Google AI Mode Ads Are Here: Should Small Sellers Test New Channels?

Should small sellers try new channels like Google AI Mode ads and ChatGPT ads? Our advice: yes, but only with money you can afford to lose. First get your conversion tracking, product data and landing pages in order, then carve a small slice of budget out of your existing Google account and run an experiment around one clear business question. Most advertisers can't use ChatGPT's new visual ads yet, so for now, watch and get your creative ready.

Both channels moved this month. In an October 9 explainer, Search Engine Land pulled together the key numbers: Google said in May that AI Mode, its conversational search experience, had passed 1 billion monthly users, and a July SE Ranking study of 50,032 US commercial keywords found text ads in 29.45% of AI Mode responses. Meanwhile, Search Engine Journal reported that starting in late October, OpenAI will test visual ads that appear while ChatGPT generates images, with a small group of US advertisers, and that "most advertisers won't have access to visual ads yet."

How do ads show up in AI Mode, and do you need a separate campaign?

According to Search Engine Land's roundup, AI Mode currently has four ad formats, most of them still in testing:

• Existing text and Shopping ads: served from eligible campaigns and shown below the AI answer or blended into it (testing).
• Conversational Discovery ads: Gemini tailors the ad creative to the user's question (testing).
• Highlighted Answers: sponsored merchants appear directly in the list of recommendations (testing).
• Direct Offers: promotions shown during shopping conversations; retailers supply the offer in their campaign settings, and Google's AI decides when to show it (pilot).

The key point: Google Ads has no separate "AI Mode campaign," and you can't pick AI Mode as a placement on its own. What gets into AI Mode is Performance Max, Shopping ads and Search ads using broad match, including AI Max for Search. In September, a small experiment also let exact- and phrase-match Search ads enter AI Mode when user intent was clear.

The article also warns that AI Mode launching in a market doesn't mean its ad formats are available there. If you sell into the US, step one is confirming whether your account can take part.

Can you run ChatGPT ads right now?

According to Search Engine Journal, ChatGPT's existing ads appear below answers, and OpenAI says ChatGPT reaches 1.2 billion people a week. The new visual ads will appear while users generate images, using pictures to show product inspiration, how a product is used or what it feels like to use it. They will be clearly labeled and kept separate from the image the user generated.

But OpenAI hasn't released targeting, pricing, creative requirements or reporting details for the new format. On measurement, the report says Hightouch, Tealium and LiveRamp can send conversion data back to ChatGPT Ads, while AppsFlyer, Triple Whale and others are attribution partners. The report also stresses that the early results several partners disclosed measured different metrics, so they can't be treated as general benchmarks.

For small sellers, the practical move right now is to prepare high-quality images of your products in use and follow the rollout, rather than pulling existing budget to chase a placement you can't get yet.

Why do new channels feel exciting and scary at the same time?

Behavioral economics has a name for this tension: loss aversion. Losing a given amount hurts noticeably more than gaining the same amount feels good.

It pushes sellers to two extremes. Some are afraid of missing the window ("everyone else is running these, so I'm losing out by sitting still") and rush money in. Others are afraid of wasting it, especially after being burned by an agency that "guaranteed results," and touch nothing at all. Both reactions come from the same psychology, just pointed in opposite directions.

It's also why many small sellers are wary of the traditional agency model: the monthly fee always gets paid, but nobody can explain the results. Skepticism about new channels is healthy. The key is to write it down as questions you can test: how much to spend, which metric to watch and how long before you decide.

So turn doubt into method. Set a test budget you won't mind losing and a stop-loss line, then let the data talk. Doubt isn't the problem; never testing it is.

Is a new channel worth testing? Start with this table

How do you test AI Mode ads with money you can afford to lose?

1. Confirm eligibility. Check whether your target market and your account can take part in the relevant ad formats.
2. Lay the foundation. As Search Engine Land recommends, fix conversion tracking, product feed errors and landing pages first, so your pages clearly say what you sell and who it's for.
3. Write down one business question. For example: "With AI Max turned on, can we win orders from new search terms without raising our cost per acquisition (CPA)?"
4. Set a stop-loss budget. Take a small share of your existing Google budget, fix a test period and a maximum loss, and stop when you hit that line.
5. Review AI Max settings one by one. Focus on brand exclusions, URL controls and text customization. The article warns that under certain asset optimization and URL settings, fixed ad copy won't be honored.
6. Judge results on profit. AI Max reports don't break out AI Mode performance separately, so don't just watch conversion counts. Look at order quality, gross margin and the share of new customers, and review search terms, landing pages and assets.
7. Decide on schedule. If the test hits its target, scale up gradually; if it doesn't, stop. Write down what you learned as the starting point for the next round.

On a tight budget, how do you rank old and new channels?

The order is simple: get your existing channels healthy first, then test new ones with what's left over. Your test budget should come from money saved by optimizing current channels, not from cutting ads that are already making money.

ToucHigh's Ads feature automatically adjusts budgets toward ROAS (return on ad spend) across Amazon, Meta and Google and makes creative-fatigue and budget-allocation calls for you, freeing up time and budget for experiments. On the other side, visibility in AI answers doesn't have to be all paid: GEO (generative engine optimization) is about getting your store cited naturally in answers from ChatGPT, Perplexity and Google AI. To see how the paid and organic tracks fit together, explore ToucHigh's products.