clicks did not increase proportionally. Real search performance is better read from Jul–Jan and Jun windows.

Daily clicks — 12 months

Daily impressions — 12 months

Avg. position — 12 months (lower = better)

Top countries by clicks

All traffic is the product's target audience (US municipal government) plus organic international. Non-US clicks are almost all informational, not commercial.

Country Clicks Impressions CTR Position

What the international traffic means

UK (97 clicks), Canada (90), Philippines (71): Not buyers. Philippines and UK traffic is likely people in local government admin roles looking for general meeting minutes tooling. They will not convert on a US municipal product.

Serbia (68 clicks, 17% CTR): HeyGov is based in Serbia. This is team/internal traffic.

Netherlands + Germany (18k impressions, near-zero clicks): Part of the bot query traffic. Automated searches triggering impressions with no real intent.

Action: No content changes needed for international traffic. The GEO work targets US government search and LLM queries specifically.

US dominance confirms product-market fit signal

84% of all clicks come from the US. The product is US municipal-only. This ratio is healthy. Non-US impressions are inflating the total impression count without contributing to pipeline.

LLM presence baseline

Run each query in ChatGPT (GPT-4o), Claude (Sonnet), Gemini, and Perplexity. Document whether ClerkMinutes appears and whether the description is accurate. This is the before-state for the Q4 GEO plan.

Status column: ✓ Appears accurately · ~ Appears inaccurately · ✗ Not mentioned · ? Not yet checked

⚠ Prompt tracking is not the primary measurement signal

Most LLM-influenced traffic arrives as direct, not referral. GA4 undercounts LLM impact. Prompt-based tools (Profound, Peec, etc.) automate the same flawed methodology. Use the query table below for diagnostic audits — not as a KPI.

Real measurement approach:
  • Brand search in GSC — branded query impressions + clicks tick up when LLMs mention ClerkMinutes. Pull start-of-Q4 vs. end-of-Q4. This is the proxy that holds up in a review.
  • Self-reported attribution at trial signup — add "how did you hear about us?" with AI tools as explicit options (ChatGPT, Claude, Gemini, Perplexity, other AI tool). One field. Captures the dark funnel path GA4 can't see.
  • Trial volume is the scorecard — LLM presence is the mechanism, not the metric. Report trial starts. The content infrastructure is the lever.

Branded search baseline (GSC)

Export from GSC: Performance → filter Queries "containing" → clerkminutes → Date range: Aug 1 – today → Export CSV. Drop the file below.

Drop GSC CSV here or click to upload
Query ChatGPT (GPT-4o) Claude (Sonnet) Gemini Perplexity Notes / accuracy issues

What to look for in each response

Accuracy flags to document:
  • Wrong pricing (current: $99/mo Starter, $249/mo Growth)
  • Missing features (AI Actions, historical search, FOIA export, Meetings Portal)
  • Wrong integration support (Zoom + Teams imports supported)
  • Wrong ICP (described as generic AI tool vs. municipal government specific)
  • Competitor mentioned in its place (Otter, Fathom, BoardBreeze)
  • HeyGov not mentioned as parent company
Queries most likely to surface ClerkMinutes:
  • "How does ClerkMinutes work" — direct brand query
  • "What is ClerkMinutes used for" — direct brand query
  • Queries most likely to miss it: Zoom integration, Teams integration, transcription-specific
Re-run this baseline after Q4 content ships:

Same 8 queries, same 4 tools, December 2026. Use as directional context alongside GSC brand search volume and trial signup attribution data — not as standalone proof of impact.

Hub and spoke tracker

Each hub needs a live, well-structured pillar page before spokes can properly reference it.

State page status

Priority states for programmatic SEO and LLM citation. Selection criteria: customer concentration + open meeting law specificity + population weight.

State Status Open meeting law Notes

Content quality audit — AI-sounding flags

Posts that need review for generic/non-specific content. LLMs deprioritize vague content. So do clerks.

Page Impressions Pos Quality flag Action
On-Page SEO
WARN Title Tag
Meeting Minutes Software ! AI Tool for Municipal Clerks - ...
Title too long (72 chars). May be truncated in search results.
WARN Meta Description
Streamline municipal operations with ClerkMinutes, the AI-po...
Too long (181 chars). Will be truncated in search results.
PASS H1 Heading
AI-Powered Meeting Minutes for Municipal Clerks. Finish your
Single H1 tag found. Perfect for SEO structure.
PASS H2 Headings
6 found
Examples: Watch ClerkMinutes in Action, How ClerkMinutes® Works...
PASS H3 Headings
12 found
Examples: Upload Your Agenda, Upload the Meeting Recording...
PASS H4 Headings
1 found
Examples: Download free guide on how to best do yo
PASS Canonical URL
https://clerkminutes.com/
Canonical URL is properly configured.
INFO Meta Keywords
Not set
Meta keywords are ignored by Google but used by some minor search engines.
Technical SEO
PASS Character Encoding
UTF-8 [OK]
Correct character encoding set.
PASS Robots Meta Tag
max-image-preview:large
Robots meta tag configured properly.
INFO Hreflang Tags
Not found
If you have multi-language content, add hreflang tags for international SEO.
PASS Robots.txt
Found & accessible [OK] -- https://clerkminutes.com/robots.txt
Robots.txt exists and does not block your site. Sitemap reference found inside.
PASS XML Sitemap
Found [OK] (80 URLs) -- https://clerkminutes.com/sitemap.xml
XML Sitemap exists and is accessible to search engines. URL: https://clerkminutes.com/sitemap.xml
INFO Server Technology
cloudflare
Server information is exposed. Consider hiding for security.
Content Quality
PASS Word Count
1354 words
Good word count for SEO rankings.
PASS Internal Links
12 links
Good internal linking structure (12 internal, 11 external).
INFO External Links
11 links
External links to authoritative sources can improve credibility.
FAIL Text-to-HTML Ratio
1.6%
Very low text-to-HTML ratio. Too much code vs actual content.
Performance
PASS Server Response Time
228ms
Excellent! Server responds very quickly.
WARN Page Size
586.7 KB
Page is getting large. Optimize images and minify CSS/JS.
WARN Compression (GZIP/Brotli)
Not enabled
Enable GZIP or Brotli compression to reduce page transfer size.
PASS Browser Caching
Configured [OK]
Cache headers are set: public, max-age=0, must-revalidate
Security
PASS HTTPS / SSL
Enabled [OK]
Site uses secure HTTPS protocol. Great for SEO and user trust.
PASS X-Frame-Options
Present [OK]
Clickjacking protection enabled.
PASS X-Content-Type-Options
Present [OK]
MIME-type sniffing protection enabled.
PASS HSTS Header
Present [OK]
HSTS is configured. Forces HTTPS connections on repeat visits.
WARN Content Security Policy
Missing
No Content Security Policy detected. Add CSP to control which resources browsers can load.
PASS Referrer-Policy
Present [OK]
Referrer-Policy set. Controls what referrer data gets shared with other sites.
PASS Permissions-Policy
Present [OK]
Permissions-Policy set. Controls browser feature access like camera, mic, and geolocation.
PASS Mixed Content
None detected [OK]
No mixed content detected. All resources load over HTTPS.
INFO Cookie Security
No cookies set
No Set-Cookie header detected on this page. If your site uses cookies elsewhere, make sure they have Secure, HttpOnly, and SameSite flags.
Social & Rich Data
PASS Open Graph Tags
5/5 tags configured [OK]
Open Graph is well configured: og:title, og:description, og:image, og:url, og:type
PASS Twitter Card Tags
Card type: summary_large_image [OK]
Twitter Card properly configured with 4/4 tags.
PASS Schema.org Markup
1 schema block(s) [OK]
Structured data found. This can generate rich snippets in search results.
PASS Favicon
Present [OK]
Favicon is configured: /assets/framerusercontent.com/images/vfp4R2X8bNkLxifgmcsFYx6
Accessibility
WARN Image Alt Text
257/258 have alt
1 image(s) missing alt text. Fix for accessibility & SEO.
PASS ARIA Attributes
5 elements [OK]
ARIA attributes found. Good for screen reader accessibility.
PASS HTML Lang Attribute
lang="en" [OK]
Language attribute is set. Helps search engines and screen readers.
Mobile Optimization
PASS Viewport Meta Tag
Configured [OK]
Viewport properly configured for mobile devices.
PASS Mobile-Friendly
Likely mobile-friendly [OK]
Viewport meta tag is set correctly, indicating mobile optimization.
PASS Tap Targets
73 interactive elements
Interactive elements found. Ensure they are at least 44x44px for mobile usability.
Priority Action Plan
1. CRITICAL -- Text-to-HTML Ratio
Very low text-to-HTML ratio. Too much code vs actual content.
2. WARNING -- Title Tag
Title too long (72 chars). May be truncated in search results.
3. WARNING -- Meta Description
Too long (181 chars). Will be truncated in search results.
4. WARNING -- Page Size
Page is getting large. Optimize images and minify CSS/JS.
5. WARNING -- Compression (GZIP/Brotli)
Enable GZIP or Brotli compression to reduce page transfer size.
6. WARNING -- Content Security Policy
No Content Security Policy detected. Add CSP to control which resources browsers can load.
7. WARNING -- Image Alt Text
1 image(s) missing alt text. Fix for accessibility & SEO.
AEO -- Answer Engine Optimization
Get cited by ChatGPT, Google AI Overviews, Gemini, and Perplexity
WARNING Title Tag

What to do: Rewrite as a direct question or definitive answer: "How to Fix X" / "What Is X: Complete Guide 2025". Put the primary keyword in the first 3 words. Keep under 60 chars.

STEP-BY-STEP:

  • Open your CMS and find the page title field.
  • Rephrase to start with the main keyword.
  • Add a power word: Complete, Ultimate, Proven, Step-by-Step.
  • Test at seotesteronline.com/serp-snippet-generator.

WHY IT MATTERS

AI answer engines use the title tag as the primary topic signal. Question-format titles match how users phrase AI queries. Pages with question-style H1s appear in AI Overviews 2.3x more often.

AI Platforms: Google AI Overviews · ChatGPT · Perplexity · Bing Copilot

COMPETITOR INTEL

Top-cited sites on Perplexity lead with "What Is", "How To", or "Best [Year]" titles -- not brand names.

WARNING Meta Description

What to do: Write a 145-155 char meta description as a standalone factual answer -- not a teaser. Start with the answer, not "Learn how to...". Include one specific number or outcome.

STEP-BY-STEP:

  • Identify the #1 question your page answers.
  • Write the answer in the first sentence (80 chars).
  • Add context or a specific stat in sentence 2.
  • Avoid "Click to learn" or marketing fluff.

WHY IT MATTERS

Perplexity and ChatGPT web browsing extract meta descriptions as content previews. Fact-dense descriptions with specific data get cited up to 3x more. AI Overviews pull descriptions verbatim as answer snippets.

AI Platforms: Google AI Overviews · Perplexity · ChatGPT web · Bing Copilot

COMPETITOR INTEL

Bank of America (32.2% AI visibility in banking) uses factual, stat-heavy descriptions -- never marketing copy.

WARNING Image Alt Text

What to do: Rewrite all alt text as descriptive sentences. For data: "Chart showing 35% conversion increase after implementing FAQ schema". For screenshots: describe what is happening on screen.

STEP-BY-STEP:

  • Run an audit with Screaming Frog to find all empty/short alt tags.
  • For every image: write alt text as a sentence describing what it shows.
  • For charts/graphs: include the key insight in the alt text.
  • For product images: include product name, model, color.

WHY IT MATTERS

Multimodal AI systems (Gemini, GPT-4o) process alt text during indexing to understand images and connect them to surrounding content. Descriptive alt text creates machine-readable data that LLMs extract and cite.

AI Platforms: Google AI Overviews · Gemini · GPT-4o -- Multimodal indexing

COMPETITOR INTEL

Top-cited sites implement this as baseline.

BEST PRACTICE Schema.org Markup

What to do: Add FAQPage schema for any Q&A content. Add Article + Author schema with credentials. Add HowTo schema for guides. Use JSON-LD format. Validate before publishing.

STEP-BY-STEP:

  • Install Rank Math or Yoast for WordPress (or add JSON-LD manually).
  • Add FAQPage schema to every FAQ section -- list the exact Q&A pairs.
  • Add Article schema with author name, credentials, dateModified.
  • Add HowTo schema for any process/guide pages.

WHY IT MATTERS

Schema is the #1 highest-ROI AEO change. AI systems use FAQPage to extract Q&A pairs directly, Article+Author for E-E-A-T verification, and HowTo for procedural queries. Early adopters appear in AI answers 10x faster.

AI Platforms: Google AI Overviews · Bing AI · All LLM indexing pipelines

COMPETITOR INTEL

Forbes, Investopedia, and Healthline -- the top three most-cited domains across AI platforms -- all have FAQPage + Article + Author schema on every page.

BEST PRACTICE HTTPS / SSL

What to do: Migrate to HTTPS immediately. Get SSL from Let's Encrypt (free). Force HTTPS via .htaccess. Update all internal links. Implement HSTS header. Check for mixed content.

STEP-BY-STEP:

  • Get SSL cert via Let's Encrypt or your host (free).
  • Add redirect in .htaccess: RewriteRule ^ https://{HTTP_HOST}%{REQUEST_URI} [L,R=301].
  • Update all hardcoded http:// links in content.
  • Add Strict-Transport-Security header.

WHY IT MATTERS

HTTPS is an absolute requirement for any AI citation. LLM training datasets are overwhelmingly HTTPS. Perplexity, Bing Copilot, and Google's AI systems filter HTTP pages. No exceptions.

AI Platforms: All AI platforms -- Universal non-negotiable requirement

COMPETITOR INTEL

Zero of the top 1,000 most-cited domains on any AI platform are HTTP.

BEST PRACTICE FAQ Content Sections

What to do: Add a dedicated FAQ section to every key page with 8-12 Q&As using the exact questions users type into AI. Wrap in FAQPage schema.

STEP-BY-STEP:

  • Use Google's "People Also Ask" and AnswerThePublic to find real questions.
  • Add FAQ section at the bottom of each key page.
  • Use question as H3, answer as paragraph (50-80 words).
  • Add FAQPage JSON-LD schema matching every Q&A.

WHY IT MATTERS

FAQ sections are the single most-cited content format in AI answers. AI engines extract individual Q&A pairs and use them as direct responses. Pages with FAQ schema are 5x more likely to appear in AI Overviews.

AI Platforms: Google AI Overviews · ChatGPT · Perplexity · Claude

COMPETITOR INTEL

NerdWallet adds 10-15 FAQs to every page -- this is the primary driver of their dominance in financial AI citations across ChatGPT, Perplexity, and Google.

BEST PRACTICE H1 Heading

What to do: Set your H1 as the exact question your target user would type into ChatGPT. Format: "How to [Action] [Topic]" or "[Topic]: The Complete [Year] Guide". Max 70 chars.

STEP-BY-STEP:

  • Search your target topic in Perplexity -- note how the AI re-states the question.
  • Use that phrasing as your H1.
  • Put the primary keyword in the first 4 words.
  • Ensure only ONE H1 per page.

WHY IT MATTERS

RAG (Retrieval-Augmented Generation) systems chunk documents at heading boundaries. The H1 becomes the AI's entire understanding of your page topic. Mismatched H1s cause AI engines to mis-file your content.

AI Platforms: ChatGPT · Claude · Gemini · Perplexity · Google AI Overviews

COMPETITOR INTEL

Every top Perplexity citation uses an H1 that is either a question or includes the year. Zero top citations use brand names as H1.

BEST PRACTICE Answer-First Structure

What to do: Add a "Quick Answer" block at the very top of each page -- 2-3 sentences that directly answer the page's primary question. Use a visually distinct styled box.

STEP-BY-STEP:

  • Identify the #1 question your page answers.
  • Add a styled box at the top labeled "Quick Answer" or "TL;DR".
  • Write 40-60 words directly answering the question.
  • Keep language simple and factual -- no marketing tone.

WHY IT MATTERS

AI answer engines extract the most direct, upfront answer. The inverted pyramid structure (answer -> evidence -> context) is the format AI systems are trained to recognise. Pages with this structure appear in featured snippets and AI Overviews.

AI Platforms: Google AI Overviews · Perplexity · ChatGPT

COMPETITOR INTEL

Every top Healthline, Mayo Clinic, and WebMD article opens with a boxed quick answer -- this is a core reason they lead medical AI citations.

BEST PRACTICE Original Data & Statistics

What to do: Include 3-5 proprietary statistics, original research findings, or unique data points in every major page. Cite all external sources with author + publication + date.

STEP-BY-STEP:

  • Conduct a survey, analyse your own data, or compile industry data.
  • Add a "Key Statistics" call-out box near the top of key pages.
  • Create a shareable infographic from your data.
  • Format: "X% of [audience] say [finding] -- [Your Brand] Survey, [Year]".

WHY IT MATTERS

LLMs heavily favour content with original data. Proprietary statistics get cited across the web -- each citation strengthens your entity authority. Conductor's own AEO report became widely cited precisely because it contains first-of-its-kind data.

AI Platforms: ChatGPT · Perplexity · Claude -- Factual citation preference

COMPETITOR INTEL

Semrush, Ahrefs, and Hubspot all produce annual research reports specifically to generate AI-cited statistics -- their proprietary data is the #1 source of their AI visibility.

BEST PRACTICE H2-H4 Headings

What to do: Restructure ALL H2/H3 subheadings as standalone questions. Each section must independently answer its own heading in 75-200 words. Use "What", "How", "Why", "When" openers.

STEP-BY-STEP:

  • Export a list of your current H2/H3 tags.
  • Convert each to a Q-format: "How does X work?" "What is Y?" "Why should you Z?"
  • Write a 75-word direct answer directly beneath each heading.
  • Add a table or bullet list to each major section.

WHY IT MATTERS

AI citation systems extract content in chunks bounded by headings. Each chunk competes independently. Self-contained answer sections are 4x more likely to be used as AI citations than continuous prose.

AI Platforms: Google AI Overviews · ChatGPT · Perplexity · Claude

COMPETITOR INTEL

Investopedia -- one of the most AI-cited domains -- uses a question H2 + 2-sentence answer + expandable detail for every single section.

BEST PRACTICE AI Bot Access (robots.txt)

What to do: Verify your robots.txt allows: GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, Google-Extended, Gemini, BingBot. Blocking these removes you from AI training and citations.

STEP-BY-STEP:

  • Visit yourdomain.com/robots.txt.
  • Search for "GPTBot", "ClaudeBot", "PerplexityBot", "Google-Extended".
  • If any are in Disallow: rules, remove them.
  • Optionally add explicit Allow: rules for each bot.

WHY IT MATTERS

Blocking AI crawlers prevents your content from ever being indexed or cited. GPTBot and ClaudeBot are opt-out by default. Many sites accidentally block them via overly-broad wildcard Disallow rules. One line in robots.txt can exclude you from all AI citations.

AI Platforms: ChatGPT · Claude · Perplexity · Google AI -- Indexing access

COMPETITOR INTEL

All top AI-cited domains have confirmed GPTBot and ClaudeBot access. Perplexity has published guidance that blocking their bot removes you from all Perplexity citations permanently.

BEST PRACTICE Author E-E-A-T Signals

What to do: Add a named author bio with credentials, LinkedIn URL, and a dedicated author page to every content piece. Add Person schema with jobTitle, knowsAbout, and sameAs.

STEP-BY-STEP:

  • Create an author page for each writer: name, headshot, credentials, bio.
  • Link every article byline to the author's page.
  • Add Person schema with: name, jobTitle, knowsAbout, sameAs (LinkedIn URL).
  • Get authors to publish or be quoted on external authoritative sites.

WHY IT MATTERS

AI systems use E-E-A-T (Experience, Expertise, Authoritativeness, Trust) signals. Named experts with verifiable credentials are cited 4x more than anonymous content. Profound's data shows author attribution is a major citation differentiator.

AI Platforms: Google AI Overviews · Bing AI · Perplexity -- E-E-A-T verification

COMPETITOR INTEL

Investopedia requires every article to be reviewed by a credentialed expert -- their "Reviewed By" badge is a key reason they top AI finance citations.

BEST PRACTICE Word Count

What to do: Add a "Quick Answer" box (40-60 words) at the very top. Then expand to 1,500-2,500 words below it. Use the inverted pyramid structure: answer first, context second, detail third.

STEP-BY-STEP:

  • Add a styled box labeled "Quick Answer:" or "TL;DR:" at the top of every key page.
  • Write 2-3 sentences directly answering the main topic.
  • Then structure the rest of the page in sections below.
  • Target 1,500+ words for competitive topics.

WHY IT MATTERS

AI engines extract the most concise upfront answer. Pages with an answer-first block get featured in AI Overviews 3x more. Long-form depth below ensures you also rank for detailed queries.

AI Platforms: ChatGPT · Perplexity · Google AI Overviews

COMPETITOR INTEL

Every NerdWallet and Bankrate page starts with a "Quick Answer" or highlighted summary box -- this is why they dominate AI financial citations.

BEST PRACTICE Canonical URL

What to do: Set a self-referencing canonical on every page pointing to the HTTPS version with consistent trailing-slash policy. Redirect all variants (HTTP, www/non-www, trailing slash) to the canonical.

STEP-BY-STEP:

  • Add <link rel="canonical" href="https://yourdomain.com/page/"> to every page.
  • Choose trailing-slash or no-trailing-slash -- be 100% consistent.
  • Set up 301 redirects for all non-canonical variants.
  • Verify with Screaming Frog or Ahrefs.

WHY IT MATTERS

AI crawlers consolidate ranking signals at the canonical URL. Without canonicals, your content authority splits across duplicate versions -- reducing the probability that any single URL gets selected for AI citations.

AI Platforms: Google AI Overviews · Bing AI · All AI crawlers

COMPETITOR INTEL

Top-cited sites implement this as baseline.

BEST PRACTICE Open Graph Tags

What to do: Set og:description as a factual standalone summary (not a teaser). Add og:type="article", og:article:author, og:article:published_time, og:article:modified_time.

STEP-BY-STEP:

  • Add og:description with your factual answer summary.
  • Set og:type="article" for content pages.
  • Add og:article:author linking to an author profile URL.
  • Add og:article:published_time and og:article:modified_time.

WHY IT MATTERS

Perplexity and ChatGPT web browsing parse OG metadata as quick content abstracts. Author attribution builds E-E-A-T. Modified time signals freshness -- AI engines prefer recently updated content.

AI Platforms: Perplexity · ChatGPT web · Bing Copilot

COMPETITOR INTEL

Perplexity's most-cited sources all have complete og:article metadata including author, published date, and modification timestamp.

BEST PRACTICE Internal Links

What to do: Build a topic cluster: one pillar page + 6-12 supporting cluster pages. All cluster pages link back to the pillar with keyword-rich anchor text. Pillar links out to all clusters.

STEP-BY-STEP:

  • Identify your 3 most important topic areas.
  • Create a comprehensive pillar page for each (2,000+ words).
  • Create 6+ cluster pages covering related subtopics.
  • Link every cluster page to the pillar using anchor text = pillar's primary keyword.

WHY IT MATTERS

LLMs map topical authority via link graphs. Dense, interconnected clusters signal domain expertise. Brands with strong topic clusters are cited 3-5x more often than brands with isolated pages.

AI Platforms: ChatGPT · Gemini · Google AI Overviews -- Topical authority signal

COMPETITOR INTEL

Hubspot's topic cluster model -- widely credited as the benchmark -- helped them achieve top AI citations for marketing, CRM, and sales topics across all major AI platforms.

BEST PRACTICE HTML Lang Attribute

What to do: Add lang="en" (or your locale) to the <html> tag. For multilingual sites: add hreflang tags for each language/region. Submit hreflang sitemap to Search Console.

STEP-BY-STEP:

  • Add lang="en" to your <html> opening tag.
  • For multilingual: add <link rel="alternate" hreflang="en-US" href="..."> in <head>.
  • Add hreflang for every language variant.
  • Generate and submit an hreflang sitemap.

WHY IT MATTERS

Language metadata helps AI engines correctly route queries to your language version. Incorrectly declared language causes AI systems to deprioritise your content for correctly-matched queries.

AI Platforms: Google AI Overviews · Gemini -- Language routing

COMPETITOR INTEL

Top-cited sites implement this as baseline.

GEO -- Generative Engine Optimization
Build LLM citation readiness across ChatGPT, Claude, Gemini & Perplexity
WARNING Meta Description

What to do: Write meta descriptions as factual citation previews -- imagine how Perplexity shows your snippet. Include specific numbers, dates, or verifiable outcomes.

STEP-BY-STEP:

  • Identify what specific fact, stat, or answer your page provides.
  • State that fact in the first sentence (80 chars).
  • Add supporting detail or benefit in sentence 2.
  • Include one specific number or outcome if possible.

WHY IT MATTERS

GenAI retrieval pipelines use meta descriptions as compressed content summaries. Fact-dense descriptions with specific data are cited 2-3x more often. Vague or promotional descriptions are rarely surfaced in AI answers.

AI Platforms: Perplexity · ChatGPT web · Bing Copilot -- Content preview

COMPETITOR INTEL

Top-cited sites implement this as baseline.

WARNING Image Alt Text

What to do: Write alt text as data-extractable sentences. For charts: "[Chart type] showing [key metric]: [specific finding with number]". For infographics: list the top 3 data points.

STEP-BY-STEP:

  • List every image showing data, processes, or results.
  • For charts: "Bar chart showing 40% increase in AI visibility after FAQ schema implementation".
  • For screenshots: describe what action is being performed.
  • For infographics: describe the top 3 data points in the alt text.

WHY IT MATTERS

Multimodal LLMs (Gemini, GPT-4o) index alt text as machine-readable data. Images with descriptive alt text create AI-extractable facts. Charts with descriptive alt text get their statistics cited in LLM responses.

AI Platforms: Gemini · GPT-4o · Google AI Overviews -- Multimodal indexing

COMPETITOR INTEL

Top-cited sites implement this as baseline.

BEST PRACTICE Word Count

What to do: Create 2,000-4,000 word cornerstone content with original data, expert quotes, cited sources, and key stats. Structure: intro -> answer -> evidence -> context -> conclusion.

STEP-BY-STEP:

  • Identify your 5 highest-traffic pages and expand each to 2,000+ words.
  • Add at least 3 original statistics or proprietary findings.
  • Include quotes from named experts with credentials.
  • Add a cited reference list at the bottom.

WHY IT MATTERS

LLMs (ChatGPT, Claude, Gemini) are trained on comprehensive, well-sourced content. Long-form pages with citations are included in AI training data 3x more often. Conductor's 2025 data shows depth is the #1 GEO factor.

AI Platforms: ChatGPT · Claude · Gemini · Perplexity -- Training data inclusion

COMPETITOR INTEL

Wikipedia -- the single most-cited source across all LLMs -- uses exhaustive citations, structured sections, and regular updates. Model your content structure after Wikipedia articles.

BEST PRACTICE Schema.org Markup

What to do: Add Organization schema with sameAs links to Wikipedia, LinkedIn, Crunchbase, Twitter/X. Add Person schema for all authors with jobTitle, knowsAbout, and sameAs to LinkedIn.

STEP-BY-STEP:

  • Create Organization JSON-LD: name, url, logo, description, sameAs array.
  • In sameAs, include: Wikipedia URL, LinkedIn Company, Crunchbase, Twitter/X, Instagram.
  • Create Person schema for each author: name, jobTitle, knowsAbout topics, sameAs LinkedIn.
  • Nest Author in every Article schema.

WHY IT MATTERS

LLMs build entity knowledge graphs from sameAs links. The more authoritative sites that confirm your entity, the stronger your brand becomes in the AI's "memory". Strong entities are cited 5x more than unverified brands.

AI Platforms: ChatGPT · Gemini · Claude -- Entity recognition & knowledge graph

COMPETITOR INTEL

Every major brand cited by ChatGPT has a verified entity graph: Wikipedia page, Crunchbase profile, LinkedIn Company, and sameAs schema linking them all.

BEST PRACTICE HTTPS / SSL

What to do: HTTPS mandatory. Implement HSTS header. Ensure certificate from trusted CA. Fix mixed content. Target A+ on SSL Labs. These are table-stakes for any LLM citation.

STEP-BY-STEP:

  • Enable HTTPS (Let's Encrypt is free).
  • Add to server config: Strict-Transport-Security: max-age=31536000; includeSubDomains.
  • Fix all mixed content warnings (use Chrome DevTools -> Security tab).
  • Test at ssllabs.com -- aim for A rating.

WHY IT MATTERS

LLM training datasets are overwhelmingly HTTPS. Conductor's benchmark of 3.3 billion sessions found zero HTTP domains in top AI citations across all 10 industries measured. HTTPS is mandatory, not optional.

AI Platforms: All platforms -- Absolute baseline requirement

COMPETITOR INTEL

Top-cited sites implement this as baseline.

BEST PRACTICE Cross-Web Entity Mentions

What to do: Get your brand mentioned and linked from: Wikipedia, Crunchbase, LinkedIn Company, industry directories, and 10+ authoritative publications in your niche.

STEP-BY-STEP:

  • Check if your brand has a Wikipedia page -- create one if eligible.
  • Claim/update your Crunchbase profile with full details.
  • Ensure your LinkedIn Company page is complete and active.
  • Get listed on 10+ industry-specific directories.

WHY IT MATTERS

LLMs build entity knowledge graphs from cross-web mentions. The more authoritative sites that mention and link to you, the stronger and more recognizable your entity becomes. Strong entities are cited 5x more often than brands with sparse mentions.

AI Platforms: ChatGPT · Gemini · Claude -- Entity knowledge graph

COMPETITOR INTEL

Relixir, Profound, and AthenaHQ -- the most AI-cited GEO tools -- all have Wikipedia citations, Crunchbase profiles, LinkedIn presence, and 100+ press mentions. Entity strength drives citation frequency.

BEST PRACTICE H2-H4 Headings

What to do: Structure every article with H2 sections as answer blocks: H2 (question) -> 75-word answer paragraph -> supporting evidence -> table or list. Each H2 must be independently citeable.

STEP-BY-STEP:

  • Map your article to "what questions does each section answer?"
  • Rewrite H2s as those exact questions.
  • Write a direct 75-word answer as the first paragraph under each H2.
  • Follow with evidence, stats, or examples.

WHY IT MATTERS

AI citation systems extract content in chunks bounded by headings. Modular, self-contained sections are 4x more likely to be cited. Conductor's 2025 report shows structured heading architecture correlates with 67% higher AI citation rates.

AI Platforms: Google AI Overviews · ChatGPT · Perplexity

COMPETITOR INTEL

Healthline uses the most consistent Q&A heading architecture of any health domain -- this is the core reason they dominate AI medical citations.

BEST PRACTICE Topical Authority & Coverage Gaps

What to do: Search your main topics in Perplexity and ChatGPT -- note which competitors get cited. Identify what content they have that you don't. Create that content. Build comprehensive coverage.

STEP-BY-STEP:

  • Search 10-20 key queries in Perplexity and ChatGPT.
  • Record which domains get cited for each query.
  • Visit those domains -- identify what pages/formats they use.
  • List content types you're missing: guides, FAQs, data pages, comparisons.

WHY IT MATTERS

AI systems prefer citing domains that cover a topic comprehensively. Topical authority (covering all angles) is the strongest long-term GEO signal. Early adopters of GEO-ready content are discovered up to 10x faster than relying on organic SEO.

AI Platforms: All AI platforms -- Comprehensive domain authority

COMPETITOR INTEL

Relixir, Profound, and AthenaHQ charge $250-$400/month for. Your competitive intelligence: search your keywords in AI tools, note who wins, reverse-engineer their content strategy.

BEST PRACTICE H1 Heading

What to do: Use an authoritative, definitive H1 that signals comprehensiveness: "The Complete Guide to [Topic] ([Year])" or "Everything You Need to Know About [Topic]".

STEP-BY-STEP:

  • Add the current year to evergreen content H1s.
  • Use signifiers: "Complete Guide", "Ultimate Reference", "Expert Analysis", "In-Depth Review".
  • Include your primary keyword.
  • Match the H1 to what AI would title a response on this topic.

WHY IT MATTERS

RAG systems anchor topical relevance to H1 tags. Comprehensive-sounding H1s signal authoritative coverage to LLM retrieval pipelines. "Complete Guide" H1s are 60% more likely to be cited by primary sources.

AI Platforms: ChatGPT · Perplexity · Claude · Gemini

COMPETITOR INTEL

Investopedia leads banking/finance AI citations. Every Investopedia article uses "[Term]: Definition, Examples, and [Year] Guide" H1 format.

BEST PRACTICE Internal Links

What to do: Build hub-and-spoke clusters: pillar page + 8-12 cluster pages + lateral cluster-to-cluster links. Use descriptive anchor text. Link depth: important pages reachable in 2 clicks.

STEP-BY-STEP:

  • Map your top 3 topic areas as topic hubs.
  • Create/designate a pillar page for each hub.
  • Create cluster pages for every subtopic.
  • Link bidirectionally: pillar>cluster, cluster>cluster.

WHY IT MATTERS

LLMs assess topical authority by crawling link graphs. Dense, interconnected clusters signal domain expertise. Brands with strong topic clusters are cited 3-5x more. Early topical authority is very hard for competitors to displace.

AI Platforms: All AI platforms -- Domain topical authority signal

COMPETITOR INTEL

Hubspot's topic cluster model (pillar pages + clusters) is the benchmark -- credited as the reason they dominate AI citations for marketing, CRM, and sales topics.

BEST PRACTICE Content Freshness & Updates

What to do: Update key pages every 30-60 days with new data, statistics, or developments. Add a visible "Last Updated: [Month Year]" date stamp. Update og:article:modified_time every refresh.

STEP-BY-STEP:

  • Create a monthly content review calendar.
  • For each update: refresh at least one statistic with current data.
  • Add a new section or paragraph for recent developments.
  • Update og:article:modified_time in the page source.

WHY IT MATTERS

Perplexity and Bing Copilot heavily prioritise recent content. Pages with modification dates in the last 30-90 days get citation preference for competitive queries. Stale content loses AI citations over time as freshly-updated competitors displace them.

AI Platforms: Perplexity · Bing Copilot · Google AI Overviews -- Content freshness

COMPETITOR INTEL

Perplexity's citation previews explicitly show the page's modification date. The most-cited pages are updated every 30-60 days without exception.

BEST PRACTICE Open Graph Tags

What to do: Set og:description as a factual answer. Add og:article:published_time + og:article:modified_time. Freshness signals increase AI citation priority for time-sensitive queries.

STEP-BY-STEP:

  • Add og:description with a factual 2-sentence summary.
  • Add og:article:published_time in ISO 8601 format.
  • Update og:article:modified_time every time you update content.
  • Add og:article:author with a URL to the author profile.

WHY IT MATTERS

Perplexity and ChatGPT web browsing use modification timestamps as freshness signals. Pages updated within 30-90 days get citation priority for topical queries. Stale OG tags lose AI citations over time.

AI Platforms: Perplexity · ChatGPT web · Bing Copilot -- Content freshness

COMPETITOR INTEL

Perplexity's most-cited sources are updated every 30-60 days -- the modification timestamp is visible in citation previews.

BEST PRACTICE Canonical URL

What to do: Self-referencing canonical on every page pointing to HTTPS version. Consistent trailing-slash policy across entire site. All duplicate variants 301-redirect to canonical.

STEP-BY-STEP:

  • Audit canonical tags site-wide with Screaming Frog.
  • Fix missing canonicals -- every page needs one.
  • Ensure canonicals point to HTTPS, correct domain (www or non-www), consistent slash.
  • Set up 301s for all HTTP/www/slash variants.

WHY IT MATTERS

AI crawlers consolidate link authority at the canonical URL. Without it, duplicate content splits authority -- reducing citation probability for any version. Canonical hygiene is a quick technical win for GEO.

AI Platforms: Google AI Overviews · Bing AI -- Authority consolidation

COMPETITOR INTEL

Top-cited sites implement this as baseline.

BEST PRACTICE Citation & Source Network

What to do: Link to authoritative external sources (.gov, .edu, peer-reviewed research) for every claim. Format citations as: Author, Publication, Year. This makes your content trust-worthy to LLMs.

STEP-BY-STEP:

  • Audit every factual claim in your content.
  • Find the primary source (.gov, .edu, research paper, official report).
  • Link directly to that primary source.
  • Format in-text: "According to [Author] ([Year])..." with hyperlink.

WHY IT MATTERS

Content that cites authoritative sources is more likely to be cited itself. LLMs use citation networks as trust signals. Research shows pages citing .gov and .edu sources are included in AI training data 2.5x more often than unsourced pages.

AI Platforms: ChatGPT · Perplexity · Claude -- Trust & authority signaling

COMPETITOR INTEL

Every Wikipedia article -- the most-cited single source across all LLMs -- uses rigorous inline citations with primary sources. Model your citation format on Wikipedia's style.

BEST PRACTICE HowTo / Structured Steps Schema

What to do: For any tutorial, guide, or process page: add HowTo JSON-LD schema with numbered steps. Each step: name, text, optional image. Makes your process directly extractable by AI.

STEP-BY-STEP:

  • Identify all pages that explain a process or steps.
  • Add HowTo JSON-LD schema to each.
  • Include: name, description, totalTime, step array.
  • Each step: @type HowToStep, name (short label), text (instruction).

WHY IT MATTERS

HowTo schema is directly extracted by Google AI Overviews and Perplexity for procedural queries. Structured step-by-step content is featured in AI answers 3x more than unstructured prose. One schema implementation can dramatically increase AI visibility.

AI Platforms: Google AI Overviews · Bing AI · Perplexity

COMPETITOR INTEL

WikiHow -- the most-cited how-to source across all AI platforms -- uses structured step schemas on every single article. This is the technical foundation of their AI visibility.

BEST PRACTICE Server Response Time

What to do: Achieve <200ms TTFB. Enable server-side caching. Use a CDN. Optimise database queries. AI bots have strict crawl time limits -- slow pages get partially indexed.

STEP-BY-STEP:

  • Measure TTFB at web.dev/measure or gtmetrix.com.
  • Enable server caching: Redis or Memcached for dynamic sites.
  • Add Cloudflare (free tier) as a CDN.
  • Optimise the slowest database queries.

WHY IT MATTERS

AI crawlers have strict time budgets. Pages with TTFB >500ms get partially crawled -- your most important content (usually below the fold) may never be indexed. Fast pages = complete indexing = higher citation probability.

AI Platforms: All AI crawlers -- Indexing completeness

COMPETITOR INTEL

Top-cited sites implement this as baseline.