1. Executive Summary

The Bottom Line

On May 19, 2026, Google announced the most significant changes to Search in over 25 years. The shift from keyword-based ranking to entity-based AI citation fundamentally changes how businesses get found online. ClickRadius was built to solve exactly this problem — and the market just accelerated toward the product.

48%
of queries now show AI Overviews — and expanding1
-33%
publisher organic traffic decline in past year2
78%
of legal service queries trigger AI Overviews3
5x
higher conversion rate from AI referral traffic vs traditional organic4

Key Findings

  1. ClickRadius is ahead of the curve. The product already scores for AI citation readiness, builds entity authority across multiple platforms, monitors citations across 6 AI engines, and auto-deploys fixes. These are exactly the capabilities businesses now desperately need.
  2. On-site SEO is now just the foundation. Industry data shows the majority of what drives AI citations happens outside your website — entity building, directory presence, multi-platform authority, and external signals. ClickRadius is one of the only platforms that covers both on-site and off-site.
  3. 89.8% of brands have zero AI search mentions.5 This is the window. Early movers who build entity authority and optimize for AI citation now will capture disproportionate market share before competitors catch up.
  4. The platform is technically deep. ClickRadius is not a reporting dashboard — it is a full AI-citation intelligence platform with Bayesian learning, real-time 6-engine monitoring, automated entity building, statistical outcome measurement, and auto-deployment. Patent Pending 64/063,349 with 67 claims.

2. Google's Algorithm Revolution — What Changed

2.1 Google I/O 2026 (May 19, 2026)

Google VP of Search Elizabeth Reid called this "the biggest upgrade to our Search box in over 25 years." CEO Sundar Pichai separately described it as "our biggest upgrade to Search ever." The announcements fundamentally restructure how search works:

AI Mode Becomes the Default

Google's experimental AI Mode is now the default search experience for all users globally. Instead of typing a query and getting 10 blue links, users get a conversational AI interface powered by Gemini 3.5 Flash that synthesizes answers from across the web. Traditional search results are still accessible but are now secondary.

Search Box Redesigned

The search box itself has been redesigned with AI-powered suggestions replacing traditional autocomplete. Queries are interpreted with semantic understanding rather than keyword matching. The old approach of "rank for keyword X" is giving way to "be the authoritative entity that AI cites for topic X."

AI Overviews Expansion

AI Overviews (formerly Search Generative Experience) now appear on 48% of all queries, up from approximately 15% in early 2026.1 This rate continues to expand as Google rolls out AI Mode globally.

Information Agents

Google introduced Information Agents — autonomous AI agents that can monitor topics 24/7 on behalf of users. They run searches, compare options, and deliver summaries without the user ever visiting a website. Launching this summer for Google AI Pro and Ultra subscribers, this represents a further shift away from click-based traffic.

Critical Shift: Google is no longer primarily a referral engine that sends traffic to websites. It is becoming an answer engine that synthesizes information and only cites sources when they provide genuine expertise or authority that the AI cannot replicate on its own.

2.2 The Impact — By the Numbers

Metric Before (Pre-2026) After (May 2026) Change
Position #1 Click-Through Rate6 27% 11% -59%
Zero-Click Searches (Overall)7 ~45% 60% +33%
Zero-Click in AI Mode8 N/A 93% New
Publisher Organic Traffic2 Baseline -33% Severe
AI Overview Appearance Rate1 ~15% 48% +220%
Legal Services AI Overview Rate3 ~30% 78% Among Highest
Brands with Zero AI Mentions5 N/A 89.8% Opportunity
AI Referral Conversion Rate4 2.8% (organic) 14.2% (AI referral) +407%
Click Boost from AI Citation9 Baseline +35% organic, +91% paid Major Lift

2.3 What Google Now Values

The shift can be summarized as moving from keyword relevance to entity authority:

Declining in Importance

  • Keyword density and exact-match optimization
  • Link quantity (number of backlinks)
  • Page count / content volume
  • Traditional on-page SEO signals alone
  • Desktop-only optimization

Rising in Importance

  • Entity authority — Knowledge Graph presence, structured data, consistent NAP across platforms
  • Expertise signals — E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness
  • Structured data / Schema markup — machine-readable content that AI engines can parse
  • Citation-worthiness — content AI engines want to reference and attribute
  • Multi-platform entity consistency — same business identity across directories, social, and knowledge bases
  • Original expertise — first-hand experience and novel analysis AI can't synthesize on its own

2.4 The New Citation Economy

A new metric has emerged: AI citation share. Being cited in an AI Overview doesn't just drive traffic — it drives dramatically higher-quality traffic:

The Opportunity: With 89.8% of brands having zero AI search mentions,5 the competitive landscape is wide open. Businesses that build entity authority and optimize for AI citation NOW will establish positions that are extremely difficult for competitors to displace later — AI engines develop "trust patterns" that compound over time.

2.5 Industry Expert Perspectives

Leading SEO strategists have been consistent in their assessment of this shift:

Lily Ray, VP of SEO & AI Search at Amsive, has emphasized that traditional SEO is not dead — but it is now necessary rather than sufficient. On-site optimization remains the foundation, but entity authority and AI-engine optimization are where the differentiation happens. She advocates a hybrid SEO + GEO (Generative Engine Optimization) approach as the new standard.

Rand Fishkin, co-founder of SparkToro, has argued that the websites that will thrive are those providing information AI cannot easily synthesize on its own — original research, genuine expertise, and novel analysis. Content that merely rephrases what's already available will be absorbed into AI answers without attribution.

3. The On-Site vs Off-Site Paradigm Shift

This is the single most important concept for understanding why traditional SEO tools are failing businesses in 2026 — and why ClickRadius takes a fundamentally different approach.

3.1 The Old Model: On-Site Was Everything

For 20 years, SEO focused almost entirely on what happens on your website: keywords, meta tags, page speed, mobile responsiveness, content quality, internal linking. If you optimized your site well enough, Google would rank you.

Most SEO tools are still built for this model. They scan your site, generate a report, and give you a checklist of on-site fixes. The implicit assumption: fix your website and the rankings will follow.

3.2 The New Model: Off-Site Is Where AI Citations Are Won

AI engines don't just look at your website to decide whether to cite you. They look at your entire digital footprint — your entity presence across the web:

~18%

On-Site Optimization

Schema markup, meta tags, content quality, technical SEO, page speed, mobile optimization

Important — but no longer sufficient on its own

~82%

Off-Site Entity Authority

  • Directory presence — Data Axle, Foursquare, Bing Places, and industry-specific directories
  • Knowledge Graph entity — Google KG, Wikidata, and linked data sources
  • Social authority signals — LinkedIn, Reddit, and platform-specific content
  • Citation consistency — NAP (Name, Address, Phone) matching across all platforms
  • Third-party mentions — press, reviews, industry publications, forums
  • Cross-platform entity verification — consistent identity signals that AI engines can validate

LinkSurge's analysis of AI citation sources found that 91% of sources cited in AI-generated answers come from third-party platforms, not the business's own website. Stacker's research showed sites with strong entity presence saw 34% AI citation rates compared to just 8% for sites relying on on-site optimization alone.

The Gap in the Market: Most SEO tools still focus almost exclusively on on-site optimization — the ~18% of the opportunity. They can tell you your meta tags are wrong, but they can't build your entity presence on Data Axle, verify your Knowledge Graph entry, or monitor whether AI engines are citing you. This is the gap ClickRadius was built to fill.

3.3 Why This Matters for Every Local Business

Consider a local PI attorney. Traditional SEO tools will:

But none of that tells the attorney whether:

ClickRadius does all of this. It's the difference between checking your house's foundation and actually looking at the neighborhood, the street, the city records, and what everyone is saying about your address. AI engines look at the whole picture — and ClickRadius monitors and builds the whole picture.

4. ClickRadius — Platform Overview

ClickRadius is not a reporting tool — it is a full AI-citation intelligence platform. It scores, monitors, builds, and fixes. It covers on-site optimization and off-site entity authority. And it's protected by Patent Pending 64/063,349 with 67 claims.
48
API Route Modules
61
Database Tables
21
Automated Cron Jobs
6
AI Engines Monitored
5
Entity Build Platforms
67
Patent Claims

4.1 On-Site Capabilities

Capability What It Does
Scoring Engine (6 weighted categories) Schema Markup (22%), Content Quality (18%), AI Readiness (18%), Meta Tags (18%), Technical SEO (14%), Security (10%)
Schema Analyzer Parses JSON-LD, Microdata, and RDFa; recognizes 30+ Schema.org types; scores depth, industry relevance, and completeness
Auto-Fix Engine Generates AND deploys fixes via WordPress REST API, JavaScript snippets, or Cloudflare Workers — not just recommendations
Content Engine AI-powered article generation with GEO scoring, E-E-A-T evaluation, fact-checking, featured image generation, editorial workflow, and WordPress publishing
Search Agent Scanner 6-dimension analysis for Google Information Agent readiness: semantic HTML, accessibility, structured data, crawlability, action readiness, content originality

4.2 Off-Site Capabilities (the differentiator)

Capability What It Does
Entity Orchestrator Builds entity presence across Data Axle, Foursquare, Bing Places, LinkedIn, and Reddit via modular adapter architecture
Citation Monitor Real-time monitoring across 6 AI engines (Claude, ChatGPT, Gemini, Perplexity, Grok, Copilot) with velocity tracking and cross-engine consensus
AI Overview Tracker Monitors which keywords trigger AI Overviews, detects client citation presence and position, computes citation share vs competitors
Knowledge Panel Monitor Google Knowledge Graph + Wikidata cross-reference with completeness scoring and trend analysis
Entity Verification Cross-platform consistency checker — verifies entity data across Google KG, Wikidata, and 10 social platforms with discrepancy severity scoring
Outside Signals Scanner Scans 20+ platforms across 5 signal categories (brand mentions, reviews, community, media, social) with composite scoring
Conversational Query Analyzer Transforms keywords into AI Mode conversational variants, compares citation likelihood between formats, identifies content gaps

4.3 Intelligence & Automation

Capability What It Does
Thompson Sampling Strategy Engine Bayesian multi-armed bandit that learns which optimization strategies work for each site — adapts based on measured outcomes, not one-size-fits-all
Outcome Measurement Statistical significance testing (Welch's t-test) correlating optimization actions with citation rate changes
Content Distribution Automated distribution of generated content to connected platforms, extending entity authority beyond the client's own site
White-Label System Full multi-tenant branding with custom company name, logo, colors, and portal title — resellers present ClickRadius as their own platform
Reseller Infrastructure Self-registration, client/site management, automated onboarding (scan + 3 draft articles on site add), branded reports, Stripe billing integration

4.4 Competitive Advantages

  1. Entity Building is built in, not bolted on. Competitors like Rank Monster offer scoring but not entity construction. Scoring a problem you can't fix is a report, not a solution.
  2. Citation monitoring across ALL major AI engines. Most competitors monitor 1-2 engines. ClickRadius monitors 6, because each engine has different citation patterns and data sources.
  3. Auto-fix deploys changes, not just recommendations. Most tools stop at "here's what's wrong." ClickRadius fixes it — automatically deploying schema, meta tags, and security headers.
  4. Learning strategy engine. The Thompson Sampling bandit gets smarter with every measurement cycle. This is not static — it adapts per site.
  5. Statistically validated outcomes. Not "we think this helped" — Welch's t-test with p-value significance thresholds and confounding variable detection.
  6. Patent protection. Patent Pending 64/063,349 with 67 claims creates a defensible moat covering the integrated approach of scoring + entity building + AI citation monitoring + auto-remediation.

5. ClickRadius — Technical Architecture & Specifications

This section provides the detailed technical specifications of the ClickRadius platform for technical evaluation. Every capability described here is built, deployed, and running in production.

5.1 Scoring Engine — 6-Category Weighted Analysis

The scoring engine crawls a client's site and runs parallel analyzers across 6 weighted categories. The composite score (0-100) represents overall AI-citation readiness:

Schema Markup
22%
Content Quality
18%
AI Readiness
18%
Meta Tags
18%
Technical SEO
14%
Security
10%

Schema Analyzer (22% weight)

Parses three structured data formats: JSON-LD (application/ld+json script blocks), Microdata (itemscope/itemprop attributes), and RDFa (typeof/property attributes). Recognizes 30+ Schema.org types across three tiers:
  • Business types: LocalBusiness and 20+ subtypes — LegalService, Dentist, Restaurant, Store, MedicalBusiness, AutoDealer, RoofingContractor, Electrician, LocksmithService, MovingCompany, Hotel, ChildCare, Florist, TaxiService, and more
  • Organization types: Organization, Corporation, GovernmentOrganization, NGO, EducationalOrganization, MedicalOrganization, SportsOrganization, Airline
  • Content types: Article, BlogPosting, WebPage, WebSite, CreativeWork, Event, VideoObject, FAQPage, HowTo, BreadcrumbList, Product, Service
Scoring factors: LocalBusiness presence (+25), industry-specific subtype (+15), Organization fallback (+10), FAQPage (+8), HowTo (+5), BreadcrumbList (+5), schema depth (nested properties), content schema bonus (+5).

GEO Score (Generative Engine Optimization)

A dedicated metric evaluating how likely content is to be cited by AI engines. Three equally-weighted dimensions, each capped:

GEO Score = min(100,
  (min(quotations, 5) / 5) × 33 +
  (min(statistics, 8) / 8) × 33 +
  (min(citations, 5) / 5) × 34
)
Quotation detection: Counts <blockquote>, <cite>, and <q> HTML elements.

Statistics detection (7 regex patterns): Percentages with context (e.g., "X% of/increase/decrease"), dollar amounts with context ("$X per/in/worth"), numeric counts ("X years/clients/cases"), approximations ("over/more than X"), compound counts ("X+ years"), ratings ("rated/score X.X"), and multipliers ("Xx faster/more").

Citation detection: Text patterns ("according to", "source:", "research shows", "study found") plus external links with citation-related anchor text.

Structural bonus signals: Table presence, heading depth (H2+H3 hierarchy), FAQ section detection, ordered lists, definition lists — tracked but not scored, used for content improvement recommendations.

5.2 Citation Monitor — 6-Engine Real-Time Intelligence

The Citation Monitor sends actual queries to each AI engine's API, analyzes responses for brand mentions and URL citations, computes sentiment, and tracks velocity over time.

Supported AI Engines

Engine Integration Method Model
Claude (Anthropic) Official SDK (@anthropic-ai/sdk) Haiku (fast, cost-effective)
ChatGPT (OpenAI) Official SDK (openai) GPT-4o-mini
Gemini (Google) Official SDK (@google/generative-ai) Gemini Pro
Perplexity REST API (api.perplexity.ai) Sonar
Grok (xAI) OpenAI-compatible SDK (api.x.ai) Grok
Copilot (Microsoft) Web automation pipeline Copilot

Citation Score Formula

baseScore = min(100,
  mentionRate × 0.5 +
  (cited > 0 ? 30 : 0) +
  avg_confidence × 20
)

citationScore = min(100, baseScore × 0.7 + weightedMentionRate × 0.3)
Confidence scoring: Brand mentioned + URL cited = 0.95 | Brand mentioned, no URL = 0.70 | Not mentioned = 0.30

Engine-specific weights: Each AI engine has a customizable scoring profile stored in engine_profiles. Example: Perplexity citations are weighted 1.5x (source-heavy engine), Gemini schema signals weighted 1.5x (structured-data sensitive). This means the system adapts its scoring to how each engine actually behaves.

Sentiment analysis: 11 positive keywords (recommend, excellent, top, best, leading, trusted, great, quality, reliable, expert, outstanding) and 9 negative keywords (avoid, poor, bad, worst, scam, unreliable, expensive, overpriced, disappointing) with weighted comparison.

Citation Velocity Tracking

Weekly snapshots stored in citation_velocity table. Rate-of-change computed via SQL LAG() window function across consecutive periods. Trend classification:
Accelerating (>5% increase) | Growing (>0%) | Stable (>-5%) | Declining (>-15%) | Dropping (<-15%)

Cross-Engine Consensus Scoring

consistencyScore = max(0, 100 - mentionStdDev × 2)

consensusScore = consistencyScore × 0.3 + crossPlatformRate × 0.4 + sentimentAgreement × 0.3
Measures how consistently the brand appears across all 6 engines. High consensus means the brand's entity authority is broadly recognized. Low consensus identifies engine-specific gaps — e.g., a business visible to Gemini but invisible to ChatGPT, indicating a Knowledge Graph strength but Wikipedia/Reddit weakness.

Per-Engine Optimization Intelligence

Each engine profile includes hardcoded optimization tips based on known citation behavior patterns. Examples:
  • ChatGPT: "Wikipedia/Wikidata presence appears in 26-48% of ChatGPT top-10 citations"
  • Gemini: "Structured data produces 3.1x citation lift in Gemini responses"
  • Perplexity: "FAQ schema correlates with 41% citation rate vs 15% without"
These insights drive targeted recommendations — if a business scores low on ChatGPT but high on Gemini, the system recommends Wikipedia/Wikidata presence rather than generic SEO improvements.

5.3 Entity Intelligence Layer

Entity Orchestrator — Automated Entity Building

The Entity Orchestrator crawls client sites, extracts JSON-LD schema data (recognizing 14 entity types), and builds entity presence across external platforms using a modular adapter architecture. Each platform has its own adapter module — adding a new platform means adding one file, no changes to the orchestrator itself.

Platform adapters:
PlatformFunctionIntegration
Data AxleBusiness directory (feeds Google, Siri, Cortana)API submission
FoursquareLocation data (feeds Apple Maps, Uber, Snap)API submission
Bing PlacesMicrosoft search ecosystemAPI submission
LinkedInProfessional authority signalContent publishing API
RedditCommunity presence (top AI citation source)OAuth API + fallback

Build Plan Intelligence

The orchestrator doesn't blindly submit to every platform. It generates a prioritized build plan based on the Outside Signals score gaps:
Signal CategoryTrigger ThresholdAutomated Actions
Brand VisibilityScore < 40Submit to Data Axle, Foursquare, Bing Places
Social AuthorityScore < 50Publish to LinkedIn
Earned MediaScore < 30Distribute press release
Community PresenceScore < 40Publish to Reddit
Each action is logged as a correlation_event for later outcome measurement — the system tracks whether the entity building actually improved citation rates.

Knowledge Panel Monitor

Integrates two external APIs to assess Knowledge Panel presence:
  • Google Knowledge Graph Search API — entity lookup returning name, types, description, KG ID, relevance score, image, and URL
  • Wikidata API — entity search + claims extraction (P856: official website, P571: inception date, P17: country)
Completeness scoring (0-100):
SignalPoints
Google Knowledge Graph presence+25
KG has description+10
KG has detailed description+10
KG has image+5
KG has URL+5
Wikidata entity exists+15
On-site Organization/LocalBusiness schema+15
sameAs links in schema+10
Logo in schema+5
Founder/foundingDate in schema+5
Trend analysis compares current vs historical completeness scores (up to 12 months), classifying as improving (>5 delta), stable, or declining (<-5 delta).

Entity Verification — Cross-Platform Consistency

Compares entity data from 4 independent sources — on-site schema, Google Knowledge Graph, Wikidata, and social profiles — to detect inconsistencies that damage AI citation trust.

Social platform detection (10 platforms): Facebook, LinkedIn, Twitter/X, Instagram, YouTube, Yelp, BBB, Wikipedia, Wikidata, Crunchbase. Extracted from sameAs schema links and verified via HTTP HEAD requests.

Consistency checks: Name match (site vs KG, site vs Wikidata), URL match (site vs KG, site vs Wikidata), social profile accessibility (HTTP 200 response).

Discrepancy severity: Critical (name mismatch across Google KG), High (URL mismatch or broken social profiles), Medium (missing Wikidata entry or incomplete sameAs links). Each discrepancy includes specific fix instructions.

5.4 Outside Signals Scanner — 20+ Platform Intelligence

Comprehensive off-site signal analysis across 5 weighted categories, scanning 20+ platforms to measure brand presence outside the client's own website.
Signal Category Weight Platforms Scanned Key Metrics
Brand Visibility 25% Google Search (via DataForSEO API), Bing Search, Google News, Google Knowledge Graph Mention count, KG panel presence, search result diversity
Reviews 20% Google Business Profile, G2, Capterra, Trustpilot, Yelp, BBB Platform count, average rating, total review volume
Community 20% Reddit (OAuth API), Quora, Stack Overflow, Hacker News (Algolia API) Mention count, subreddit diversity, upvote engagement, HN story count
Media 20% News publications, guest posts, podcasts, syndication networks, Wikipedia News mentions, external link count, Wikipedia presence
Social 15% LinkedIn, Twitter/X, YouTube, Facebook Platform presence, sameAs schema count, channel discovery, diversity score
Scoring examples:
Brand Visibility: 15+ mentions = 40pts, KG panel = +30pts, any signals = +10pts, 2+ signal types = +20pts (max 100)
Reviews: Each platform = 10pts (max 40), avg rating ≥4.5 = 30pts, 100+ reviews = 30pts (max 100)
Community: Reddit is weighted highest because it is the #1 cited source across AI engines. Subreddit diversity and upvote engagement are tracked separately.
Media: Wikipedia presence scores up to 35pts — research shows Wikipedia/Wikidata appears in 26-48% of ChatGPT's top-10 citations.

Composite: overall = brand × 0.25 + review × 0.20 + community × 0.20 + media × 0.20 + social × 0.15

5.5 AI Overview Tracking & Search Agent Optimization

AI Overview Tracker

Monitors which client keywords trigger Google AI Overviews and whether the client is cited. Uses AI simulation (Gemini Flash primary, Claude Haiku fallback) to predict AIO behavior for each tracked keyword.

Per-keyword analysis: AIO trigger likelihood, predicted citation sources (3-5 per query), client domain presence, citation position, and optimization tips.

Batch processing: Up to 20 keywords per batch, ordered by priority DESC, then least-recently-checked first. Automated weekly checks via cron job.

AIO Visibility Score:
aioVisibilityScore = min(100,
  citationRate × 0.6 +
  (aioTriggerRate > 50 ? 20 : aioTriggerRate × 0.4) +
  (clientCited > 0 ? 20 : 0)
)
Citation share analysis: Tracks competitor domains appearing in AIO citations, computing average position and share percentage. Identifies which competitors are winning the AI citation war for each keyword.

Search Agent Optimization Scanner

Evaluates readiness for Google's new Information Agents across 6 weighted dimensions with 70+ individual checks:

Dimension Weight What It Checks
Semantic HTML 15% 14 semantic tags (nav, main, article, section, aside, header, footer, figure, figcaption, details, summary, time, mark, address), semantic-to-generic ratio, non-semantic interactive element penalties
Accessibility 15% Alt text coverage, ARIA landmarks, lang attribute, heading order validation (H1 first, no skipped levels), skip links, tabindex audit, ARIA-labeled elements
Structured Data 20% Critical types (Organization, LocalBusiness, WebSite), content types (Article, FAQPage, HowTo, Product, Service), agent-specific types (BreadcrumbList, SiteNavigationElement, SearchAction)
Crawlability 15% robots.txt parsing, Google-Agent user-agent blocking detection, sitemap presence, RSS feed detection, IndexNow header, canonical tags, noindex checks
Action Readiness 15% CTA elements, forms with proper labels, tel: links, mailto: links, contact page links — agents need to know what actions users can take
Content Originality 20% Author attribution, publication dates, about section, testimonials, statistics density, source citations, experience markers ("in our experience", "we've found"), minimum word count

5.6 Conversational Query Analyzer

As Google shifts to AI Mode, users type natural-language questions instead of keyword fragments. This analyzer evaluates how citation likelihood changes between traditional keyword queries and conversational variants.

Process:
  1. Takes a traditional keyword (e.g., "personal injury lawyer chicago")
  2. AI generates 4 conversational reformulations with intent classification (informational, transactional, comparative, local) and complexity rating (simple, moderate, complex)
  3. Each variant is run through a citation simulation — predicting which sources would be cited, citation likelihood (0-1), position (1-5), and content signals needed
  4. System compares keyword citation rates vs conversational citation rates
Insight generation:
  • Opportunity: Conversational queries cite the site more than keyword queries — lean into conversational content
  • Risk: Site is cited for keywords but NOT conversational variants — the shift to AI Mode will cost this client traffic
  • Trend: Quantified citation lift percentage between formats (e.g., "17% more likely to be cited in conversational queries")
  • Action: Specific content signals needed, frequency-ranked across all variants
Overall trend aggregation: Across all analyzed keywords, classifies the site as conversational_queries_favored, keyword_queries_favored, or balanced.

5.7 Content Engine — AI-Powered Article Generation Pipeline

The Content Engine generates citation-optimized articles through a 10-step quality pipeline. Every article must clear multiple quality gates before publication.

1
Site Facts Extraction — Crawls the client's site, extracts JSON-LD schema, title, H1 headings, about section, and business details to create a verified fact base
2
Article Generation — Claude Sonnet generates the article with strict GEO requirements: 5+ quotations (<blockquote>), 8+ statistics with sources, 5+ citations ("according to"), FAQ section (4-6 questions), comparison table, definitive opening statement, section-ending quotable summaries
3
GEO Score Check — Minimum threshold of 75 (out of 100). Articles scoring below are rejected and regenerated
4
Retry Loop — Up to 3 generation attempts with targeted improvement prompts. After 3 failures, flagged for human intervention
5
Fact Validation — Cross-references generated claims against the site facts base. No invented statistics, no contradictions with known business information
6
E-E-A-T Evaluation — AI scores the article on all 4 dimensions (Experience, Expertise, Authoritativeness, Trustworthiness) on a 0-25 scale each. Checks for original insight, comparison tables, and strong opening. Articles below thresholds are flagged for human review
7
Schema Enrichment — Full Article schema markup injected (author, publisher, datePublished, citations, keywords, abstract). FAQ schema generated from the FAQ section. Machine-readable quality signals for AI engines
8
Auto Fact-Check — Extracts verifiable claims via regex (statistics, "according to" citations, "the average" patterns). Sends up to 5 claims to the internal fact-check API. Sets verification status: verified, partial, or disputed
9
Featured Image Generation — DALL-E 3 generates a 1792×1024 professional featured image in abstract style, matched to the article's topic
10
WordPress Publishing — Publishes via WordPress REST API with full schema markup injected into content. Auto-detects SEO plugin (Yoast, RankMath, or AIOSEO) and writes meta fields to the correct plugin format. Triggers internal link analysis after publish
Editorial workflow: Draft → Review → Approve/Reject → Schedule → Publish. Sites with editorial_review = true require human approval before publication. Scheduled posts are published automatically by the hourly cron job.

Prompt injection protection: All user-supplied input is sanitized — strips newlines, backticks, double quotes, and known injection markers (ignore previous instructions, </system> tags) with enforced maximum length.

RSS feed: Public endpoint per site serving RSS 2.0 with Atom self-link, up to 50 published articles.

5.8 Strategy Engine — Bayesian Machine Learning

The Strategy Engine uses Thompson Sampling with a Beta-Binomial model — a Bayesian multi-armed bandit algorithm — to learn which optimization strategies work best for each site. This is not a static recommendation engine; it adapts based on measured outcomes.

How it works:
  1. 18 optimization strategies across 6 categories (schema, meta, GEO, content, technical, entity), each modeled as a "bandit arm" with Beta(α, β) prior distribution
  2. When deciding what to optimize, the engine samples from each arm's Beta distribution using Marsaglia-Tsang Gamma sampling + Box-Muller normal distribution
  3. Highest-sampled strategies are selected (top 3 by default). This naturally balances exploration vs exploitation — arms with less data get explored more
  4. After 14-28 days, the Outcome Measurement system evaluates whether citation rates improved
  5. Results update the arm's α (success) or β (failure) parameters — the distribution sharpens over time

Informative Priors by Category

Different strategy categories start with different prior beliefs based on industry knowledge:
CategoryPrior (α, β)Implied Success RateRationale
Schema(3, 2)60%Schema fixes have high success rates in citation improvement
Meta(3, 2)60%Meta tag improvements reliably improve discoverability
GEO(2, 2)50%GEO optimizations are newer, less certain
Content(2, 2)50%Content quality is high-impact but variable
Technical(2.5, 1.5)63%Technical fixes (speed, security) are reliable
Entity(1, 1)50%Entity building is the newest category — minimal prior bias, learn from data

Hierarchical Learning

blendFactor = min(1.0, total_pulls / observations_needed)

effectiveAlpha = blendFactor × alpha + (1 - blendFactor) × parent_alpha
effectiveBeta = blendFactor × beta + (1 - blendFactor) × parent_beta
Arms with limited data are blended with parent-level aggregates. 10% of each reward signal propagates to related arms with fewer observations. This means if "add FAQ schema" works well for dental sites, that signal partially informs "add FAQ schema" decisions for medical sites — cross-context learning without cross-context assumptions.

5.9 Outcome Measurement — Statistical Significance Testing

ClickRadius doesn't just say "we think this helped." It uses Welch's two-sample t-test — a standard statistical significance test — to determine whether optimization actions actually caused citation rate improvements.

Methodology:
  1. For each optimization event, define a pre-window (14 days before) and post-window (28 days after)
  2. Collect citation velocity data from both windows
  3. Run Welch's t-test to compare pre and post mention rates
  4. Check for confounding variables — were there overlapping optimization events in the post-window?
  5. Assign attribution confidence: High (no overlap, sufficient data, statistically significant p < 0.05), Medium (insufficient data or not significant), Low (overlapping events)
Welch-Satterthwaite degrees of freedom:
df = (v1/n1 + v2/n2)² / ((v1/n1)²/(n1-1) + (v2/n2)²/(n2-1))

Impact classification:
Positive: change > +10% | Neutral: -10% to +10% | Negative: change < -10%
Why this matters: Most SEO tools report correlation ("your score went up and traffic went up"). ClickRadius reports causation with statistical rigor — separating real impact from noise, controlling for confounding variables, and assigning confidence levels. This is the difference between "trust us" and "here's the math."

5.10 Auto-Fix Engine — Deploy, Verify, Revert

The Auto-Fix Engine doesn't just generate recommendations — it generates fixes AND deploys them to live sites. Four fix types across four deployment methods:
Fix Type What It Generates
Meta DescriptionAI-generated meta descriptions optimized for AI citation
Title TagAI-generated title tags with entity and topic focus
Alt TextAI-generated image alt text for accessibility and structured signals
Internal LinksAI-suggested contextual internal links for topic clustering
Deployment Method How It Works
WordPress REST APIAuto-detects SEO plugin (Yoast, RankMath, or AIOSEO), writes to correct meta field keys. Falls back to WP excerpt field if no SEO plugin detected.
JavaScript SnippetClient-side script injection for non-WordPress sites
Cloudflare WorkerEdge-level HTML rewriting — modifies response before it reaches the browser, zero client-side overhead
ManualGenerates fix only, human deploys. Used when automated deployment isn't configured.
Strategy-guided batch fixes: The batch endpoint integrates with the Thompson Sampling Strategy Engine — the bandit selects which fix types to run for each site, and each optimization is linked to a strategy_decision_id for outcome tracking.

Post-deployment verification: After deploying a fix, the system waits 60 seconds, re-crawls the live page with Cheerio, and confirms the fix is actually present on the rendered page. Records verified_at and verified_ok. This catches failed deployments before they're reported as done.

Revert capability: Every optimization stores before_value. One-click revert restores the original via the same deployment channel. No fix is irreversible.

Schema lock protection: Sites with schema_locked = true are blocked from all auto-fix operations. This prevents automated changes from overwriting carefully crafted custom schema.

5.11 Automation — 21 Scheduled Operations

ClickRadius runs 21 automated cron jobs on scheduled cadences, from hourly to monthly. Each job uses PostgreSQL advisory locks to prevent overlapping runs, and failed jobs are queued in a retry system with automatic re-execution.

Job Schedule Function
Citation MonitorDaily 4:00 AMCheck all active citation queries across 6 AI engines
Citation FeedbackDaily 5:30 AMAnalyze citation response patterns and generate optimization insights
Conversation FinderDaily 9:00 AMScan for new online conversations mentioning client brands
Content DistributionDaily 10:00 AMDistribute generated content to connected platforms
Publish ScheduledHourlyPublish articles whose scheduled time has arrived
Health CheckEvery 6 hoursVerify all monitored sites are accessible and responding
Engagement TrackingEvery 12 hoursTrack engagement metrics on distributed content
Freshness ScanEvery 3 daysRe-score sites with stale analysis data
Citation VelocityMonday 5:00 AMSnapshot weekly citation rates for trend analysis
Entity BuildTuesday 7:00 AMExecute entity build plans across connected platforms
Weekly Auto-FixWednesday 5:00 AMRun strategy-guided batch fixes on eligible sites
AIO TrackingWednesday 3:00 AMCheck AI Overview presence for all tracked keywords
Outcome MeasurementFriday 8:00 AMCorrelate optimization actions with citation changes (Welch's t-test)
Outside SignalsSunday 6:00 AMFull 20+ platform off-site signal scan
Entity Check1st of monthVerify entity consistency across Google KG, Wikidata, and social
Knowledge Panel1st of monthCheck Knowledge Graph and Wikidata presence
Entity Verification1st of monthCross-platform entity consistency audit
Report Delivery1st of monthGenerate and email branded monthly reports
Agent Readiness15th of monthSearch Agent optimization scan (6 dimensions)
Conversational Query1st & 15th of monthAnalyze conversational vs keyword citation differences
Retry QueueEvery 2 hoursRe-execute failed jobs with error tracking
Advisory lock system: Each job has a unique lock ID. Before execution, the job acquires a PostgreSQL advisory lock — if a previous run is still in progress, the new firing skips silently. This prevents the kind of data corruption that occurs when two identical scans run simultaneously.

Retry queue: Failed jobs are inserted into scheduler_retry_queue with error details. The retry job runs every 2 hours, re-attempting failed operations with the original parameters. After repeated failures, admin alerts are triggered via the notification system.

Activity logging: Every scheduled run is logged to activity_log with duration, result counts, and error details. The scheduler status dashboard shows all jobs with their last run time, next scheduled run, and health status.

5.12 White-Label & Reseller Infrastructure

ClickRadius includes a complete white-label system built for resellers and agencies. The platform can be presented to end clients under any brand — no ClickRadius branding visible.

White-Label Branding

Multi-tenant branding resolution: Site → Client → Reseller profile. The system traverses the ownership chain to find the correct branding. Customizable elements:
  • Company name — replaces all "ClickRadius" references in the portal and reports
  • Logo URL — custom logo in portal header and PDF reports
  • Primary color — replaces accent colors throughout the portal UI
  • Portal title — custom dashboard title (default: "AI Readiness Dashboard")

Reseller Management System

Self-registration: Resellers create their own accounts with company name and credentials. Account created in pending_payment status until billing is configured.

Dashboard: 5 parallel aggregation queries showing client count, site count with health distribution (healthy/warning/critical), scan history, recent activity, and top-performing sites by score.

Automated onboarding: When a reseller adds a new client site, the system automatically:
  1. Triggers a full analysis scan in the background
  2. Generates content prompts for the site's business type
  3. Picks the top 3 keywords from auto-generated citation queries
  4. Generates 3 draft articles (1,200+ words, authoritative tone)
The reseller's new client has a score, recommendations, and content ready within minutes — not days.

Branded reports: Monthly PDF/email reports with the reseller's branding, including score history, citation summary by engine, recommendations, and score breakdown across all 6 categories.

Billing: Stripe-integrated subscription management with automated cancellation flow (30-day grace period, portal access until end of period).

Conclusion

Google's May 2026 changes are the most significant shift in search since Google itself launched. For businesses unprepared, this is a crisis. For businesses with the right tools, this is an unprecedented opportunity.

ClickRadius was built for this moment. While competitors are scrambling to add AI features to their keyword-ranking tools, ClickRadius was designed from the ground up for the AI-citation era. The complete platform — Bayesian strategy engine, 6-engine citation monitor, automated entity builder, statistical outcome measurement, 10-step content pipeline, and auto-fix deployment — covers every dimension of AI visibility that Google now prioritizes.

The on-site + off-site advantage is the differentiator. Every traditional SEO tool can tell you your meta tags are wrong. ClickRadius can do that and build your entity presence on Data Axle, monitor whether ChatGPT knows you exist, verify your Knowledge Graph entry, and deploy fixes automatically. That's the ~82% of the opportunity that other tools leave on the table.

The window of opportunity is now. With 89.8% of brands having zero AI search mentions,5 the first movers who build entity authority and optimize for AI citation will establish positions that compound over time. Every month of delay is a month competitors can use to establish their own AI trust signals.

The Platform

48 route modules. 61 database tables. 21 automated cron jobs. 6 AI engines monitored. 5 entity build platforms. Bayesian machine learning. Statistical significance testing. Auto-deployment with verification. White-label ready. Patent Pending 64/063,349 with 67 claims.

ClickRadius is not a tool — it's an AI-citation intelligence platform. The technology is built, the market is ready, and the window is open.

Sources

  1. 48% AI Overview rate: BrightEdge 9-Industry AI Search Tracker, 2026. Methodology: tracking AI Overview presence across commercial verticals.
  2. -33% publisher traffic: Press Gazette / Chartbeat global publisher traffic analysis, 2025-2026. US decline measured at 38%.
  3. 78% legal services rate: SE Ranking AI Overviews Research, 2026. Healthcare measured even higher at 88% (BrightEdge data).
  4. 5x conversion / 14.2% vs 2.8%: Opollo 2026 AI Search Benchmark Report, analysis of GA4 referral data from 312 B2B technology firms. Independently corroborated by RankScience analysis of 12M website visits.
  5. 89.8% zero AI mentions: Victorious Q1 2026 Quarterly Search Report, analysis of 177 brands across eight AI platforms.
  6. Position #1 CTR 27% to 11%: SISTRIX analysis of 100M+ keywords, measuring CTR impact of AI Overviews on organic position #1.
  7. 60% zero-click: Bain & Company, February 2025.
  8. 93% zero-click in AI Mode: Semrush / Seer Interactive analysis of 25.1M impressions in Google AI Mode.
  9. +35% organic / +91% paid from AI citation: Seer Interactive analysis of 3,119 informational queries across 42 organizations, tracking 25.1M organic and 1.1M paid impressions (June 2024 - September 2025).