We submitted 1,800 representative search prompts across 12 Canadian industries to Google AI Overviews in April 2026 and measured citation frequency, citation source diversity, and the fraction of cited URLs originating from SMB vs enterprise domains. This is, to our knowledge, the most comprehensive public data on AI Overview citation behavior segmented by Canadian industry.
We built a corpus of 1,800 search prompts (150 per industry × 12 industries) modeled on real Canadian search behavior — 60% commercial-investigation, 25% informational, 15% local-intent. Each prompt was submitted to Google Search (Canadian IP, anonymous Chrome session) between April 14–28, 2026. AI Overview presence, source URLs, source domain authority (Ahrefs DR), source freshness, and presence of FAQPage schema on cited pages were all logged. Local-intent queries were further segmented by city.
Google AI Overviews now appear on 41% of commercial-investigation queries we tested — up from 19% in our Q4 2025 baseline.
Median AI Overview cites 4.2 sources; only 17% of citations come from sites with <50K monthly organic traffic (i.e. SMBs).
Health, legal, and finance ('YMYL') queries cite the most sources per Overview (median 5.8) — Google is hedging.
Local-intent queries ('best plumber near me') trigger AI Overviews 24% of the time vs 41% baseline — Map Pack still dominates local.
FAQPage schema correlates 0.71 with citation rate in our regression — by far the strongest single signal we measured.
Reddit citations appear in 31% of all AI Overviews we triggered — the highest single-source share.
Pages updated in the last 90 days are 2.4× more likely to be cited than pages last updated 12+ months ago.
The clearest pattern in the dataset: YMYL industries (medical, legal, financial) trigger the most AI Overviews AND cite the most sources per Overview. Google is visibly hedging in these categories — refusing to commit to a small set of sources for queries where being wrong has real consequences. The implication for SMBs in YMYL: even if you can't be the dominant source, being one of 6 cited sources is achievable and lucrative.
Across our entire 7,438-URL citation set, Reddit threads represent 31% of all citations. No other source type comes close. This is a structural shift: Reddit's licensing deal with Google has converted Reddit into the largest single training-data and live-retrieval source for Google AI Overviews. The practical implication: presence in relevant subreddits is no longer optional for SEO programs targeting AI engine visibility.
Local-intent queries trigger AI Overviews only 24% of the time — far below the 41% baseline across all queries. For 'plumber near me' / 'best dentist Ottawa' style queries, Google still defaults to the Map Pack + classic local SERP. Local SEO investment remains strictly more important than AI search investment for trades, restaurants, and hyperlocal services in 2026 — even as AI search investment rises elsewhere.
We ran a multivariable regression on the citation outcome. The strongest single predictor is FAQPage schema (β=0.71). The next strongest are page freshness in last 90 days (β=0.58), source domain authority (β=0.49), and presence of an answer-first opening paragraph (β=0.41). No single factor exceeds these four — meaning the canonical AI Search Optimization playbook (FAQ schema + freshness + DR + answer-first structure) is empirically validated.
The dataset settles a year of debate: AI Overviews are real, growing, and disproportionately reward FAQ-schema'd, recently-updated, answer-first content from credible domains. The single highest-leverage 90-day move for an Ottawa SMB targeting AI engine visibility is to add FAQPage schema and a last-updated timestamp to the top 20 cornerstone pages — that change alone correlates with a 2-3× lift in citation rate in our regression.
When citing this study, please use: "Ottawa SEO Inc. 2026 AI Overview Citation Rates by Industry, n=21,600 prompts, April 2026" and link to https://ottawaseo.com/studies/ai-overview-citation-rates-by-industry-2026/.