Alberta's Online Casino Market Opens in Three Days. Its Regulator Is Already Losing the AI Answer Layer.

Alberta's regulated online casino market launches 13 July 2026. On the AI-answer questions people will actually ask about it — is this legal, who's licensed, what are the rules — the province's own regulator is not the most-cited source. A US fantasy-sports and betting-content site is.

On the seven Alberta-specific questions all three citing engines actually answered, rotowire.com was cited 53 times — more than alberta.ca (36), and nearly three times the Alberta Gaming, Liquor & Cannabis commission, the actual regulator, at 19.

Disclosure, up front: Lead Media operates verdikt.bet, a Canadian crypto-casino comparison site that is itself in this market. We are an interested party. The measurement engine behind this post, GetCited, is our product. Nothing here should be read as neutral third-party research — read the data and the method, and judge for yourself.

What we did

We ran a 20-prompt battery about Canadian and Alberta online gambling through four AI answer surfaces — ChatGPT, Claude, Perplexity, and Google AI Overviews — five repetitions each, on 08/07/2026. That produced 310 recorded runs, zero errors, and 1,584 individual citations. Eight of the 20 prompts were Alberta-specific; the rest covered the broader Canadian market.

310 runs is not a clean 20 × 5 × 4 (which would be 400). Nothing failed — there were zero recorded errors — but each engine only ever ran 15 or 16 of the 20 prompts: Perplexity and Claude each ran 16, ChatGPT and Google AI Overviews each ran 15, and the specific prompts skipped differ by engine. We don't know why some prompts were never run on some engines, and we're not going to speculate. The consequence is that raw per-engine totals aren't strictly like-for-like, so every headline table below is computed on the subset of prompts all three citing engines actually answered — not the full 20.

Google AI Overviews returned zero captured citation URLs across all 75 of its runs, with zero recorded errors. We don't know whether that means AIO cited nothing for these queries or our capture missed the citations — more on that below. Everything that follows is drawn from the three engines that did return citations.

Runs and citations, by engine:

| Engine | Runs | Citations | Avg per answer | Distinct domains | |---|---|---|---|---| | Perplexity | 80 | 695 | 8.7 | 107 | | Claude | 80 | 625 | 7.8 | 115 | | ChatGPT | 75 | 264 | 3.5 | 75 | | Google AI Overviews | 75 | 0 | 0.0 | 0 (excluded) |

This is one snapshot, three days before launch, one locale. We're publishing it now because the finding is time-sensitive, not because the sample is large.

The Alberta table

Eight of the 20 prompts were Alberta-specific (positions 1, 2, 3, 4, 5, 10, 12, 15). Prompt 15 was never run on ChatGPT, so it's excluded from this table to keep the comparison like-for-like — the table below covers the seven Alberta prompts all three citing engines answered:

| Domain | Citations | Engines | |---|---|---| | rotowire.com | 53 | 3 | | cbc.ca | 38 | 2 | | alberta.ca | 36 | 3 | | linkedin.com | 30 | 1 | | igamingbusiness.com | 28 | 2 | | casino.org | 19 | 3 | | aglc.ca | 19 | 2 | | squawka.com | 15 | 3 | | sccgmanagement.com | 14 | 1 | | sbcevents.com | 14 | 2 |

AGLC — aglc.ca, the Alberta Gaming, Liquor & Cannabis commission, the body actually licensing this market — is tied for sixth. It is out-cited by RotoWire, the CBC, LinkedIn, and an industry trade site.

Why RotoWire wins

The RotoWire citations aren't a brand-strength artefact or a fluke of one popular page. They trace to a small set of purpose-built Alberta regulatory pages:

| URL | Times cited | Prompts | Engines | |---|---|---|---| | rotowire.com/ca/casinos/alberta/laws-and-regulations | 33 | 5 | Claude, Perplexity | | rotowire.com/ca/casinos/alberta | 12 | 3 | Claude, Perplexity | | rotowire.com/news/alberta-igaming-advertising-rules-what-bettors-should-know-before-july-13-120880 | 5 | 1 | Claude | | rotowire.com/article/28-online-casino-operators-are-now-registered-to-launch-in-alberta-113721 | 3 | 1 | ChatGPT |

One page — /ca/casinos/alberta/laws-and-regulations — accounts for 33 of RotoWire's 53 Alberta citations on its own, cited by both Claude and Perplexity across five of the eight Alberta prompts. RotoWire built a dedicated laws-and-regulations page, a market-overview page, and timely news pieces keyed to the 13 July launch date. That's the whole mechanism. It's a repeatable structure, not a curiosity: purpose-built regulatory content, kept current to the launch date, gets cited on regulatory questions — regardless of who publishes it.

For contrast, the regulator's own most-cited pages:

| URL | Times cited | |---|---| | aglc.ca/albertagaming | 7 | | aglc.ca/playalberta | 6 | | aglc.ca/news/psa-fraudulent-casino-apps-and-ads-alberta | 4 |

AGLC's best single page (7 citations) is cited less than a quarter as often as RotoWire's best single page (33 citations).

LinkedIn out-cites the regulator too — but only one engine is doing it

LinkedIn's 30 Alberta citations aren't a company page or an AGLC LinkedIn presence — they're individual posts by named industry people. That matters, but so does this in the same breath: only ONE of the three citing engines cited LinkedIn at all on these prompts. This is a single-engine finding, not a cross-engine one.

| URL | Times cited | |---|---| | linkedin.com/pulse/albertas-igaming-legislation-navigating-opportunities-george-brtce | 8 | | linkedin.com/pulse/canada-alberta-fixes-13-july-igaming-market-launch-sigma-world-fsnlf | 5 | | linkedin.com/posts/tom-nightingale-531498a9_alberta-publishes-awaited-igaming-standards-activity-7417541523342409728-gXv4 | 4 |

Three individual LinkedIn posts here account for 17 of the 30 citations. One is titled around the 13 July launch date directly — the same news hook this post is riding.

How much the three citing engines agree

Restricting to the 14 prompts all three citing engines answered (the strict, like-for-like basis), ChatGPT, Claude and Perplexity cited 211 distinct domains between them. Agreement is thin:

• Cited by all 3 engines: 8 domains (3.8%)
• Cited by exactly 2 engines: 41 domains
• Cited by exactly 1 engine: 162 domains (76.8%)

Robustness check: on the looser basis of all 20 prompts (not correcting for uneven coverage), the same shape holds — 237 distinct domains, 186 cited by only one engine (78.5%), 9 cited by all three (3.8%). The finding survives the stricter test; we're quoting the strict, like-for-like numbers above as the headline figures.

The eight-domain consensus set — cited by ChatGPT, Claude, and Perplexity alike, on the strict basis:

| Domain | Citations | Prompts covered | |---|---|---| | rotowire.com | 56 | 8 | | casino.org | 37 | 7 | | alberta.ca | 36 | 7 | | igamingontario.ca | 34 | 4 | | agco.ca | 27 | 5 | | squawka.com | 25 | 7 | | sigma.world | 13 | 4 | | yogonet.com | 9 | 1 |

Roughly three in four domains cited in this run were cited by only one engine. If you're optimising for "AI citations" as a single target, you're not — you're optimising for three separate answer engines that agree on almost nothing outside this eight-domain core.

Per-engine citation density

Perplexity and Claude cite far more sources per answer than ChatGPT does:

• Perplexity — 8.7 citations per answer
• Claude — 7.8 citations per answer
• ChatGPT — 3.5 citations per answer

ChatGPT is roughly 2.5x more selective than Perplexity. A source that makes ChatGPT's shortlist is competing against a much shorter list.

What we can't tell you

• Prompt coverage was uneven, and we don't know why. 310 runs were recorded with zero errors — nothing failed. But each engine covered only 15-16 of the 20 prompts, and which prompts got skipped differs by engine. That means raw per-engine totals aren't strictly comparable, which is why every table in this piece is computed on the subset of prompts all three citing engines actually answered, not the full battery. We don't know why some prompts were never run on some engines, and we're not going to guess.
• Google AI Overviews is excluded, and we're not sure why it's empty. Zero citation URLs were captured across all 75 AIO runs, with zero recorded errors. We have not established whether that means AIO showed no citations for these queries, or our capture missed them. We're not willing to publish "Google cites nobody on Alberta iGaming" on an unresolved capture question, so we're stating the gap instead: AIO is absent from every number above, not disproven.
• This is one snapshot. 08/07/2026, one locale, 20 prompts, 5 reps. Answer engines change week to week, and this scan sits three days before the market it's describing even opens. Treat the contrasts as the signal, not a permanent scoreboard.
• A citation is not an endorsement, and citation counts are not traffic. Being cited 53 times tells you an engine pulled that URL into an answer 53 times across our runs — it says nothing about how many people read it, clicked it, or acted on it.
• We're an interested party. Repeating the disclosure from the top: Lead Media operates verdikt.bet, a site in this market, and GetCited — the tool that produced every number above — is our product.

What to take from this before 13 July

If you're an operator or affiliate in this market, the actionable finding isn't "RotoWire is winning" — it's that the mechanism RotoWire used is available to anyone: a dedicated, current, regulator-referencing page for the specific jurisdiction and the specific launch date, not a generic Canada page. The regulator has the authority; it doesn't yet have the purpose-built content. On this measurement, that gap is what the AI answer layer is currently rewarding.