Every Major AI Search Engine Just Started Selling Ads — Here's What That Means If You Don't Have a Million-SKU Catalog

Here's a sentence that would have sounded like science fiction two years ago: every major AI assistant your customers use to find things is now, or is about to be, selling ad space inside the answer itself. Google AI Mode. ChatGPT. Perplexity. Microsoft Copilot. Amazon's Rufus. Not "exploring it." Doing it, this year, with real advertisers already live.

If you run a small business, that's not a tech headline to skim past. It's the starting gun on the same pay-to-play arms race that happened to Google Search over the last fifteen years — except this time it's happening in months, not years, and the early infrastructure is being built for advertisers with a million SKUs in their product feed, not the business with twelve. Here's exactly what changed, why it quietly favors the biggest catalogs in the room, and what still gives a smaller business a real shot at showing up anyway — while that door is still open.

What Actually Happened (And It's Not Just One Platform)

This isn't a single rumor. It's five platforms moving in the same direction inside a few months of each other:

• Google has started showing ads inside AI Mode. A Google spokesperson confirmed it's part of a "monthslong test" that's actively running right now — ads with a "sponsored" label appearing at the bottom of AI Mode results. Google is one of the first major players to monetize AI search directly, alongside Microsoft's Copilot and Amazon's Rufus.
• OpenAI just made ChatGPT shopping ads dramatically easier to run at scale. Retailers can now connect their existing product catalog — the same structured feed many already send Google — and ChatGPT auto-generates ads from it. The platform is built to handle up to a million SKUs per advertiser. New advertisers have to submit a 100-product sample before OpenAI will take their full catalog.
• Perplexity has launched its own shopping feature and is actively testing ad monetization on top of it.
• This follows the ChatGPT Ads self-serve launch we covered back in our ChatGPT ads breakdown — OpenAI dropped its $50,000 minimum spend and opened a self-serve Ads Manager to any US advertiser. That was step one. Catalog-driven ads at a million-SKU scale is step two, and it's a very different kind of advertiser OpenAI is now building for.

Put plainly: the AI answer box — the thing that replaced the blue links your customers used to click through — is turning into the same kind of auction-driven, pay-for-placement real estate that the top of a Google results page has been for over a decade. It just got there faster, and the infrastructure being built right now is sized for catalogs most small businesses don't have.

Do this now: Don't wait for "the AI ads rollout" to feel finished before you pay attention. By the time it's fully live everywhere, the early-mover pricing advantage — the cheaper clicks that always exist before a platform gets competitive — will already be gone, the same way it vanished on Google Ads a decade ago and on ChatGPT Ads within months of launch.

Why This Quietly Favors Big Catalogs, Not Small Budgets

Here's the part that doesn't get said out loud in most of the coverage of this shift: the tooling being built for AI search ads is catalog infrastructure, not small-business infrastructure.

• OpenAI's new system is explicitly built to ingest up to a million SKUs and reuses the structured product feed format big retailers already maintain for Google Shopping. If you're a local service business with no product catalog at all — a plumber, a dentist, a consultant — there's currently no equivalent ad product built for you yet inside these systems.
• Ad tech partners quietly plugging into this ecosystem — names like Criteo and StackAdapt have come up in early reporting — are built to manage large-catalog, programmatic bidding at scale. That's not hostile to small business, but it's not designed around a business with a five-product lineup and a modest monthly budget either.
• Google's own documentation so far covers AI Overviews eligibility more clearly than AI Mode — meaning even the advertiser control and reporting tools for this new inventory are still immature. Early-stage ad inventory with immature reporting tends to get bought up fastest by advertisers big enough to absorb some wasted spend while they test. That's rarely the small business with a tight monthly budget.
• Early analysts are telling advertisers point-blank: "don't panic or overspend," test at a comfortable pace, and expect early ROAS to look weak. That's sound advice — but it's advice aimed at agencies and brands who can afford to run a controlled experiment. A solo operator with a $500/month ad budget doesn't have a lot of room for "testing at a comfortable pace."

None of this means paid AI ads are a bad idea forever. It means the floor is being built right now by and for the biggest players first, the same way Google Ads auctions have steadily priced out smaller bidders over fifteen years. If your entire 2027 marketing plan depends on getting in early and cheap on AI search ads, you're competing for inventory against advertisers who can treat a six-figure test budget as a rounding error.

Do this now: If you want to test any of this paid inventory, treat it exactly like the analysts recommend to the big guys — a small, bounded test budget with a hard 30-day review, not a reallocation of your core spend. But don't treat paid AI ads as your main lever. There's a bigger one still sitting wide open.

The Opportunity This Creates: The Organic Answer Slot Just Got More Valuable

Here's the part worth sitting up for. Every time a search channel gets crowded with ads, the organic result sitting next to those ads becomes more valuable, not less — because it's the trusted-looking answer in a sea of sponsored ones, and because fewer competitors are willing to do the unglamorous content work to earn it.

That's exactly the dynamic opening up inside AI answers right now:

• Decision-grade AI visibility tracking is still shockingly rare. Industry data we flagged in our ChatGPT ads piece puts systematic AI visibility tracking at only 16% of brands — most of the market isn't even measuring whether they show up in AI answers, let alone optimizing for it.
• Getting cited organically inside an AI answer still works, and it's measurably valuable. We covered the data on this in our Google AI Mode breakdown: pages that get cited inside an AI Mode answer get 120% more clicks than uncited pages in the same results. That's organic lift, with no CPC, no auction, no million-SKU catalog required.
• AEO and GEO — the discipline of writing content that gets accurately cited inside an AI-generated answer — reward exactly the kind of direct, fact-forward writing a small business with real expertise and real client results can produce. You don't need Amazon's catalog. You need a page that answers the real question your customer is asking, in the first sentence, backed by something true and specific to your business. That's the core of a real multi-platform GEO strategy, and it's still a wide-open lane.
• As paid inventory fills the top and bottom of these AI answers, the organic citation sitting in the middle becomes the thing that reads as credible — the same psychological effect that's kept organic search results valuable even as Google stacked more ads around them for twenty years.

The businesses that treat "AI search now has ads" as a reason to panic about being priced out are missing the actual move: double down on the free lane while the paid lane is still being built for someone else's catalog size.

Do this now: Run the same audit we recommend before any paid AI test — ask ChatGPT, Gemini, Claude, and Perplexity the top five questions your buyers actually ask before they hire or buy from someone like you. Note who gets cited. If it's not you, that gap is your highest-leverage content priority this quarter, and it costs nothing but the work.

What To Actually Do This Week

• Don't let "AI ads exist now" talk you into redirecting budget away from content and AEO/GEO work. The data says the organic citation lane is getting more valuable as the paid lane fills up, not less.
• If you want to test paid AI search ads, cap it hard. A small, bounded test budget with a 30-day review — never your core spend, and never with the expectation of early-platform pricing lasting.
• Run the manual AI visibility audit this week, not "eventually." Five real buyer questions, four AI tools, one hour. You'll learn more from that hour than from any paid AI ads dashboard.
• Prioritize original data and real client outcomes over generic advice content. AI systems are built to prefer citing a primary source over the fifth article saying the same generic thing — and that's a fight a small business with real results can actually win.
• Revisit this quarterly, not annually. This landscape moved from "ChatGPT ads with a $50K minimum" to "catalog-driven ads at a million SKUs" in about five months. A strategy built on last year's rules is already out of date.

FAQ

Are there ads inside AI search results now, like ChatGPT and Google AI Mode?
Yes. Google has started testing sponsored ads inside AI Mode results, OpenAI has built catalog-driven shopping ads for ChatGPT that can handle up to a million SKUs per advertiser, and Perplexity has launched its own shopping feature with ad monetization in testing. Microsoft Copilot and Amazon's Rufus are also monetizing AI search answers.

Does this mean small businesses can't compete in AI search anymore?
No — it means the paid lane is being built first for large-catalog advertisers, but the organic citation slot inside AI answers is still free, still measurably valuable (cited pages get 120% more clicks than uncited ones in the same AI Mode results), and still winnable with the right content strategy. That's a bigger near-term opportunity for most small businesses than paid AI ads.

What's the difference between regular SEO and optimizing to get cited inside an AI answer?
Regular SEO optimizes a page to rank and hold a visitor's attention once they click. AEO (answer engine optimization) and GEO (generative engine optimization) optimize content to be accurately quoted inside an AI-generated answer — which rewards clear, question-shaped structure and original, fact-specific content over keyword density alone.

Should I spend money testing ads on ChatGPT or Google AI Mode right now?
If you want to test, cap it at a small, bounded budget with a hard 30-day review — the same advice analysts are giving large advertisers. Don't treat it as your primary growth lever yet, since the early infrastructure and reporting tools are still immature and built around catalog-scale advertisers.

The Bottom Line

The AI search industry didn't wait to figure out the "right" way to monetize AI answers before building the ad infrastructure — Google, OpenAI, Perplexity, Microsoft, and Amazon are all live or testing right now, within months of each other. That's genuinely fast, and it's genuinely built, for now, around advertisers with catalogs most small businesses don't have.

But that's exactly why the organic answer slot — the one you earn with real expertise and content built to be cited, not clicked — is worth more right now than it's been in years. Most of the market still isn't measuring whether they show up there. The businesses that fix that gap this quarter will be sitting in the answer long after the paid lane gets as crowded and expensive as Google Ads did.

Figuring out exactly where your marketing budget should move right now — content and AEO/GEO work, a bounded paid AI test, or both — is precisely the kind of AI strategy work cCoreVentures' AI Strategy & Implementation practice does for small business clients every day. Book a free AI strategy session and we'll tell you straight where your gaps are and what to fix first.