The Toronto ChatGPT Ads Index

Last updated: July 2026. We refresh this every month.

This is a running measure of where ChatGPT actually shows ads to Toronto buyers, industry by industry. Each month we run a fixed set of real Toronto buying questions through ChatGPT and record which industries draw an ad, who is advertising, how well the ad matches the question, and what the market is charging. We have not found anyone else publishing a Toronto-specific version of this, so we built it. We work in AI search and paid media every day, and the honest picture of this new channel did not exist anywhere we could point a client to.

One thing this is not: a pitch to go spend on ChatGPT ads. For most Toronto local businesses right now, this channel is early and thin, and being recommended organically matters more (we measure that separately in the Toronto AI Citation Tracker). This index exists so you can see, for your specific industry, whether ChatGPT ads are worth a look yet, or whether the smarter move is to get your organic house in order first and revisit when your category heats up.

Start here: ChatGPT ads are real, thin, and easy to miss

The thing almost everyone gets wrong is this: ChatGPT ads are real, but they are thin and volatile, and they are easy to miss. They show up for everyone, signed in or not, we have watched the same Toronto query serve an ad both ways. What decides whether you see one is not your login state, it is the question and the moment: ads cluster on commercial buying questions, they rotate from one session to the next, and they sit in a card below the answer where a quick glance skips right over them.

That volatility explains a result that threw us at first. A July spot-check of 50 hyper-local Toronto questions returned an ad on just 2 of them, while a broader June study of 90 commercial buying questions found 48% carried an ad. The gap is not about who was signed in. Local service questions simply draw far fewer advertisers than commercial ones, and any single pass misses ads that rotate in and out. So if you check your own industry once and see nothing, treat that as a floor, not proof the lane is empty.

So the map below is measured from a signed-in Free-tier session with web search on, the same setup for every question so the numbers stay comparable.

How ChatGPT ads work right now, the 2026 facts

Before the map, the ground rules, because most of this changed in the last few months:

• Ads launched inside ChatGPT in February 2026, at roughly a $60 CPM with a $200,000 minimum commitment, aimed at large brands.
• Self-serve buying opened on May 5, 2026, and the minimum spend was removed. Any approved advertiser can now start at whatever budget they choose.
• Cost-per-click bidding arrived in April 2026, with a recommended starting bid of $3 to $5 per click. It runs as a second-price auction, so you typically pay just enough to beat the next bid, not your ceiling.
• Current rates, from third-party reporting since OpenAI does not publish official numbers: roughly $3 to $5 per click or $25 to $60 per thousand impressions in general, with software and finance higher (around $8 to $18 per click) and e-commerce lower (around $2.50 to $5). CPMs have compressed quickly from that $60 launch figure toward $25 in some categories as inventory scaled.

The plain version: the door just opened to small and mid-sized advertisers, and it is cheaper to walk through now than it will be once more of them notice.

The map: 18 Toronto industries, most to least competitive

This is the running month-over-month map. June is our five-question-per-industry baseline; July is a lighter two-question spot check, both run signed in on the Free tier. Each row puts the two months side by side: the share of an industry’s buying questions that returned a labelled ad, who advertised this month, and our one-word read on what to do. A July zero means no ad appeared on those two questions, not that the category emptied out, because presence rotates week to week. Advertiser tags are tinted green when the ad was on-target and amber when it was off-target.

CPC and CPM are third-party market estimates. OpenAI does not publish official rates, so treat them as directional. Competition and relevance are from the Stratezik logged-in ChatGPT ad study, 90 buyer questions across 18 industries, June 19, 2026.

What the map actually tells you

Three patterns matter more than the tiers themselves.

A crowded-looking category can be wide open

A high ad rate is not the same as high competition. Online education and professional services both showed ads on most questions, but almost none of those advertisers are in the category. A data-pipeline tool and a design app ran against "best bootcamp" and "where to learn UX" questions. A forms app and an HR system ran against "best small business lawyer" and "accountant for a startup." The category looks busy and is wide open to anyone who actually does the work.

In local categories, your competition is not other local businesses

The ads that show up in local-service questions are mostly not local providers. Lyft alone advertised against a Botox question, an eye-exam question, and a movers question. A pet-products brand advertised against vet visits. A forms tool advertised against renovation and roofing questions. These are broad-reach advertisers buying cheap, uncontested local intent, precisely because no dentist, optometrist, or mover is bidding against them. A real local provider would win relevance on the first day.

The relevance gap is the opportunity

Across the 43 ads we logged, 51% had an advertiser whose core business matched the question, 35% were loosely related, and 14% were flat mismatches. But the quality splits sharply by category type. In commercial-intent industries (finance, SaaS, travel and the like) 59% of ads were well matched. In local-service industries, only 36% were. In other words, the more local the question, the more likely the ad you would compete against is off-target.

Who is actually buying

A few advertisers show up again and again, and they tell you who understands this channel already. Lyft ran across three unrelated local categories. A lead-generation forms tool ran across legal and home services. One insulation company owned every HVAC ad we saw. In finance it was the big banks and insurers on nearly every question, and in software it was the large horizontal platforms. The pattern is clear: sophisticated national advertisers are treating cheap, uncontested Toronto intent as inventory to scoop up, while the local specialists who would obviously win on relevance have not arrived yet.

Do you actually need ChatGPT ads? A straight answer by business type

If you are in a saturated category (B2B SaaS, finance and insurance, travel). It is already table stakes. Your relevant competitors are on nearly every buying question. If you are absent, you are the only one not in the room. This is the one group for which "wait and see" is the wrong answer.

If you are in a white-space category (fitness and gyms, pest control, electrical, and the near-empty dental, optometry, auto repair, moving). Nobody relevant is bidding. You could own the lane for the price of showing up, and the ad you would outrank is often a rideshare company or a pet-food brand, not a competitor. This is the cheapest, least contested entry point on the whole map.

If you are in a crowded-but-non-endemic category (online education, professional and legal services). The category looks competitive, but the advertisers are not in your business. The relevance gap is your opening: a real school or a real law firm bidding here starts with an advantage the current advertisers cannot match.

For almost everyone, one caveat. Getting recommended organically still matters more than ads for most local searches right now, and it is free. Before you buy placement, make sure AI systems can find, read, and cite you at all. We track that side in the Toronto AI Citation Tracker, and for most local businesses it is the higher-return move this year.

When is the right time

Now is the cheapest this will be. Self-serve opened in May, the minimum spend is gone, and entry CPCs sit around $3 to $5. But two honest cautions. There is no proven return yet, no measured lift the way there is on Google or Meta. And the inventory is volatile: an ad we saw on a sedation-dentistry question on July 3 was gone from the identical question two days later. Treat this as a calculated early bet with a small budget, not a channel you shift real spend into yet.

What to actually do

Three tactics travel well across industries. Target the questions where ChatGPT names nobody and gives generic advice, because there your ad becomes the recommendation instead of fighting a named list. Make your geography unmistakable, since the current ads waste money crossing city and category lines and you can beat them simply by being precisely relevant. And start small, measure clicks against real bookings or calls, and scale only what pays back, because nobody yet knows the true return and your own numbers will be the first honest read you get.

What this index cannot tell you yet

We would rather state the limits than dress them up. This edition adds July beside the June baseline, our first real month-over-month read, and ad presence is volatile, we watched a sedation-dentistry ad appear one day and vanish two days later, and the same question serve different advertisers from one session to the next. So any single reading is a snapshot, and a single "no ad" is a floor, not proof. That is exactly why this is built as a running index: from the next edition on we sample each question several times to average out that churn, and carry a comparison to the month before. Relevance is our editorial call, advertiser core business against the question, not a number handed to us. The CPC and CPM figures are third-party market estimates, since OpenAI publishes no official rates.

Methodology

We ran 90 real Toronto buyer questions across 18 industries, signed in to ChatGPT on the Free tier with web search on, in fresh chats so context did not leak between questions, on June 19, 2026. For each we recorded whether a labelled ad appeared, the advertiser, and whether the advertiser matched the question. We used the actual "Ad" label ChatGPT shows, not the organic map or citation cards that can look similar at a glance. Separately, we spot-checked the 50 hyper-local questions from our monthly citation tracker on July 3 and July 5. Those returned very few ads, which is what local service questions do, not evidence that ads are rare everywhere. The datasets behind every number here are available on request.

How this index updates

We refresh this page every month, on the same schedule as the Toronto AI Citation Tracker. Each update re-runs the logged-in ad check, updates the map, and, from the next edition on, shows what moved. A short dated changelog at the bottom records every revision, so you can always see when a number last changed.

Changelog:

• July 2026: inaugural index. Baseline map from the June 19 logged-in study (90 questions, 18 industries, 48% ad rate) plus the July logged-out cross-engine scan.

About Stratezik

Stratezik is a Toronto marketing agency that runs on its own AI agent system. We work across AI search visibility and paid media, and we publish what we learn so Toronto operators can make this call with real numbers instead of hype. If you want your industry tracked in a future edition, or you want to talk through what this means for your business, email dave@stratezik.com.

Sources

1. Stratezik ChatGPT Ad Competitiveness Study, June 2026: 90 buyer questions across 18 industries, run in logged-in ChatGPT (Free tier, web search on) on June 19, 2026, scored for ad presence, advertiser, and relevance. Dataset available on request.
2. Stratezik ChatGPT Ad-Presence Dataset, July 2026: 50 frozen Toronto buyer questions checked logged out for a labelled ChatGPT ad on July 3 and July 5, 2026. Dataset available on request.
3. Stratezik Toronto AI Citation Tracker: the companion monthly study of whether AI assistants name Toronto businesses organically. stratezik.com/blog/toronto-ai-citation-tracker-july-2026.
4. ChatGPT ad pricing, targeting, and self-serve access, third-party market reporting, 2026: OpenAI Help Center (help.openai.com/en/articles/20001207), WebFX (webfx.com/blog/ai/chatgpt-ads-cost), eSEOspace (eseospace.com/blog/chatgpt-ads-pricing), The Next Web (thenextweb.com/news/openai-chatgpt-cpc-ads-launch). OpenAI does not publish official ad rates; figures are directional.