Tell us about your project
A six-month-old ad network.
A scoped test, judged on your own figures.
AIxH, a digital advertising agency in Luxembourg, is testing advertising in ChatGPT within its Digital advertising service, with a capped budget and rigorous measurement. OpenAI's ad network has been open in Luxembourg since 24 August 2026 and in self-service for eligible European advertisers since 31 August. We treat it like the established networks: an eligibility question about your audience before any budget, chosen commercial queries, reviewed ads, and an arbitration against Google Ads and social campaigns on absolute figures.
After a few ad network launches observed over 27 years of digital acquisition, OpenAI's deserves a test, with expectations set on visibility, before any immediate conversion. ChatGPT claims around one billion weekly active users, and its ad network passed one billion dollars in annualised revenue less than two hundred days after its US launch. Inventory on that scale changes how an acquisition budget is allocated, in Luxembourg as internationally.
The management tools remain basic: no Quality Score, no search terms report, no visibility on the conversations where the ad appears. A network that chooses and publishes its own indicators, such as the 50% drop in the dismissal rate announced without the starting rate, is judged on the figures it does not disclose. We measure them on our side.
ChatGPT Ads is one of the levers of our Digital advertising service, alongside Google Ads & SEA and Social Ads. It also complements the organic citation work of our GEO & AI visibility service: appearing in ChatGPT's answers involves two separate jobs, paid presence and citation among the sources.
Eligibility
Ads are shown only to users on the Free and Go plans. Plus, Pro and Enterprise subscribers will never see your advertising. That is the first question, before the budget.
A large share of the market uses ChatGPT for free. According to Similarweb, around 26% of answers served to Free and Go users contain an ad, a third-party estimate not confirmed by OpenAI.
The test makes sense on commercial queries where your offer is compared: insurance, energy, home, tourism, consumer services.
Decision-makers often pay for their subscription, so fall outside the advertising perimeter. SE Ranking checked it on its own budget: around 97,000 impressions and 1,263 clicks in two weeks, for very few sign-ups.
We first check your audience's actual usage before committing a single euro, and we say so when the channel is not suited to it.
Scope
Your audience's usage of the free versions, categories approved by OpenAI in Europe, commercial queries on which your brand should appear. We also record what ChatGPT's answers already say about your offer and your competitors.
Writing and review of the ads, consistency with your prices and mandatory notices, dedicated landing pages with tracked URLs. No ad goes live without human approval, as on all our networks.
Short duration, capped budget, cost-per-click bids followed day by day. The test is run in OpenAI's Ads Manager or with its Ads Solutions team depending on your category.
Visits, contacts and cost per contact in GA4, compared with your Google Ads and social campaigns over the same period. At the end of the test, a decision: stop, extend or widen, with the figures that justify it.
A test promises no result. It produces figures that belong to you, on your market, and a decision based on something other than the indicators chosen by the network.
Read: ChatGPT Ads passes $1 billion, and OpenAI targets $100 billion by 2030 →
What is known
Coverage
SE Ranking analysed 50,006 commercial queries across 20 sectors: ChatGPT shows an ad on 25.94% of them, a rate close to the 29.45% measured on Google's AI Mode.
Targeting
14.35% of the ads shown have no thematic link with the query. In 96.37% of placements, the advertiser appears nowhere in the sources cited by the answer above its ad.
Cost
Range reported by the first US advertisers. Third-party measurement is taking shape: OpenAI has partnered with AppsFlyer, Adobe, Criteo and LiveRamp for attribution.
Read: ChatGPT ads arrive in Luxembourg with one ad in seven off topic →
Scoping
A first conversation is enough to decide whether the channel deserves a test. It goes faster if you come with:
Together we set the duration, the budget cap and the indicator that will decide. Fees are quoted after the scoping.
FAQ
Yes. OpenAI opened ChatGPT Ads in 31 European countries, Luxembourg included, on 24 August 2026, first through its Ads Solutions team and partner agencies, then in self-service for eligible European advertisers since 31 August 2026, in approved categories. Ads are shown only to users on the Free and Go plans; Plus, Pro and Enterprise subscribers see no advertising.
It depends on your audience. Ads are reserved for the free versions of ChatGPT: a consumer brand reaches a large share of its market, a B2B advertiser whose decision-makers pay for their subscription may reach almost nobody. SE Ranking measured it on its own budget: around 97,000 impressions and 1,263 clicks in two weeks for very few sign-ups, on an audience of subscribed professionals. So we start with that question before any budget.
Little so far, and mostly figures from the network itself or from third-party studies. SE Ranking observed an ad on 25.94% of 50,006 commercial queries, 14.35% of ads unrelated to the query, and the advertiser absent from the sources cited by the answer in 96.37% of placements. CPCs reported by early advertisers range from 1.50 to 4 dollars. OpenAI reports a 50% drop in the dismissal rate, without the absolute rate. No client result can therefore be promised: a test is judged on its own figures.
The network provides no Quality Score, no search terms report and no visibility on the conversations where the ad appears. So we measure on our side: dedicated landing URLs, tracking of visits and contacts in GA4, comparison of the cost per contact with your Google Ads and social campaigns over the same period, and a reading of the commercial queries on which your brand is cited or absent in the answers. The test is stopped, extended or widened on those figures.
AIxH does not publish a minimum budget and the network imposes no public threshold in Europe to date. A test is scoped over a short period, with a capped budget and a presence objective on precise commercial queries, without expecting immediate conversions. Fees are quoted after the scoping. Write to hello@aixh.com to talk about it.
Client reviews
★★★★★ 5.0 / 5 on Google (45 reviews) · See our Google profile →
“Being visible on Google is good. Being visible everywhere clients search is better. We called on AIxH to work on the overall visibility of one of our clients' websites, with a Global Search strategy. The idea is simple: stop limiting organic search to Google. You also have to work on your presence on other search engines, social platforms and, now, AI assistants such as ChatGPT, Claude or Gemini.On this subject, the AIxH team has a great deal of experience, real expertise and, above all, a results-driven approach.”
“Little by little, we widened the collaboration: SEA campaigns, community management, AI training... and at every step I found the same approach: clear advice, concrete recommendations and a genuine focus on results. The team is very available, responsive and very easy to deal with. I know I can count on them whenever I have a question or a need, and that makes a real difference day to day.”
Contact
A first conversation is enough to check it and scope a test, duration, cap and indicator included.
Scope a ChatGPT Ads test →