Skip to content
ChatGPT Ads: What US Search Data and a Real Account Say About Testing This New Channel
Digital Marketing

ChatGPT Ads: What US Search Data and a Real Account Say About Testing This New Channel

November 20, 20267 min read

People in the US are typing “chatgpt ads” into Google about six times more often than they did a year ago. For a leadership team, the question is simple: should we put money into this yet?

Intelegencia Market Signal, November 2026:

  • US searches for “chatgpt ads” rose from about 4,400 a month in October 2025 to about 27,100 in October 2026.
  • Searches for “chatgpt ads manager”, the self-serve tool advertisers use to run campaigns, went from about 10 a month to about 3,600.
  • Over the last six months, searches for “google ads” were still about 26 times larger. ChatGPT Ads adds to search; it does not replace it.
  • The four everyday searches we checked as a baseline (“pizza near me”, “car insurance”, “accountant near me”, “hotels”) moved less than 10% either way over the same six months.

Data: Google Keyword Planner, US monthly search volumes. Compared by Intelegencia.

In my view, ads in ChatGPT deserve a small, controlled test if your buyers research by asking questions and you can measure the real outcome, such as approved leads. Set a budget limit first, work on the account every week, and at 90 days judge whether results are moving toward your benchmark.

Two line charts of US Google searches per month from November 2024 to October 2026. Searches for chatgpt ads peaked at 27,100 in January 2026, fell to 9,900 in April and returned to 27,100 by October 2026. Searches for chatgpt ads manager rose from about 10 a month to 3,600 in October 2026.

Source: Google Keyword Planner (US), monthly. Chart by Intelegencia, November 2026.

What ChatGPT Ads Are and Who Sees Them

Ads in ChatGPT, sometimes called OpenAI ads, are sponsored messages shown next to the answers. OpenAI began testing them in the US in early 2026, and they appear only for people on the Free and Go plans. Plus, Pro, Business, Enterprise and Edu users see no ads (OpenAI Help Center).

OpenAI says ads are labeled, kept separate from answers, and do not change what ChatGPT says. It also says advertisers see only totals such as views and clicks, never anyone’s chats.

The biggest difference from search advertising is targeting, because there is no keyword field. Instead, you write short “context hints” that describe the situations your product suits, and OpenAI treats them as a signal of relevance, not a strict rule (OpenAI Help Center).

Our setup guide covers the account steps.

Why US Interest in ChatGPT Ads Is Rising, and Why to Keep It in Proportion

Searches for “chatgpt ads” climbed sharply in December 2025 and peaked in January 2026, the month OpenAI announced ads. They fell by about two-thirds by April and have risen or held every month since May.

The more telling number is “chatgpt ads manager”, which went from about 10 searches a month to about 3,600. People search for a tool when they are trying to use it, so this suggests more businesses are running campaigns, although search data cannot prove it. Longer phrases such as “ads in chatgpt” have not returned to their early-2026 peaks.

Just over eight months after launch, OpenAI said more than 100,000 advertisers had run campaigns on ChatGPT, and ads were available in 89 countries (OpenAI, late 2026).

Searches for established ad platforms moved like this:

US Google searchesSix months to October 2026Change vs the same months a year earlier
"google ads”3,423,000Up 18% (523,000 more)
Meta (“meta ads” and “facebook ads” together)596,000Up 18% (88,800 more)
“chatgpt ads”130,700New channel, launched in early 2026
“amazon ads”112,700Down 4% (4,100 fewer)
“microsoft ads”95,400Down 3% (3,300 fewer)
“linkedin ads”55,800Down 24% (17,300 fewer)
“perplexity ads”1,990Down 54% (2,360 fewer)

Search interest is not rising for every AI ad platform, as Perplexity shows.

Some buyers shop this way. A Gartner survey of 328 US consumers, published in June 2026, found 16% use AI chatbots to search for new products or services to buy (Gartner, June 2026). That is a minority, but they are the buyers an ad inside a chatbot can reach while they are still choosing.

What One Real Account Taught Us About a New Ad Channel

We run this channel for a client, a US financial advisor matching platform, and the first month looked bad. Cost per approved lead, meaning a lead the client actually accepted, came in at more than three times their paid search benchmark. The case study says a report that credits only the last ad someone clicked would have recommended shutting the test down in week three.

The case study traces the early cost mainly to two set-up problems. The account was built like a search campaign, and it measured form fills instead of approved leads. Once those were fixed and the team tested hints, copy and images every week, cost per approved lead crossed below the search benchmark in month four and settled about 35% below it.

One account proves little on its own. The ChatGPT account ran at under a third of the search budget, and cost per lead usually rises as budget grows, so the result is not proof the channel holds at larger spend. The full story is in our case study and in our post on the first six months of results.

How to Test a New AI Ad Channel Fairly: The Fair Test

For any new channel, the question is the same: did it earn its budget? I use a simple rule for this, which I call the Fair Test. It follows the same idea as the Four Proofs Test I use for choosing an AI SEO agency: proof before more money.

A fair test has four parts:

  1. Score the channel on the outcome that pays you, such as approved leads or sales, using the same definition as your other channels. In our client’s case, that meant monthly spend divided by the approved leads the CRM (the customer database) credited to each channel.
  2. Set the budget limit, the benchmark to beat, and the result that would make you stop, before you spend anything.
  3. Work on the account every week and review it at 90 days, judging the direction, not only the number. Our client’s account was still behind search at 90 days, but its cost had improved for three months, and it crossed below search in month four. An account still flat at 90 days, though, has a structural problem that more patience will not fix.
  4. Decide which conversations your brand should and should not appear in.

If your tracking cannot yet connect ad spend to real outcomes, fix that first with proper conversion tracking. If you want the test set up this way from day one, our ChatGPT Ads team builds accounts with the measurement and the 90-day review in place before the first dollar is spent.

Where I Disagree With the Usual Advice on New Ad Channels

The usual advice is either to get in early, before the channel fills up, or to wait until it matures.

I do not agree with either. Being early helps only if you can measure what you are buying, and our client’s first month showed how expensive an early start can look without the right set-up. Waiting has a cost too, because you learn nothing about how your own buyers respond.

Start when you can run a fair test, meaning you can measure approved leads or sales.

The CFO Read: How Much to Risk on a New Ad Channel

Expect the first month to look expensive, as it did for our client, and keep the test small compared with your main channel (our client’s ran at under a third of its search budget). Decide the test budget before launch, and compare cost per approved lead with your other channels using one definition. Scale in steps, because cost per lead usually rises as budget grows.

Three questions protect the budget:

  • What outcome will we judge at 90 days, and who confirms it in the CRM?
  • By day 90, is cost per approved lead moving toward the benchmark, and which channel sets that benchmark?
  • What happens to the budget if the trend is flat?

Deciding Where Your Brand Appears Is a Customer Trust Decision

With ads inside conversations, your brand can appear while someone is working through a real problem, such as a money decision. Choosing which of those moments you want to be part of is a customer trust decision as well as a media decision. Customer experience leaders should help make that call alongside marketing.

Gartner reported in November 2026 that 42% of CMOs are monitoring brand misinformation in AI-generated content (Gartner, November 2026). Set your rules before you scale, and use a short brand audit to decide which topics fit your brand.

Plan a Fair Test With Intelegencia

Intelegencia’s ChatGPT Ads team can set up the measurement, write the first context hints, and agree on the 90-day review with you, so the verdict rests on real outcomes. Talk to our team.

FAQs

Frequently Asked Questions

Advertising on ChatGPT can pay off if your buyers research by asking questions, and you can measure the real results. Start with a small, capped test, work on it weekly, and at 90 days judge whether it is moving toward the results of your existing channels. Results from one business may not carry over to yours.