Original research report from Ottawa SEO Inc. — Canadian-domain sites earn just 6.8% of citation slots inside AI-generated answers for Canadian-intent queries — a gap of 41 points behind US-domain incumbents on the same query set.
**Canadian-domain sites earn just 6.8% of citation slots inside AI-generated answers for Canadian-intent queries — a gap of 41 points behind US-domain incumbents on the same query set.**
This report is the result of a systematic study by Ottawa SEO Inc. examining AI citation share across the Canadian business landscape. The data, methodology, and findings below are released under a free-attribution license — Canadian and international publishers, journalists, and researchers are welcome to cite this report with link attribution to ottawaseo.com.
For questions about methodology, data access, or interview availability with the report's author (Martin Vassilev), email research@ottawaseo.com.
- 6.8% — Share of AI-Overview citations earned by Canadian-domain sites for Canadian-intent commercial queries (n=8,420) - 47.9% — Share earned by US-domain sites on the same Canadian-intent query set - 23.1% — Share earned by Wikipedia, Reddit, and other community/reference domains - 11.4% — Share earned by Canadian government (.gc.ca, provincial) and educational (.edu) sites - Legal services: highest Canadian-domain citation share at 14.2% (driven by lawyer-directory aggregators) - SaaS and B2B technology: lowest Canadian-domain citation share at 1.9% - Perplexity cites 2.4× more Canadian-domain sources per answer than Google AI Overviews
### Finding 1
**Canadian sites are systematically under-cited.** Across 24 verticals and 8,420 queries, Canadian-domain sites captured the cited-source slot in just 1 in 14 AI answers — even when the query explicitly named a Canadian city or used Canadian English spelling. The dominant pattern is US incumbents being cited on Canadian-intent queries because their content depth and citation graph outweighs anything Canadian competitors have shipped.
### Finding 2
**Wikipedia and Reddit are the universal anchors.** Reference and community sources captured 23.1% of all citation slots, the second-largest bucket after US commercial domains. The implication for Canadian operators: optimizing for Wikipedia mentions, earning Reddit thread presence on Canadian-relevance threads, and seeding authoritative third-party references is now a primary citation lever — not a side activity.
### Finding 3
**Perplexity is the friendliest engine for Canadian publishers.** Perplexity's source-diversity bias surfaces Canadian sources at 2.4× the rate Google AI Overviews does. For Canadian SMBs, optimizing first for Perplexity citation patterns produces the fastest citation-share lift, then transfers (with effort) to Google AI Overviews.
### Finding 4
**The vertical gap is real.** Legal services — where directory aggregators (CanLII, Lawyer-Lookup, Avvo Canada) carry citation weight — produces the highest Canadian-domain capture rate at 14.2%. SaaS and B2B technology produce the lowest at 1.9%, where US-incumbent content depth is decisive.
### Finding 5
**Citation share predicts brand-search lift.** Sites that earned ≥3 AI-Overview citations across the sample saw a median 23% lift in branded organic search volume month-over-month. Sites with zero citations saw a median 1.4% lift over the same window.
Sample collection between January 6 and March 28, 2026. Query set was assembled from three sources: (a) Semrush Canadian commercial-intent keywords ranking on top-10 SERP positions (n=4,200), (b) Google Trends rising queries scoped to Canada (n=2,100), and (c) People-Also-Ask harvested questions from Canadian-targeted SERPs (n=2,120). Each query was issued to all four target engines (Google AI Overviews, ChatGPT Search, Perplexity Pro, Gemini 2 Pro) and the cited sources extracted. Domain classification used WHOIS and Schema.org Organization markup with manual review for ambiguous cases. Citation share is computed at the answer-level (a domain cited once per answer counts once, regardless of how many times it appears in that answer). Confidence interval at 95% for headline figures: ±0.9 percentage points.
This report adheres to the research-publication standards we apply to all Ottawa SEO Inc. original research: a clearly defined sample, a documented data-collection window, transparent statistical controls, and an explicit confidence interval where applicable.
If you would like to replicate, extend, or audit any of the findings in this report, please contact research@ottawaseo.com — we make our raw data available to qualified researchers and journalists under a non-disclosure framework that protects participant privacy.
**For Canadian operators, AI citation work is no longer optional.** The compounding cost of every quarter of inaction is a US incumbent locking down the citation slots in your category. The realistic path to citation share for a Canadian operator in 2026: (1) build the schema-markup foundation that AI engines parse for source attribution, (2) earn Wikipedia and Reddit presence on category-relevant threads, (3) ship original-research content that AI engines preferentially cite as primary sources, and (4) optimize first for Perplexity (the friendliest engine for new Canadian publishers) before pursuing Google AI Overviews citation share.
This report was authored by **Martin Vassilev**, founder of Ottawa SEO Inc. (est. 2014). Martin has personally led SEO engagements across more than 500 Canadian businesses over twelve years, with operating depth in local SEO, technical SEO, and the emerging discipline of Generative Engine Optimization (GEO).
Ottawa SEO Inc. publishes original Canadian-market SEO research quarterly. Subscribe to the Ottawa SEO Inc. research newsletter to receive the next report at publication, or follow @ottawaseoinc for findings as they're released.
For media inquiries, interview requests, or commentary on this report, contact martin@ottawaseo.com directly.
8,420 generative-engine answer queries sampled across Google AI Overviews, ChatGPT Search, Perplexity, and Gemini between January and March 2026, scoped to Canadian-domain (.ca and Canadian-business) citations across 24 industry verticals.
Yes — this report is released under a free-attribution license. We ask that citing publications include a clear attribution to Ottawa SEO Inc. and a working link to the report URL (https://ottawaseo.com/reports/ai-search-citation-share-canada-2026/). For substantive use of the data set, we appreciate advance notice via research@ottawaseo.com so we can support fact-checking.
This report was published in Q1 2026 with the most recent data available at publication. We refresh the underlying data set quarterly. The next scheduled update for this report is Q3 2026 — sign up at /contact/ to receive the next edition automatically at publication.
Yes — we make raw data available to qualified researchers and journalists under a non-disclosure framework that protects participant privacy. Contact research@ottawaseo.com with a brief description of your intended analysis to request access.
This research was funded entirely by Ottawa SEO Inc. as part of our quarterly research-publication program. No external sponsors, no advertiser influence, no paid placements. The methodology and findings are independent.
This is part of an ongoing research-publication series from Ottawa SEO Inc. covering Canadian SEO, local search, AI search citation, and emerging organic-visibility disciplines. Browse our full report archive for related studies, or see our State of AI Search Citations Canada 2026 Report for the foundational citation-pattern research that informs this work.