91 published pieces. A first-party AI citation study, AI-search field guides, the 12 Vectors, every major AI engine, vertical playbooks, and the agency-pattern catalogue.
Generative Engine Optimization (GEO) is what we actually do. ChatGPT, Perplexity, Gemini, Microsoft Copilot, Apple Intelligence, and Google AI Overviews now answer queries before users click. Every piece below is built from primary sources, our own client data, or the published work of researchers we cite by name. No AI slop, no recycled SERP summaries.
Our May 2026 study of 1,732 real AI citations across nine Ontario cities, and the playbooks built from it. Why the four engines recommend different businesses, and how to get named.
Across 1,732 real citations in nine Ontario cities, only 16.3% of sources were shared by two or more engines.
ChatGPT130 of ChatGPT's local citations resolved to google.com. Your Google Business Profile is the lever.
GeminiGemini routed 384 citations through the vertexaisearch wrapper. What that opacity means for tracking.
ClaudeClaude leaned on curated directories, citing threebestrated.ca 116 times. The most editorial engine.
PerplexityThe most diversified engine, spreading citations across HomeStars, Opencare, BBB, and firm sites.
DentalAI routes best-dentist answers through Opencare and RateMDs, not practice sites. A YMYL-health read.
LegalThe Google-number-one firm is often absent from the AI answer. Directories win, within LSO rules.
RoofingHomeStars and ThreeBestRated dominate roofer answers. A trades, high-consideration read.
HVACFurnace-quote aggregators and HomeStars own the HVAC answer. Most contractors are invisible.
AgenciesHow agencies surface in AI search, FD included. Lists and awards move the needle; backlinks barely register.
DirectoriesReview directories quietly became the source every engine pulls from. Why, and what SMBs do about it.
ReviewsHomeStars topped every other domain in our scrape. The role of Yelp, RateMDs, and BBB.
GBPGBP is the single biggest lever for local AI visibility, especially for ChatGPT. The fields engines reuse.
PlaybookThe operational sequence to earn citations across all four engines, from one unambiguous entity outward.
ListingsA concrete map: GBP feeds ChatGPT, ThreeBestRated feeds Claude, the review layer feeds Perplexity.
MeasurementAI-citation share of voice, measured per engine. The four-engine method we ran on 1,732 citations.
EntityWhat an engine resolves about your business before it ranks anything. NAP, category, sameAs, and the MID.
TorontoThe largest, most splintered market. The four engines name almost no business in common.
HamiltonMid-market and directory-dominated. HomeStars, ThreeBestRated and local directories gate each trade.
AI vs MapsRanking in the Maps 3-pack does not carry to the AI answer. Two separate selection events.
Engine behaviour, the directories AI trusts, local-market patterns, and vertical playbooks, all built on the same 1,732-citation study.
The full methodology: the prompts, the four engines, nine cities, and how 1,732 citations became 583 cited hosts.
Rank vs CitationRetrieval and query fan-out, not rank, decide the AI answer. The evidence, with the numbers.
CitationsA backlink pools authority; an AI citation hands over the answer. What changes for off-site strategy.
Tie-BreakersWhen reviews and profiles match, verifiability is the tie-breaker. The five signals that settle it.
RAGHow grounding, not training data, decides whether an engine can name you. In plain language.
VarianceAI engines are non-deterministic. What run-to-run variance means for honest tracking and sample size.
Zero-ClickThe answer resolves without a click. Why being named in the answer becomes the goal.
CanadaStatistics Canada and CIRA data, .ca directory weighting, and why US guidance misleads Ontario SMBs.
ROILead value, citation share, and an honest time horizon. How to think about return without overpromising.
DirectoriesCited 116 times, Claude's favourite. The 50-point inspection model and how a business earns a place.
TradesHow HomeStars' invoice-verified reviews became the trust signal AI borrows for the trades.
HealthOpencare and RateMDs own AI answers for dentists and clinics. The trust model, with YMYL caution.
ComparisonChatGPT grounds in Google Maps; Claude in curated directories. Two engines, opposite habits.
SeasonalityFurnace and roofing intent spikes shift which sources AI cites. Prepare content ahead of the season.
Waterloo RegionIn the tech corridor, directory dominance is weaker and ChatGPT names local operators directly.
MississaugaCanada's sixth-largest city leans on GTA-wide directories. How its AI answers differ from Toronto's.
PlumbingEmergency intent, licensing as a verifiable trust signal, and why a local plumber can out-recommend a national brand.
ElectricalECRA/ESA licensing as machine-readable trust, and structuring pages for panel, EV, and generator questions.
Real EstateRECO and TRESA rules, brokerage-versus-agent entity signals, and why four engines name different agents.
Our 12-part framework for AI-search visibility. Each vector is its own discipline.
Diagnose, Anchor, Resonate, Embed, Cite, Structure, Distribute, Refresh, Cluster, Localize, Measure, Iterate. Full framework.
Vector 1Diagnose how ChatGPT, Perplexity, Gemini, and AI Overviews currently describe your brand. The baseline.
Vector 2Wikidata feeds ChatGPT, Gemini, Perplexity, and Apple Intelligence. One corrected entity propagates everywhere.
Vector 3From keyword research to prompt research. The conversational queries prospects send AI engines.
Vector 4The 40-60 word block, Aggarwal optimization tactics, why position-1 pages can lose AI visibility.
Vector 5Tier-1 citations lift AI selection probability by 89%. The four-tier hierarchy and the discipline.
Vector 665% of AI Mode citations and 71% of ChatGPT citations include structured data. The JSON-LD graph.
Vector 7Common Crawl, Reddit, industry press, and the two distribution layers AI engines actually pull.
Vector 876.4% of ChatGPT-cited pages were updated within 30 days. The substantive-update discipline.
Vector 9Clustered content earns 3.2x more AI citations than standalone. The pillar-plus-cluster architecture.
Vector 10AI Overviews trigger on 40%+ of local queries. The three pillars of local AI visibility.
Vector 11Brand Visibility, Citation Rate, AI Share of Voice. The measurement framework and the case story.
Vector 12Quarterly iteration that closes the methodology. AI engines update faster than annual planning.
Engine-by-engine playbooks. Each AI search surface has its own selection logic.
AI Overviews appear in 33-50% of queries. Layered framework from organic foundation to passage-level structure.
GeminiGemini is Google's most multimodal engine, pulling from YouTube, Knowledge Graph, and live web.
PerplexityPerplexity-referred traffic converts at 14.2% vs Google's 2.8%. The 6 signals that drive citation rate.
ComparisonDifferent architectures, audience sizes, citation behaviour, and source preferences. The honest comparison.
ComparisonPerplexity wins on deep research; Google wins on speed, local, and multimedia. What it means for visibility.
CopilotOptimizing for Bing in 2026 cites you in 3 AI engines simultaneously. The multiplier playbook.
AppleApple's 2026 Siri rebuild ('World Knowledge Answers') is Gemini-backed. The convergence playbook.
MechanicAI Overviews synthesize from 8-13 sources; Featured Snippets extract one verbatim. The unified optimization.
The honest 2026 framing of how AI search differs from traditional SEO and what it actually rewards.
Not the same job. Buyer journey, click economics, and optimization mechanics have all changed.
FoundationsSEO optimizes for clicks. GEO optimizes for citations inside AI answers. When each matters.
DataVerified, sourced AI search statistics. Adoption, citation behavior, market share, conversion. Cite these.
VoiceVoice queries are 30-60% longer than typed and structurally different. How voice search SEO works now.
SchemaArticle + Person + Organization + LocalBusiness + FAQPage + HowTo in a connected @graph. Copy-paste ready.
Knowledge GraphThe actual process to get into the Knowledge Graph: timeline, source consistency, schema, GBP, Wikidata.
WikidataWikidata feeds every major AI engine. Almost no SEO agency knows how to leverage it. Here's how.
TrustWhat each E means in 2026, how AI engines weight it, and how to demonstrate it operationally.
FD SystemOur named multi-agent orchestration system. Specialized agents pass work between each other under human oversight.
Industry-specific guides. Each vertical has its own SERP physics, regulatory layer, and lead-flow gap.
Vertical SEO playbooks for home inspectors, accountants, contractors, chiropractors, and seven more.
VerticalWhy home inspector SEO is structural. How to flank InterNACHI and earn AI Overview citations.
VerticalWhy CPA Ontario eats your traffic, why niche specialization is an SEO move, and the bookkeeper-handoff funnel.
VerticalCCO-compliant patient acquisition. Condition-specific SEO, insurance pages, MVA referral funnels, YMYL schema.
VerticalLead-broker math, Google Local Services Ads, past-customer referral funnel, trust signals that beat horror-story bias.
VerticalOut-trust the scam pool. Google Guarantee, brand-authorized signals, scam-defensive content.
VerticalOutflank 1-800-GOT-JUNK, College Hunks, and Junk King with estate-cleanout long-tail and neighbourhood pages.
VerticalOntario seasonal curve, material-specific keyword pools, pool-fence regulatory layer driving install jobs.
VerticalWin commercial accounts and route-density residential without burning cash on Angi resold leads.
VerticalWin fleet contracts and residential route density. Service-area schema, sub-niche pages, booking-flow gap.
VerticalBrand-specific organic traffic against Sears, Best Buy, manufacturer service. Brand-page matrix, authorization flow.
Documented patterns from real agencies, the contracts they use, and the numbers behind the harm.
Eight bad-agency patterns documented across BlitzMetrics, Stan Ventures, SEO Sherpa, with the verification checklist.
ContractsHow 12-month no-exit contracts work, why agencies write them, and the negotiation language that breaks them.
Contracts11 clauses every SEO agreement should carry, the 4 that should never appear, and the verification language.
ExitHow agencies hold websites, GBPs, and content hostage at exit. The prevention checklist that kills the pattern.
DecisionSix diagnostic conditions that justify firing. Breakup checklist and migration plan with the lock-in math.
ReportingWhy impressions, keyword count, and traffic charts hide underperformance. The four outcome metrics instead.
Service ModelWhy agencies sell fixed packages without auditing your business. The cost of generic work, and the diagnostic-first alternative.
StaffingHow SEO agencies actually staff their work, why concealment is the issue, and how to verify who is producing your deliverables.
StaffingAgencies marking up The HOTH and SEOReseller 200-300%. The disclosure problem, and how to verify the work.
ContentWhy agency 'AI content engines' produce a 28% organic traffic drop. The data, and the honest content alternative.