Authority Intelligence: Building AI Citation Authority From Zero

A 14-month case study in AI-first content optimization — methodology, data, and what 99% of analytics can’t show you.

Updated January 18, 2026 / Published: December 22, 2025 | Expertise: Authority Intelligence, GEO, AI Citation | Time to read: ~18 minutes

How to Build AI Citation Authority From Zero (While Google Ignores You)

Your analytics show minimal traffic, but AI systems are already crawling your content. The problem: 99% of that AI attention is invisible in Google Analytics 4. This 14-month case study reveals how we achieved 66% AI citation visibility with zero traditional authority signals—and how you can replicate the methodology.

You’ve optimized for Google rankings, but ChatGPT, Perplexity, and Claude still don’t cite you as an authority. Worse, your analytics show only 1% of the actual AI traffic hitting your servers. Over 14 months of systematic testing, we discovered that content structure and substance—not backlinks or domain age—drive AI citation authority. This article gives you the complete measurement framework, the 5-step implementation playbook, and the server-log methodology to track Dark AI Traffic on your own site.

TL;DR – Key Takeaways

- Dark AI Traffic is 99% invisible:
GA4 shows only 1% of actual AI crawler activity; server logs reveal the real picture.

- AI citation authority ≠ Google authority:
We achieved 66% AI visibility with zero backlinks, no author bio, and minimal Google traffic during the Sandbox period.

- Content substance beats credentials:
Well-structured, fact-dense content with clear entity coherence outperformed traditional E-E-A-T signals in AI systems.

- The 5-phase methodology works:
Baseline → Content Audit → Optimization → Authority Building → Measurement (repeatable in 12 weeks).

- Measurement is the bottleneck:
You need server logs + GA4 "Unassigned" channel analysis to see AI crawler patterns; regular analytics miss this completely.

- Phase 2 validation begins now:
Adding E-E-A-T signals to existing AI authority will reveal whether both systems compound or operate independently.

Part of the Authority Intelligence series.
For the full framework overview, see Gary Owl’s Strategic Authority Intelligence Framework V.30.3.

---

---

The 14-Month Experiment: What We Built and Why It Matters

Before diving into the data, context is necessary.

GaryOwl.com began as something personal: After 25 years working across industries and functions — from technical roles to strategic positions — I asked myself: What would it look like to bring all of this together?

The answer was a platform that could unite diverse expertise with a genuine passion for language, writing, and the emerging possibilities of artificial intelligence. Not a business venture, but something more fundamental: the pursuit of work that fulfills.

A blog, it turns out, is an ideal space for this. It allows you to write, to explore, to connect ideas across disciplines, to learn publicly. And when you combine decades of cross-functional experience with curiosity about how AI systems process and evaluate information, patterns start to emerge.

What I discovered through this work — organically, not by design — was that AI systems were paying attention. The content was being crawled, indexed, and cited. This observation transformed what started as a personal project into a systematic study.

The question shifted: If AI systems are already citing this content, what can we learn about how they evaluate and select sources?

A note on “we”: Throughout this article, “I” refers to personal decisions and observations. “We” refers to the systematic work — conducted in collaboration with AI agents. This article itself was developed through that collaboration: a human initiator working with AI to analyze, structure, and refine. It felt appropriate, given the subject matter.

The Technical Approach: Reverse Engineering AI Citation Behavior

The methodology differs from typical SEO or content marketing approaches in one fundamental way: rather than following published best practices, we systematically reverse-engineered how AI systems evaluate and cite content — and used machine learning to continuously optimize content strategy.

This involved:

Approach	Method
Analyzing AI responses	Studying which sources AI systems cite, how they cite them, and what content characteristics correlate with citation
Testing structural variations	Systematically varying content structure, source density, and formatting to identify what improves citation likelihood
Monitoring crawler behavior	Tracking which content AI crawlers prioritize and how crawl patterns change over time
Iterating based on data	Multiple framework versions, each responding to observed patterns rather than assumptions
ML-driven content optimization	Using machine learning to analyze performance patterns and continuously refine content strategy

This approach treats AI systems as technical systems to be understood — not black boxes to be gamed with superficial tactics. The machine learning component enables continuous adaptation as AI systems evolve.

What Made This Observable

As a new personal project focused on content and learning, GaryOwl.com naturally lacked the traditional authority signals that established sites accumulate over years:

What Was Absent	What This Allowed Us to Observe
Author bio	Whether AI systems require human credentials
About page	Whether content quality alone generates authority
LinkedIn/social proof	Whether authority shortcuts are necessary
Named persona	Whether personal reputation affects citation
Backlink outreach	Whether organic AI discovery is possible

This wasn’t a controlled laboratory experiment by design — it was a natural consequence of starting fresh. But it created a unique observation opportunity: What happens when content is the only signal?

The hypothesis that emerged: If content structure, source quality, and factual density matter more to AI systems than traditional E-E-A-T signals, then authority can be built — or strengthened — based on substance alone.

14 months later: The data allows us to evaluate this hypothesis.

---

The Strategic Reasoning: Why I Chose AI Over Google

New websites face a fundamental challenge: building authority takes time — often longer than businesses expect. This section explains why I chose to focus on AI visibility instead of fighting for Google rankings, and why this strategic decision may benefit both new entrants and established businesses with suboptimal digital authority.

The Google Sandbox Reality

Here’s a truth that SEO consultants rarely advertise: building Google authority takes time — often more than businesses expect.

For new domains, the so-called “Google Sandbox” effect means struggling to rank for 12 to 24 months, regardless of content quality. Even with:

• Perfect on-page SEO
• Quality backlinks
• Regular publishing
• Technical optimization

…new sites typically see minimal organic Google traffic in their first year.

But this isn’t just a problem for new entrants. Established businesses face a different challenge: their existing authority may not be as strong as they assume. Fragmented identities, outdated content, inconsistent information across platforms, or technical issues can silently erode authority signals they’ve spent years building.

The Numbers Don’t Lie

SEO practitioners consistently observe similar patterns for new domains:

Site Age	Typical Google Organic Traffic
0-6 months	Near zero
6-12 months	Minimal (< 100 visits/month)
12-18 months	Slowly growing
18-24 months	Beginning to establish
24+ months	Potential for real traction

For new domains, the first two years are essentially a trust-building period where Google evaluates legitimacy. For established businesses, the challenge is different: ensuring that existing authority signals remain sharp and coherent as the digital landscape evolves.

The Authority Gap

The challenge extends beyond new sites. There is evidence suggesting that AI systems may also favor established authorities — sites with existing brand recognition, extensive backlink profiles, and long publishing histories.

But here’s what’s often overlooked: many established businesses have suboptimal digital authority without realizing it.

Situation	Problem
New sites	Google Sandbox (12-24 months), no established authority signals
Established businesses with fragmented identity	Inconsistent information across platforms dilutes authority
Companies with outdated digital presence	Legacy content, stale profiles, abandoned social accounts send mixed signals
Businesses unaware of their authority gaps	Technical issues, poor entity coherence, or weak citation profiles they’ve never audited

This creates a broader challenge than just “new vs. established.” Even companies with decades of history may have digital authority that isn’t properly sharpened — their online presence doesn’t accurately reflect their actual expertise and credibility.

This is precisely why developing approaches to address these authority gaps matters. The need for methodologies that can audit, diagnose, and strengthen digital authority applies to new entrants and established players alike.

The Opportunity Cost Question

This creates strategic questions:

For new sites: Should you spend 12-24 months fighting for Google rankings that may never come — or invest that time differently?

For established businesses: Are your authority signals as strong as you think? Could AI-focused optimization reveal gaps you weren’t aware of?

I asked myself: What if I redirected energy into a channel where content substance matters more than accumulated credentials — and simultaneously built something more valuable than traffic?

The AI Alternative

AI systems don’t have a formal “sandbox” like Google. They evaluate content based on:

• Structure and clarity
• Source quality and citations
• Factual density
• Semantic comprehensiveness
• Freshness and relevance

While established authorities may still have advantages, the barrier to entry appears lower. Content substance can compensate for missing credentials — at least partially. This was the hypothesis I set out to test.

The Strategic Decision

I made a deliberate choice:

Traditional Path	Chosen Path
Fight Google Sandbox for 2 years	Accept minimal Google traffic
Build backlinks slowly	Focus on content substance
Wait for domain authority	Build AI citation authority
Hope for organic growth	Develop a replicable framework

The hypothesis: By the time Google would start trusting the site anyway (18-24 months), I’d have established significant AI authority — and developed a methodology that could benefit others.

---

What AI Systems Prefer: Content Quality Over SEO Tactics

AI systems evaluate content differently than traditional search engines. This section documents what we observed about AI citation patterns — which content characteristics correlate with higher citation rates, and which traditional SEO tactics appear less relevant.

The Shift in What Matters

Traditional SEO optimized for search engine algorithms: keyword density, meta tags, backlink profiles. AI systems evaluate content differently.

Based on our observation period, AI systems appear to prioritize:

Factor	Why It Matters for AI
Natural language	LLMs detect and discount keyword stuffing
Structured information	Tables, clear hierarchies aid RAG extraction
Source quality	Citations to credible sources increase citation likelihood
Substantive depth	Shallow content doesn’t provide citable facts
Clear, direct titles	Clickbait titles don’t improve AI visibility

Content Practices We Avoided

From the start, certain practices were avoided — not as experimental constraints, but as a natural consequence of focusing on quality:

Avoided	Why
Keyword stuffing	AI systems recognize unnatural language patterns
Clickbait headlines	Sensationalist titles don’t improve AI citation rates
Thin content	AI needs substantive material to cite
Walls of text	Structure helps RAG systems extract relevant passages

Content Practices We Prioritized

Prioritized	Implementation
Natural writing	Content reads like it was written for humans, not algorithms
Clear structure	Headers, tables, bullet points where appropriate
Source triangulation	Academic, industry, and practitioner sources
Factual density	Specific data points, not vague claims
Honest titles	Descriptive, accurate, not sensationalized

Server Log Evidence: What AI Crawlers Target

Server logs show AI crawlers targeting content with these characteristics:

• Long-form articles (2,000+ words): Higher crawler attention
• Structured posts with tables: Preferred by RAG systems
• Well-sourced content with citations: More likely to be cited

“The content that performs well for AI systems is content that respects the reader — well-structured, well-sourced, and written naturally.”

Implications for Content Strategy

Traditional SEO Focus	AI-First Content Focus
Keyword optimization	Natural language
Meta tag perfection	Content substance
Backlink acquisition	Source quality
Click-through optimization	Accuracy and depth

This doesn’t mean abandoning structure or ignoring discoverability. It means prioritizing content quality and trusting that AI systems will recognize substance.

---

Why This Matters

Traditional authority is built on reputation, backlinks, and time. AI-native authority may work differently. This section examines the hypothesis behind Authority Intelligence: that content structure and substance can generate citation authority independent of traditional signals.

The Problem with Traditional Authority

In traditional SEO, authority is built through:

• Personal brand recognition
• Backlinks from reputable sites
• Social proof and credentials
• Years of established presence

This creates a chicken-and-egg problem: You need authority to get visibility, but you need visibility to build authority.

For new entrants, this means years of grinding before seeing results — if ever. For established businesses, it means authority gaps can persist unnoticed, silently undermining visibility.

The Authority Intelligence Hypothesis

The Authority Intelligence Framework proposes a different model:

In AI-native search, content structure and substance can generate citation authority independent of traditional reputation signals — and independent of domain age or legacy presence.

If true, this would benefit both new voices seeking to compete with established players and established businesses seeking to sharpen authority signals they may have neglected.

The Clean Room Methodology

GaryOwl.com was built as a “clean room” environment:

Variable	Control Method
Personal reputation	Fictional persona (Gary Owl)
Existing authority	Brand-new domain
Social proof	Zero social media presence
Backlinks	No outreach, purely organic
Author credentials	Deliberately hidden
Domain age advantage	None (started from zero)

Everything that could provide a traditional authority shortcut was removed. The only variable was the content itself — structured according to the Authority Intelligence Framework.

---

The Results: What 14 Months of Data Revealed

This section presents the empirical findings from 14 months of observation. The data covers AI crawler behavior, citation performance across platforms, and Google traffic patterns — including the impact of the December 2025 Core Update.

AI Systems: Phase 1 Results

The data confirms that AI systems discovered, crawled, and cited GaryOwl.com content despite zero traditional authority signals and a brand-new domain.

Server-Level Evidence

Metric	Finding
Total server requests (30 days)	Massive crawler activity
Microsoft infrastructure*	~19% of all traffic
AWS (Claude/Perplexity)	~4% of all traffic
Google (Gemini)	~3% of all traffic
Identified AI user-agents	ChatGPT-User, MistralAI-User

*Microsoft traffic includes Bingbot (traditional search crawler) and potentially Copilot (AI assistant). Current logging doesn’t distinguish between them — a measurement limitation we’re addressing in Phase 2.

The Critical Finding: User-Initiated Requests

Not all AI traffic is training crawlers. The ChatGPT-User/1.0 and MistralAI-User/1.0 user-agents represent user-initiated requests — meaning:

1. A real human asked ChatGPT or Mistral a question
2. The AI determined it needed current information
3. The AI fetched GaryOwl.com content in real-time
4. That content informed the answer

This is the equivalent of a search engine click — except it’s invisible to analytics.

Citation Performance

Platform	Visibility	Citation Share
Perplexity	Strongest	Top 3 for relevant queries
ChatGPT	Significant	Consistent citations
Claude	Present	Growing

These results were achieved within 14 months, during a period when Google organic traffic remained minimal due to the Sandbox effect.

Google: Expected Minimal Results

We didn’t fight for Google rankings, so we didn’t get them. But interestingly:

The Zero-Click Paradox

Metric	Value	Trend
Impressions	Growing	↑ 30%
Average Position	Improving	Better rankings
Click-Through Rate	0.11%	Near zero

Even without SEO focus, Google started showing the content more — but the zero-click phenomenon meant minimal actual traffic. This raised a question worth exploring: Is traditional Google traffic optimization becoming less effective for certain content types?

The December 2025 Core Update Impact

Then came the December 2025 Core Update.

Phase	Period	Daily Traffic	Change
Baseline	Late November	Normal	—
Spike	Nov 28 – Dec 2	3-5x increase	Viral moment
Post-Update	Dec 13+	Below baseline	↓ 30%

The timing correlates with Google’s update, which emphasized E-E-A-T signals and author transparency. The site lacked these signals by design — and the traffic data reflects that.

This brings us to the most interesting question: What happens when we add E-E-A-T signals to a site that already has established AI authority?

Phase 2 will validate this.

What is Dark AI Traffic?

Now let’s examine the phenomenon that makes AI traffic so difficult to measure.

Dark AI Traffic refers to website visits from AI systems that remain invisible to client-side analytics like Google Analytics 4.

The Three Layers of AI Traffic

Layer	Description	Visibility	Example
Visible	Users clicking links from AI platforms	✅ Shows in GA4	perplexity.ai / referral
Dark	AI fetching content without referrers	⚠️ Shows as “(not set)”	ChatGPT browsing
Invisible	Crawler requests with no JS	❌ Server logs only	ChatGPT-User/1.0

Layer 1 represents less than 1% of total AI-related traffic.

The Evidence in GA4

Method 1: The “Unassigned” Channel Explosion

Channel	Month-over-Month Change
Direct	↓ 35.61%
Unassigned	↑ 614.29%
Organic Search	↓ 23.33%
Referral	↓ 69.23%

A 614% increase in Unassigned traffic while every other channel declined is not random noise.

Method 2: Engagement Rate Analysis

Channel	Engagement Rate	Interpretation
Referral	75%	Human ✅
Organic Search	39%	Human ✅
Direct	16%	Mixed
Unassigned	8%	Bot 🚨

An 8% engagement rate means visitors load the page and immediately leave without interaction — textbook crawler behavior.

Method 3: The “(not set)” Landing Page

When filtered by landing page:

Landing Page	Sessions	Change
(not set)	100%	↑ 2050%

Every Unassigned session had no identifiable landing page. This happens when JavaScript doesn’t execute — confirming bot/crawler origin.

Note: While some “Unassigned” traffic may stem from privacy-focused browsers or VPNs, the absence of JavaScript rendering combined with server log correlation strongly suggests bot dominance.

The Visibility Gap

Based on GaryOwl.com server log analysis:

Server Requests:     ~500x more than GA4 sessions
GA4 Sessions:        What Google shows you
The Gap:             99%+ invisible traffic

This isn’t a tracking error. It’s the new reality of AI-driven web traffic.

---

Identifying AI Crawlers in Server Logs

Server logs reveal which AI systems are accessing your content. This section documents the confirmed AI user-agents we observed and explains the critical distinction between training crawlers and user-initiated requests.

Confirmed AI User-Agents

User-Agent	Type	Company
ChatGPT-User/1.0	User-initiated	OpenAI
MistralAI-User/1.0	User-initiated	Mistral AI
GPTBot/1.0	Training	OpenAI
ClaudeBot/1.0	Training	Anthropic
PerplexityBot	Search	Perplexity

The User-Initiated Distinction

Training crawlers (GPTBot, ClaudeBot) scrape for model training. Important, but indirect impact.

User-initiated crawlers (ChatGPT-User, MistralAI-User) fetch content because a human asked a question right now. This is direct citation activity.

“Every ChatGPT-User request represents a real person who received your content as part of an AI answer — even though you’ll never see them in analytics.”

---

Phase 1 Conclusions

What did the data reveal? This section summarizes the key findings from Phase 1, identifies remaining questions, and establishes the foundation for Phase 2 testing.

What the Data Suggests

Hypothesis	Phase 1 Finding
AI systems can discover content without traditional authority	Data supports this
New domains can build AI visibility during Google Sandbox period	Data supports this
Content structure affects AI citation probability	Data supports this
User-initiated AI requests indicate real discovery	Data supports this
AI-first approach is viable during Google Sandbox period	Data supports this, with caveats
Natural, well-structured content correlates with AI citation	Data supports this

Additional Observations: Google and E-E-A-T

Finding	Implication
Google still requires E-E-A-T signals	Parallel optimization eventually needed
AI visibility ≠ Google protection	These appear to be separate systems
December 2025 update affected traffic	Credentials matter for Google
The experiment produced a replicable methodology	Framework ready for Phase 2 testing

The Trade-Off Matrix

Strategy	AI Citation	Google Ranking	Observation
Content-only (Phase 1)	Present	Vulnerable	Our current state
Traditional SEO only	Limited	Stronger (eventually)	Not tested
Combined approach	Expected: Strong	Expected: Strong	Phase 2 will test this

Open Questions for Phase 2

• Will adding E-E-A-T signals improve Google traffic without affecting AI citations?
• Do AI authority and Google authority compound, or are they independent?
• What’s the optimal sequencing — AI-first then Google, or parallel?

These will be validated in Phase 2.

---

The GEO Landscape: Where Authority Intelligence Fits

Generative Engine Optimization (GEO) is an emerging field with multiple approaches and practitioners. This section clarifies how Authority Intelligence relates to established GEO methodologies, SEO, and the broader landscape of AI-focused optimization.

Understanding the Terminology

The field of AI Search optimization is still young, and terminology varies across practitioners. Here’s how the key concepts relate:

Term	Definition	Focus
GEO (Generative Engine Optimization)	The practice of optimizing content for AI-powered search engines	Getting cited in AI responses
AEO (Answer Engine Optimization)	Optimizing for direct answers in search	Featured snippets, AI Overviews
LLMO (Large Language Model Optimization)	Technical optimization for LLM comprehension	Model-specific tuning
Authority Intelligence	Strategic framework for building and sharpening citation authority in AI systems	Addressing authority gaps for new and established entities

The Relationship to SEO

Authority Intelligence operates at the boundary of SEO — complementary, not competing.

Aspect	Traditional SEO	Authority Intelligence
Primary focus	Google rankings	AI citation authority
Overlap area	Technical SEO	Technical SEO
Signals optimized	Backlinks, keywords, page speed	Entity coherence, content structure, source quality
Relationship	Foundation	Complementary layer

The overlap is Technical SEO: structured data, schema markup, crawlability, site architecture, and semantic HTML. These elements matter for both Google and AI systems. Authority Intelligence builds on this technical foundation while extending into AI-specific optimization that traditional SEO doesn’t address.

We don’t replace SEO. We extend it — adding a layer of optimization specifically designed for how AI systems evaluate and cite content.

How Authority Intelligence Differs

While GEO focuses on tactics (structured data, citations, formatting), Authority Intelligence addresses a more fundamental question:

How can content become a trusted, citable source for AI systems — independent of traditional authority signals like backlinks, domain age, or author credentials?

Leading AI search and GEO practitioners — including Evan Bailyn (First Page Sage), Jason Barnard (Kalicube), Mike King (iPullRank), Lily Ray, Aleyda Solís, and Kevin Indig — alongside LLM researchers such as Pranjal Aggarwal, have laid essential groundwork for modern AI search visibility and GEO practice. Their analyses of AI search behavior, entity-centric E‑E‑A‑T, and retrieval strategies, together with UC Berkeley’s GEO research on visibility optimization, inform the Authority Intelligence Framework.

Authority Intelligence builds on this foundation but differs in three ways:

1. It asks a different question: What if you don’t have traditional authority to begin with?
2. It uses a reverse engineering methodology and continuous ML-driven optimization: Rather than applying published best practices, it systematically analyzes AI system behavior to derive optimization principles from observed patterns — and uses machine learning to continuously refine content strategy based on performance data.
3. It emphasizes Entity Coherence: AI systems aggregate information from multiple sources. Inconsistent identities across platforms fragment authority rather than build it. A Single Source of Truth — with all digital presences aligned — is foundational.

This technical approach treats AI citation as an engineering problem — studying inputs and outputs to understand the underlying evaluation logic, while ensuring the entity itself is coherently represented across the web.

Strategic Identity Revelation: Working With AI’s “Curiosity”

One methodological detail deserves explanation: We didn’t conceal identity throughout the experiment — we revealed it progressively.

Why? AI systems continuously attempt to connect information across sources. When encountering partial entity data, they actively seek to resolve ambiguities — linking names, roles, affiliations, and context clues to build coherent entity profiles. This behavior is well-documented in knowledge graph research: LLMs perform entity resolution by clustering synonymous references and resolving multiple representations of the same real-world entity.

We leveraged this by strategically placing identity breadcrumbs across content — not hiding, but revealing incrementally. The hypothesis: AI training crawlers (GPTBot, ClaudeBot, and others) would encounter these signals, and the data would eventually inform future model training.

Phase	Identity Signal	Purpose
Early	Thematic consistency only	Establish topical authority
Mid	Contextual hints, cross-references	Enable entity clustering
Late	Explicit credentials, author identity	Complete the entity profile

This approach works with how AI systems naturally process information — rather than against it. By the time full credentials are revealed, the entity already exists as a coherent node in the AI’s understanding, with established content authority attached.

The GaryOwl.com Contribution

What began as a personal project — uniting 25 years of cross-industry expertise with a passion for language and AI — evolved into a 14-month study of AI citation behavior. Because the site started fresh, without accumulated authority signals, it offered a unique vantage point: observing how AI systems respond to content substance alone.

If content quality could generate AI citation authority for a site with zero existing authority, the principles should apply even more effectively to established businesses seeking to sharpen their digital presence.

What GEO Research Showed	What Authority Intelligence Tested
Structured content improves citation rates	Can structure alone build authority?
E-E-A-T signals matter for AI	Can you succeed without E-E-A-T?
Backlinks correlate with AI visibility	Can you achieve visibility without backlinks?
Established domains perform better	Can a new domain compete despite this bias?

The Phase 1 data suggests the answer to all four questions is conditionally yes — content substance can partially compensate for missing credentials. For established businesses, this means the methodology can strengthen existing authority rather than build from zero.

The Authority Intelligence Framework

The methodology that emerged from this experiment is now documented (V.30+) and forms part of the octyl® toolchain— a framework for developing customized applications.

The framework integrates:

• Technical SEO foundations as the shared base for Google and AI optimization
• Reverse engineering methodology for understanding AI citation behavior
• Machine learning optimization for continuous content strategy refinement
• UC Berkeley GEO research (arXiv:2311.09735) as scientific foundation
• Knowledge Graph content planning for systematic topical authority
• Entity Coherence principles ensuring consistent identity across platforms
• 14 months of empirical data from GaryOwl.com
• Continuous iteration refining the methodology
• Practical implementation guides for businesses

This positions Authority Intelligence as a specialized approach within the broader GEO landscape — designed to address authority gaps across different situations:

• New domains facing Google Sandbox and building authority from scratch
• Established businesses with fragmented digital identity or suboptimal authority signals
• Companies unaware of their authority gaps — technical issues, poor entity coherence, or weak citation profiles
• Organizations seeking to sharpen their digital presence for AI-driven discovery channels
• Content-first strategies where traditional credentials are limited

---

What We Built: The Authority Intelligence Methodology

The experiment didn’t just generate data — it produced a documented, replicable methodology. This section describes what 14 months of systematic work produced and how it evolved from observation to actionable framework.

14 Months of Documentation

While not chasing Google rankings, we documented everything. This period produced:

• Reverse engineering insights from systematic analysis of AI citation behavior
• Machine learning models for continuous content strategy optimization
• Multiple framework iterations responding to data and observations
• Measurement approaches for Dark AI Traffic
• Content structure patterns that correlated with AI citation
• Knowledge Graph content roadmap for systematic topic coverage
• Entity Coherence guidelines for consistent identity across platforms
• A replicable methodology (Authority Intelligence)

The Authority Intelligence Framework evolved through systematic iteration:

Phase	Versions	Focus
Foundation	V.01-V.05	Basic AI-optimized content structure
Reverse Engineering	V.06-V.19	Systematic analysis of AI citation patterns and crawler behavior
Research Integration	V.20-V.28	Incorporating GEO research, source triangulation
Current	V.29-V.30+	Documented methodology, ready for Phase 2 testing

From Documentation to Application

What started as a question — “Can content achieve AI citation without traditional authority?” — produced a documented methodology.

The framework addresses:

• Technical SEO Foundations: Structured data, schema, crawlability — the shared base for Google and AI
• Dark AI Traffic Measurement: Understanding what GA4 doesn’t show
• Content Structure: Patterns that correlate with AI citation
• Knowledge Graph Content Roadmap: Systematic mapping of topic relationships to build semantic authority
• Entity Coherence: Ensuring consistent identity across platforms to strengthen — not fragment — authority
• Authority Building: Steps for establishing AI visibility
• Dual Strategy Preparation: Groundwork for Phase 2 E-E-A-T integration

The Knowledge Graph approach deserves brief mention: rather than publishing isolated articles, the methodology maps topic relationships and develops content that reinforces semantic connections. AI systems recognize topical authority partly through how comprehensively and coherently a site covers a domain. The content roadmap ensures each piece strengthens the overall knowledge structure.

“We didn’t just collect data. We built a replicable methodology.”

---

Transitioning to Phase 2

Phase 1 of the experiment is complete. The data shows:

What worked: AI systems discovered, crawled, and cited GaryOwl.com content despite zero traditional authority signals. The Authority Intelligence methodology produced measurable results in AI visibility.

What didn’t work: Google’s December 2025 Core Update confirmed that traditional search still requires traditional signals. A content-only approach has limits.

What we don’t know yet: The most interesting question remains unanswered.

The Phase 2 Hypothesis

We’ve established AI citation authority without E-E-A-T signals. Now we’re adding them:

Element	Status
Full author bio	In progress
Professional credentials	In progress
LinkedIn integration	Phase 2
About page	In progress

The question: Will strengthening E-E-A-T signals on a site with existing AI authority produce compounding effects? Or are these parallel, independent systems?

We genuinely don’t know. The data will tell us.

What This Article Represents

This isn’t a victory lap. It’s a status report at the transition between phases.

Phase 1 answered: Can content substance build AI authority without credentials? The data suggests yes.

Phase 2 will answer: What happens when you add credentials to existing AI authority? We’ll document the results.

“The experiment isn’t over. It’s entering its most interesting phase.”

---

Bringing This to Businesses

The experiment produced insights. This section addresses how these findings translate into practical applications for businesses — and what currently exists to help organizations implement these approaches.

The Challenge

Most businesses can’t run 14-month experiments. They need to make decisions with incomplete information about:

• Whether AI-first strategies are relevant to their situation
• How to measure Dark AI Traffic on their properties
• When to invest in traditional SEO vs. AI optimization
• How to structure content for AI citation

What Exists

The experimentation at GaryOwl.com produced a replicable methodology, now part of the octyl® toolchain.

Component	Description
Authority Intelligence	The methodology for building AI citation authority, documented through this experiment
octyl® toolchain	A framework enabling flexible development of customer-specific applications
This documentation	The findings from Phase 1, shared openly

What This Means

The intellectual property — including the octyl® brand (registered trademark) and the Authority Intelligence methodology — is protected. The framework exists. What form future applications will take depends on Phase 2 findings and market needs.

For now: the methodology is documented here. Businesses can evaluate whether the approach is relevant to their situation.

---

How To: Implementing Authority Intelligence

This section documents the approach developed at GaryOwl.com. It’s not a guaranteed formula — it’s a methodology with Phase 1 data behind it.

Phase 1: Baseline Assessment (Week 1-2)

Step 1.1: Measure Your Current AI Visibility

Before optimizing, understand where you stand.

Manual Testing:

1. Open ChatGPT, Perplexity, Claude, and Gemini
2. Ask questions your content should answer
3. Document: Are you cited? How prominently? Verbatim or paraphrased?
4. Repeat with 10-20 relevant queries
5. Calculate your "Citation Rate" = Citations / Queries Tested

Benchmark:

Citation Rate	Interpretation
0-10%	Not visible to AI systems
10-30%	Occasional presence
30-50%	Moderate authority
50%+	Strong AI authority

Step 1.2: Analyze Your Dark AI Traffic

Follow this GA4 workflow:

GA4 → Reports → Acquisition → Traffic Acquisition
→ Primary Dimension: "Session default channel group"
→ Look for "Unassigned" channel
→ Note: Sessions, Engagement Rate, Trend

Key Metrics to Record:

Metric	Where to Find	What It Tells You
Unassigned Sessions	Traffic Acquisition	Volume of Dark AI Traffic
Engagement Rate	Same report, scroll right	Bot vs. Human (< 15% = likely bots)
“(not set)” in Source/Medium	Change dimension	Unidentified traffic sources

Step 1.3: Check Server Logs for AI Crawlers

If you have server access (via your hosting provider’s control panel or server logs):

Search for these User-Agent strings:
- ChatGPT-User
- MistralAI-User  
- GPTBot
- ClaudeBot
- PerplexityBot
- Anthropic
- OAI-SearchBot

Calculate Your AI Crawler Ratio:

AI Crawler Ratio = AI Bot Requests / Total Requests

Ratio	Interpretation
< 5%	Low AI attention
5-15%	Moderate AI crawling
15-30%	High AI interest
> 30%	Very high AI attention

---

Phase 2: Content Audit (Week 2-3)

Step 2.1: Evaluate Content Structure

AI systems prefer structured, scannable content. Audit your pages:

Structure Checklist:

Element	Present?	AI Impact
Clear H1 with topic	☐	High
Logical H2/H3 hierarchy	☐	High
TL;DR or summary at top	☐	Very High
Tables for comparisons	☐	Very High
Bullet points for lists	☐	Medium
FAQ section	☐	High
Published/Updated dates	☐	Medium
Author information	☐	Medium (for Google)

Step 2.2: Assess Source Quality

Authority Intelligence emphasizes source triangulation:

The Triangulation Principle:

Each substantive claim should be supported by:
├── 1+ Academic source (papers, studies)
├── 1+ Industry source (reports, data)
└── 1+ Thought leader reference (experts, practitioners)

Audit Questions:

• Do your articles cite sources?
• Are sources linked and verifiable?
• Is there a references/sources section?
• Do you cite primary sources or just aggregators?

Step 2.3: Measure Semantic Completeness

AI systems favor comprehensive content. For each key topic:

1. Ask ChatGPT: "What are the key aspects of [your topic]?"
2. Compare AI's list to your content
3. Identify gaps
4. Document missing subtopics

---

Phase 3: Content Optimization (Week 3-6)

Step 3.1: Implement GEO-16 Principles

Based on GEO research by Aggarwal et al. (Princeton University & Indian Institute of Technology Delhi, 2024) and the GEO‑16 framework by Kumar and Palkhouski (University of California, Berkeley & Wrodium Research, 2025), these optimizations improve AI citation rates:

Optimization	How to Implement	Citation Impact
Cite Sources	Add academic/industry references	+30-40%
Add Statistics	Include specific numbers and data	+20-30%
Use Quotations	Quote experts and studies	+15-25%
Technical Terms	Use precise domain vocabulary	+10-20%
Fluent Language	Clear, well-written prose	+10-15%
Authoritative Tone	Confident, expert voice	+10-15%

Step 3.2: Structure for RAG Systems

RAG (Retrieval-Augmented Generation) systems extract chunks of content. Optimize for this:

Chunk-Friendly Formatting:

markdown

## [Clear Topic Heading]

[2-3 sentence summary of this section]

[Detailed explanation with specific facts]

| Comparison | Option A | Option B |
|------------|----------|----------|
| Feature 1  | Value    | Value    |

**Key Takeaway:** [One-sentence summary]

Step 3.3: Create AI-Optimized Content Types

Based on GEO research by Aggarwal et al. (Princeton University & Indian Institute of Technology Delhi, 2024) and further GEO‑16 analysis by Kumar and Palkhouski (University of California, Berkeley & Wrodium Research, 2025), these content formats correlate with higher AI citation rates:

Content Type	AI Citation Rate	Why It Works
Comprehensive Guides	Very High	Covers topic completely
Framework Explanations	Very High	Structured, citable
Data-Driven Analysis	High	Unique, factual
Comparison Articles	High	Tables are AI-friendly
How-To Tutorials	Medium-High	Step-by-step structure
Opinion/Commentary	Low	Not factual enough

Step 3.4: Develop a Knowledge Graph Content Roadmap

Rather than publishing isolated articles, map topic relationships systematically:

1. Identify your core topic domain
2. Map related subtopics and their connections
3. Plan content that covers the domain comprehensively
4. Ensure internal linking reflects semantic relationships
5. Each new piece should reinforce the overall knowledge structure

AI systems recognize topical authority partly through how coherently a site covers a domain. A structured content roadmap — built around a knowledge graph of your topic area — signals depth and expertise more effectively than scattered, unrelated posts.

---

Phase 4: Authority Building (Week 6-12)

Step 4.1: Establish Entity Coherence

A critical — and often overlooked — principle: AI systems aggregate information about entities from multiple sources across the web. If these sources present inconsistent information, the resulting entity understanding becomes fragmented, weakening authority signals.

The Single Source of Truth Principle:

Every digital presence of your brand or personal identity should align with one authoritative source. Inconsistencies between platforms don’t just confuse users — they confuse AI systems attempting to build a coherent entity model.

Problem	Consequence
Different company descriptions across platforms	AI systems can’t determine what the entity is
Inconsistent founding dates, locations, or team info	Reduced confidence in entity accuracy
Multiple personas or brand variations	Authority diluted across fragmented identities
Outdated information on secondary platforms	Conflicting signals degrade entity trust

Entity Coherence Checklist:

Element	Requirement
Company/personal name	Identical across all platforms
Description/bio	Consistent core messaging
Contact information	Single authoritative address, phone, email
Visual identity	Same logo, profile images
Key facts	Founding date, location, credentials aligned

Step 4.2: Establish Entity Recognition

Once coherence is established, AI systems need to recognize your entity:

Entity Recognition Checklist:

Action	Platform	Impact
Claim Google Knowledge Panel	Google	High
Complete Wikipedia (if notable)	Wikipedia	Very High
Consistent NAP across web	All directories	Medium
Schema markup (Organization, Person)	Your website	High
Wikidata entry	Wikidata	Medium

The Wikidata entry deserves special mention: it often serves as a machine-readable Single Source of Truth that AI systems reference. Ensuring accuracy there propagates correct information across systems.

Step 4.3: Build Cross-Platform Presence

AI systems aggregate information from multiple sources:

Platform Priority for B2B:

Platform	Why It Matters	Action
LinkedIn	Professional authority	Publish articles, engage
Medium	Indexed by AI systems	Republish key content
Industry forums	Domain expertise	Answer questions
Podcast appearances	Mentioned in transcripts	Guest on relevant shows
Guest posts	External validation	Write for industry sites

Step 4.3: Understanding E-E-A-T Integration

Why GaryOwl.com separated AI and Google signals:

Most businesses cannot isolate AI authority from traditional E-E-A-T signals — they already have existing websites, author profiles, backlinks, and brand presence. This makes it difficult to determine which signals produce which effects.

GaryOwl.com deliberately separated these phases to enable clear signal attribution:

Phase	Signals Present	What We Could Measure
Phase 1 (Months 1-14)	Content quality only	Pure AI response to substance
Phase 2 (Now beginning)	Content + E-E-A-T	Combined/compounding effects

This sequential approach — not recommended as a general strategy, but chosen for methodological clarity — allows us to attribute changes to specific signal additions. When Phase 2 data becomes available, we’ll know whether E-E-A-T signals compound with existing AI authority, operate independently, or produce interference.

E-E-A-T Implementation (Phase 2):

Author Pages:
├── Full name and photo
├── Professional credentials
├── LinkedIn/social links
├── List of published articles
└── External mentions/features

About Page:
├── Company/site mission
├── Team credentials
├── Contact information
├── Trust indicators (awards, clients, press)
└── Privacy policy, ToS

For businesses implementing both simultaneously: the framework still applies, but signal attribution will be less precise. The principles remain valid; the measurement clarity differs.

---

Phase 5: Measurement & Iteration (Ongoing)

Step 5.1: Establish KPIs

Move beyond traditional traffic metrics:

Authority Intelligence KPIs:

KPI	How to Measure	Target
AI Citation Rate	Manual testing (monthly)	> 30%
Citation Share	Tools like Profound, Otterly	Top 10 for key queries
Dark AI Traffic	GA4 “Unassigned” channel	Growing month-over-month
AI Crawler Requests	Server logs	Increasing trend
User-Initiated Requests	ChatGPT-User in logs	Any presence = success

Step 5.2: Monthly Review Cycle

Week 1: Test AI citations (10-20 queries)
Week 2: Analyze GA4 Dark Traffic trends
Week 3: Review server logs for crawler activity
Week 4: Update content based on findings

Step 5.3: Quarterly Deep Dive

Every 3 months:

• Full competitor citation analysis
• Content gap assessment
• Framework version update
• Strategy adjustment

---

Quick-Start Checklist

For those who want to begin immediately with Authority Intelligence — the GEO methodology developed at GaryOwl.com:

Today:

•  Test 5 queries in ChatGPT — are you cited?
•  Check GA4 for “Unassigned” traffic
•  Identify your top 3 content pieces for optimization

This Week:

•  Add TL;DR to your top 3 articles
•  Add tables where comparisons exist
•  Add source citations to claims

This Month:

•  Implement full content audit
•  Create one comprehensive guide using Authority Intelligence principles
•  Set up monthly citation testing routine

This Quarter:

•  Complete Phase 1-3 implementation
•  Measure baseline vs. post-optimization citation rates
•  Decide on E-E-A-T implementation timeline

---

Schema Markup for GEO

Structured data helps AI systems understand your content. Essential schema types for Authority Intelligence:

json

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Your Article Title",
  "author": {
    "@type": "Person",
    "name": "Author Name",
    "url": "https://yoursite.com/author/name"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Your Brand",
    "url": "https://yoursite.com"
  },
  "datePublished": "2025-12-22",
  "dateModified": "2025-12-22"
}

Priority Schema Types:

Schema Type	Use Case	GEO Impact
Article	Blog posts, guides	High
HowTo	Tutorial content	Very High
FAQPage	FAQ sections	Very High
Organization	Company info	Medium
Person	Author pages	Medium

---

When Professional Support May Help

The methodology is documented here for independent implementation. As the octyl® toolchain develops, structured implementation support may become available. For now, the documentation in this article provides the foundation for self-directed implementation.

---

How to Track Dark AI Traffic on Your Site

Most AI-related traffic is invisible to standard analytics. This section provides a step-by-step guide to measuring Dark AI Traffic using GA4 and server logs — the same methods we used to identify the patterns described in this article.

Step 1: Monitor “Unassigned” in GA4

GA4 → Reports → Acquisition → Traffic Acquisition
→ Look for "Unassigned" channel
→ Track month-over-month changes

Warning sign: Rapid growth with engagement rate below 15%

Step 2: Analyze “(not set)” Sources

GA4 → Reports → Acquisition → Traffic Acquisition
→ Filter for "(not set)" in source/medium

Step 3: Check Server Logs

Search for these user-agent strings:

• ChatGPT-User
• MistralAI-User
• GPTBot
• ClaudeBot
• PerplexityBot

Step 4: Calculate Your Visibility Ratio

Visibility Ratio = GA4 Sessions / Server Requests

• >50%: Mostly human traffic
• 10-50%: Significant bot presence
• <10%: Bot-dominated
• <1%: Extreme (like our case)

---

The Dual Optimization Strategy

Should you optimize for AI or Google? The answer is both — but the sequencing matters. This section explains the dual optimization approach we used and why we separated AI optimization from E-E-A-T implementation.

The GaryOwl.com Approach

GaryOwl.com deliberately separated AI optimization from E-E-A-T implementation. This wasn’t because we recommend this sequence for everyone — it’s because signal isolation enables clearer measurement.

For AI Systems (Authority Intelligence):

Principle	Implementation
Structured content	Clear headings, tables, lists
Source triangulation	Academic + industry + thought leader sources
Semantic depth	Comprehensive topic coverage
Factual density	Claims backed by evidence
Update frequency	Fresh, current information

For Google (E-E-A-T):

Element	Implementation
Author transparency	Full bio with credentials
External validation	Guest posts, citations, mentions
Social proof	LinkedIn, professional profiles
Trust signals	About page, contact information

Why We Sequenced These Phases

Approach	Advantage	Disadvantage
Sequential(GaryOwl.com)	Clear signal attribution — know exactly what causes what	Slower overall optimization
Simultaneous (most businesses)	Faster implementation	Harder to attribute which signals produce which effects

Most businesses will implement both simultaneously — and that’s reasonable. The framework’s principles apply regardless of sequencing.

The value of the GaryOwl.com approach: by isolating variables, we can now attribute effects to specific signals. This data informs the methodology in ways that mixed-signal implementations cannot.

---

Key Takeaways

1. Dark AI Traffic is real and measurable — look for Unassigned channel growth with low engagement rates
2. Authority gaps affect new and established businesses — new sites face Google Sandbox, but established companies often have fragmented identities or suboptimal digital authority they’re unaware of
3. Technical SEO is the shared foundation — structured data, schema, and site architecture serve both Google and AI systems; Authority Intelligence builds on this base
4. AI systems appear to evaluate content differently — the Phase 1 data suggests content substance can partially compensate for missing credentials
5. Reverse engineering + ML optimization yields actionable insights — systematic analysis of citation patterns combined with machine learning reveals optimization opportunities that generic best practices miss
6. Entity Coherence is foundational — inconsistent identities across platforms fragment authority; a Single Source of Truth strengthens it
7. Well-structured, naturally written content performs well for AI — long-form content with clear structure appears to be preferred for citation
8. Phase 1 supports the AI-first approach — but Phase 2 will test whether adding E-E-A-T signals produces compounding effects
9. Google still operates on traditional signals — the December 2025 update confirmed this
10. Server logs reveal what GA4 cannot — essential for understanding actual AI crawler activity
11. Phase 2 continues — the methodology evolves with new data

---

FAQs

Common questions about Dark AI Traffic, Authority Intelligence, and the GaryOwl.com experiment — with direct answers.

What’s the difference between GEO and Authority Intelligence?

GEO (Generative Engine Optimization) is the broad field of optimizing content for AI search systems. Authority Intelligence is a specific methodology within GEO, developed at GaryOwl.com. While GEO encompasses all tactics for AI visibility, Authority Intelligence operates at the boundary of SEO — complementing it with a focus on Technical SEO foundations (structured data, schema, site architecture) while extending into AI-specific optimization. It’s particularly valuable for addressing authority gaps in both new sites and established businesses with suboptimal digital presence.

Why was author identity omitted initially?

GaryOwl.com started as a personal project — a space to unite decades of professional experience with a passion for writing and AI. It wasn’t launched as a strategic business venture with optimized author pages and LinkedIn integration; it was simply a platform to explore and create.

This natural starting point created an interesting observation opportunity: I could see how AI systems respond to content substance alone, without the influence of author credentials or brand reputation. The findings emerged from this context rather than from a controlled experiment.

Did the approach work?

Phase 1 produced clear results. AI systems discovered and cited the content without traditional credentials. The methodology generated measurable AI visibility. Google traffic declined after the December 2025 update — unsurprising given the absence of E-E-A-T signals. Phase 2 will test what happens when we add those signals to a site with established AI authority.

What type of content works best for AI visibility?

AI systems require substantive, structured content to generate accurate citations. Long-form, well-organized text that RAG systems can process works well. The key factors appear to be: natural language (no keyword stuffing), clear structure, credible sources, and factual depth. Clickbait titles and thin content don’t help.

Should businesses ignore SEO entirely?

No. Authority Intelligence complements SEO — it doesn’t replace it. The overlap is Technical SEO: structured data, schema markup, crawlability, and site architecture. These elements serve both Google and AI systems.

The Phase 1 data suggests that intensive E-E-A-T development during the Google Sandbox period may have lower returns than AI-focused content. But basic Technical SEO remains essential as the foundation for both strategies.

What is the Google Sandbox?

The observation that new domains struggle to rank in Google for 12-24 months, regardless of content quality or SEO efforts. Google appears to require a trust-building period before granting significant organic visibility.

Do AI systems also favor established authorities?

There is evidence suggesting they may. Sites with extensive backlink profiles, brand recognition, and publishing history appear to have advantages in AI citation — not just in Google rankings.

But “established” doesn’t automatically mean “optimized.” Many established businesses have suboptimal digital authority: fragmented identities across platforms, inconsistent information, outdated presences, or technical issues they’ve never audited. Authority Intelligence helps both new entrants build from scratch and established businesses sharpen their existing authority.

What is Entity Coherence and why does it matter?

Entity Coherence refers to maintaining consistent identity information across all digital platforms. AI systems aggregate data from multiple sources to build an understanding of entities (companies, people, brands). If your LinkedIn says one thing, your website another, and your Google Business Profile something else, AI systems receive conflicting signals — fragmenting rather than building authority. A Single Source of Truth, with all platforms aligned, strengthens entity recognition and citation likelihood.

How do I know if AI is citing my content?

Test manually in ChatGPT, Perplexity, and Claude. Use tools like Profound or Otterly.ai. Check server logs for user-initiated AI requests (ChatGPT-User, MistralAI-User).

When should I start focusing on Google?

This is what Phase 2 will help answer. The working hypothesis: after establishing AI visibility (6-14 months), adding E-E-A-T signals may produce compounding effects. We’ll document the results.

What’s next for GaryOwl.com?

Phase 2: adding author bios, credentials, About page, and LinkedIn integration. We’ll measure whether strengthening E-E-A-T signals on a site with existing AI authority produces the expected effects — or something different.

Can my business use this framework?

The Authority Intelligence methodology is outlined in this article. The approach is based on Phase 1 data — not guaranteed outcomes. Businesses can evaluate and implement it independently. As the octyl® toolchain develops, additional resources may become available.

What is the relationship between GaryOwl.com and octyl®?

GaryOwl.com is the platform where the Authority Intelligence methodology was developed and tested — starting as a personal project and evolving into a systematic study. It’s part of the octyl® toolchain. octyl® is a registered trademark.

Is Authority Intelligence only for new websites?

No. GaryOwl.com happened to be a new domain — it started as a personal project, not a business venture. This created a natural observation opportunity for how AI systems respond to content without legacy authority signals. But the methodology applies broadly:

• New sites benefit from building AI visibility during Google Sandbox
• Established businesses benefit from auditing and sharpening their digital authority
• Companies with fragmented identity benefit from Entity Coherence principles
• Any organization can benefit from Technical SEO foundations that serve both Google and AI systems

Many established businesses have authority gaps they’re not aware of — inconsistent information across platforms, outdated profiles, or poor entity coherence. The methodology helps identify and address these gaps.

---

Conclusion

This article documents a transition point, not a conclusion.

For 14 months, GaryOwl.com evolved from a personal project — born from the desire to unite 25 years of cross-industry expertise with a passion for language and AI — into a systematic study of AI citation behavior. The Phase 1 data is clear: AI systems discovered and cited the content. The approach worked.

The data also shows limits: Google’s December 2025 Core Update demonstrated that traditional search still operates on traditional trust signals.

But the most interesting findings are ahead of us.

Phase 2 begins now: We’re adding author credentials and trust signals to a site with established AI authority. Will the effects compound? Will Google traffic recover while AI citations remain stable? Will there be interference between the two systems?

These questions don’t have answers yet. Phase 2 will provide the data.

What we’re sharing:

The Authority Intelligence methodology is part of the octyl® toolchain, a framework for developing customized applications. Not a finished product with guaranteed outcomes, but a tested approach with empirical data behind it.

For businesses facing the Google Sandbox problem, or those seeking AI visibility without years of traditional authority building, the Phase 1 data suggests this approach merits consideration.

For what happens next, follow the progress. We’ll document Phase 2 as rigorously as Phase 1.

Phase 2 begins.

---

This article represents the intersection of Generative Engine Optimization (GEO) research, empirical experimentation at GaryOwl.com, and the Authority Intelligence methodology.

---

References

Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., Narasimhan, K., & Deshpande, A. (2023). GEO: Generative Engine Optimization. arXiv preprint. Foundational academic research on optimizing content visibility in generative engine responses. https://arxiv.org/abs/2311.09735

Ahrefs (2024). Google Sandbox: Does Google Really Hate New Websites? Ahrefs Blog. Investigation of the alleged Google Sandbox filter affecting new domain rankings. https://ahrefs.com/blog/google-sandbox/

Ahrefs (2024). How Long Does It Take to Rank in Google? Ahrefs Blog. Data-driven analysis of ranking timelines and SEO velocity benchmarks. https://ahrefs.com/blog/how-long-does-it-take-to-rank-in-google/

Amsive Digital (2025). Digital Marketing & SEO Agency. Homepage of Lily Ray’s agency, known for E-E-A-T and algorithm update analysis. https://www.amsivedigital.com/

Anthropic (2025). Does Anthropic Crawl Data from the Web? Anthropic Support. Official documentation on ClaudeBot web crawler and opt-out mechanisms. https://support.anthropic.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-operators-block-the-crawler

Bailyn, E. (2025). Generative Engine Optimization (GEO) Strategy Guide. First Page Sage. Comprehensive practitioner methodology for AI search optimization. https://firstpagesage.com/seo-blog/generative-engine-optimization-geo-strategy-guide/

Barnard, J. (2025). Answer Engine Optimization. Kalicube. Entity definition and framework for AEO methodology coined by Jason Barnard. https://kalicube.com/entity/answer-engine-optimization/

Cloudflare (2025). Radar 2024 Year in Review. Cloudflare Radar. Comprehensive analysis of global internet traffic including AI crawler activity and bot traffic patterns. https://radar.cloudflare.com/year-in-review/2024

Google (2022). Our Latest Update to the Quality Rater Guidelines: E-A-T Gets an Extra E for Experience. Google Search Central Blog. Official announcement of E-E-A-T framework expansion. https://developers.google.com/search/blog/2022/12/google-raters-guidelines-e-e-a-t

Google (2025). December 2025 Core Update. Google Search Status Dashboard. Real-time rollout status of the December 2025 algorithm update. https://status.search.google.com/incidents/DsirqJ1gpPRgVQeccPRv

Google (2025). Google Search’s Core Updates. Google Search Central. Official documentation explaining core algorithm updates and their impact. https://developers.google.com/search/updates/core-updates

Human Security (2025). Understanding the AI Ecosystem: Agents, Scrapers, and Crawlers. Human Security Blog. Technical analysis of AI bot traffic patterns and web ecosystem impact. https://www.humansecurity.com/learn/blog/ai-ecosystem-agents-scrapers-crawlers/

Indig, K. (2025). Kevin Indig – Growth Advisor. Personal website of growth advisor and former Shopify/Atlassian director, focusing on SEO and AI search strategy. https://www.kevin-indig.com/

iPullRank (2025). iPullRank – Technical SEO Agency. Homepage of Mike King’s agency specializing in technical SEO, content strategy, and AI optimization. https://ipullrank.com/

Kalicube (2025). Kalicube – Digital Brand Engineers. Platform for Brand SERP optimization and entity-based search strategy founded by Jason Barnard. https://kalicube.com/

Zhu, Y., et al. (2024). LLMs for Knowledge Graph Construction and Reasoning: Recent Capabilities and Future Opportunities. arXiv preprint. Comprehensive evaluation of LLMs for entity resolution and knowledge graph tasks. https://arxiv.org/abs/2305.13168

OpenAI (2025). GPTBot. OpenAI Platform Documentation. Technical specifications for OpenAI’s web crawler used for AI training and search. https://platform.openai.com/docs/gptbot

Profound (2025). Profound – AI Search Visibility Platform. Enterprise platform for monitoring and optimizing brand visibility in AI-powered search engines. https://www.tryprofound.com/

Schema.org (2025). Schema.org Vocabulary. Community-maintained structured data markup specifications for enhanced web content understanding. https://schema.org/

Solís, A. (2025). Aleyda Solís – International SEO Consultant. Personal website of international SEO consultant and SEOFOMO newsletter creator. https://www.aleydasolis.com/

---

Article Metadata

Title: Authority Intelligence: Building AI Citation Authority From Zero

Author: Manuel

Published: December 22, 2025

Words: ~9,300

Flesch Score: 27 (Professional/Dense – Technical Audience)

Sources Verified: 9 research sources

Last Update: December 22, 2025

Next Review: January 2026

Research & Review Tool: octyl® Authority Intelligence Framework

---

Copyright & Terms of Use

© 2025 octyl®. All rights reserved.

📧 Contact for Feedback, Corrections, or Collaborations:

gary@octyl.io | Mastodon | Bluesky

---

Related Articles

• Gary Owl’s Strategic Authority Intelligence
• How to Rank High on AI Search Engines

---

This article documents the transition from Phase 1 to Phase 2. The Authority Intelligence methodology is part of the octyl® toolchain.