GEA: advertising inside AI answers
People no longer search for terms alone. They hold multi-step conversations with AI systems such as ChatGPT or Google, describe complex problems and have answers generated for them. For advertisers, the line between SEO and SEA is dissolving.
GEA, generative engine advertisement, describes ads that no longer sit in fixed blocks but appear dynamically inside the running conversation, as a solution to the problem at hand. These new touchpoints let brands be present at the moment a buying decision is made.
ChatGPT Ads: the way into advertising on AI surfaces
On 24 August 2026 OpenAI launched ChatGPT Ads in Germany, Austria, Switzerland and 31 European countries in total. What began as a US pilot in February 2026 has become, within months, an advertising platform with access to more than 900 million weekly active users.
A context with clear intent
By OpenAI's data, around 20 percent of all conversations carry direct commercial intent: researching products, comparing alternatives, making decisions. The intent differs markedly from classic search: an average ChatGPT prompt is 17 times longer than a Google query.
Reach and audience
Ads run only in the free tiers; paying subscribers see no advertising. For B2C and e-commerce that means a broad, highly engaged audience. B2B companies should test carefully: many professional heavy users work in the paid version and are harder to reach through this channel.
Formats and campaign setup
Since 31 August 2026 the self-serve ads manager has been open in beta to eligible advertisers in Germany, Austria, Switzerland and 28 further European markets. Access currently runs through the OpenAI Ads Solutions team and selected agency and technology partners. Two formats are available, both clearly separated from the AI answer and labelled "Sponsored":
Sponsored answers
Text ads below the AI answer: company name, favicon, a very short headline (16 to 24 characters recommended, 50 maximum), a description (32 to 48 recommended, 100 maximum) and a link to the landing page. That brevity forces precise, benefit-led language.
Shopping carousel
A scrollable row of sponsored product cards at the foot of the chat. They draw directly on the product feed, Shopify or Etsy for instance, and show brand, product name, price, availability and delivery time live.
Billing and bidding
Billing runs on CPM or on a CPC / oCPC model. For CPC campaigns OpenAI recommends starting with a maximum bid of 3 to 5 US dollars per click. Early reports show actual click prices often well below that, with impression costs competitive against more mature platforms. By 2030 OpenAI aims to reach 100 billion US dollars in advertising revenue with this system.
Google Ads in the AI search features
Google is expanding its advertising options in the AI-driven surfaces too, with good news and awkward news at once: there are no separate AI placements and no dedicated AI campaigns you can switch on or off. Google uses the existing auction and ranking system.
To appear in the new surfaces at all, you need AI-driven campaign types:
- search campaigns with broad match and smart bidding
- AI Max for Search: broad match combined with keyword-less dynamic search, adapting headlines and landing page URLs in real time
- Performance Max campaigns
- classic shopping campaigns with optimised product feeds
Anyone still working exclusively with rigid exact-match setups or highly granular campaign structures is cutting themselves off from future traffic. The AI needs consolidated data and flexibility to understand the context of modern queries.
Ads around AI Overviews
AI Overviews are AI-generated summaries at the top of the results page. They have been active in Germany since spring 2025 and shape web search for around two billion people. Ads appear in three areas:
- Above the summary
- For searches with clear, high purchase intent, Google still places classic text or shopping ads at the very top.
- Below the summary
- For informational or mixed searches the AI box dominates the visible area. Ads move down and only become visible on scroll.
- Inside the AI answer
- The most prominent integration: searches such as "best CRM provider" generate an AI comparison list in which sponsored platforms appear as product cards or contextual sponsored links, as part of the generated answer. Currently rolling out market by market.
The effect on performance campaigns is real: because the summary satisfies curiosity on informational searches immediately, click volume in the research phase collapses: nobody needs to leave the results page. Stable impressions in the account can mask this while click-through rates fall across the board. Advertisers respond by shifting budget to transaction-near keywords, which drives click prices up at the bottom of the funnel in competitive sectors.
Ads in Google AI Mode
AI Mode goes a step further: a standalone chat surface built on Gemini, where a question can be refined step by step. Important for the German market: these ad formats are not yet available here. They are in early testing or run exclusively in the US. The formats shown at Google Marketing Live are future rollouts, not options for today:
- Conversational discovery ads
- Ads that adapt their creative and selling points in real time to the specific question asked.
- Highlighted answers
- If the AI recommends "the best language apps for a trip to Italy", paid and highly relevant apps can appear as a highlighted recommendation in that list.
- AI-powered shopping ads
- Gemini pulls matching products from the Merchant Center feed and automatically writes accompanying copy explaining why this particular model fits the requirement.
- Business agent for leads
- Instead of a static form, a chatbot starts inside the ad slot, answers questions from the company website and qualifies the lead.
- Direct offers and native checkout
- Dynamic discount bundles and purchase completion inside the search surface via the Universal Commerce Protocol.
The foundation stays organic
Anyone expecting to buy success in generative search with budget alone is mistaken. Google and OpenAI stress the same principle: paid campaigns and organic visibility are inseparable. It pays to optimise the content organically alongside the ads.
The black box: monitoring and reporting
The biggest hurdle at the start is how little the platforms show. If you are used to complete attribution chains and granular reports, you face a black box:
- ChatGPT
- The ads manager currently gives no insight into the chats, queries or conversation histories an ad appeared in. OpenAI withholds that data for privacy reasons. Conversion-based optimisation and third-party measurement are still being built, so in practice you work with proxy metrics such as CTR or manually placed pixel conversions.
- There is still no separate reporting for AI Overviews or AI Mode. Impressions and clicks from the AI surfaces are diluted into the standard search and shopping reports. You cannot bid on these placements specifically, cannot exclude them, and cannot evaluate their performance in isolation.
Three ways to measure anyway
Implement server-side tracking
Browser cookie tracking is going blind through ad blockers, iOS updates and cookie restrictions. Server-side tracking bypasses those blocks and, by vendor benchmarks, often delivers more than ten percent additional data straight to Google smart bidding and the OpenAI conversions API. Feeding the AI without that data merely automates the burn.
Move to a blended measurement framework
Last-click attribution fails when people research in ChatGPT for days, continue in Google AI Mode and finally convert through a classic search. Look at total SERP visibility instead: the combined effect of organic AI citations, paid GEA ads and classic impressions.
Feed CRM data back
In B2B especially, a lead is not a lead. Integrate your CRM deeply into the Google account. Signalling which leads actually turn into deals trains the algorithm to look for high-value decision-makers in AI environments.
What this means in Germany today
ChatGPT Ads can already be booked here: access runs through the OpenAI Ads Solutions team and agency and technology partners, open self-service is not yet available. Google is testing AI Mode ads in the US only. What works here today are AI Overviews and the AI features inside Google Ads. Does that mean waiting? No. It means laying the foundation while competitors still sleep.
Secure access
Register at ads.openai.com or check whether your agency already has partner access. Business verification takes time. Start now.
Set up tracking
Implement server-side tracking and plan for the OpenAI conversions API. Every month of clean conversion data is a head start latecomers cannot make up.
Make the Google account AI-ready
Open campaigns to broad match, test AI Max, activate Performance Max. Only these campaign types appear in the AI surfaces later. Switching at the EU rollout is switching too late.
Clean up product data
Check titles, images, prices and attributes in the Merchant Center and your shop feed. For ChatGPT as for Google: the feed is the new creative.
Sources
- OpenAI: ChatGPT Ads expands across Europe
- OpenAI Help Center: Launch Campaigns
- Marketing Dive: What defines ChatGPT ads
- Similarweb via OfficeChai: ChatGPT queries 17x longer than Google searches
- Digiday: OpenAI brings product carousels to ChatGPT ads
- Search Engine Land: Google Marketing Live 2026
- PPC Land: Google Marketing Live 2026
Your contact
Project Strategist
Project strategist and netzstratege of the first hour. Building online shops since 2003, he guides digital commerce, with depth in Google Ads, Google Shopping, Merchant Center, Search Console and Analytics. In Karlsruhe he leads retail clients as project manager and consultant.
Where do you stand on AI advertising?
In the GEA potential analysis we review access, tracking and campaign structure. We say where starting is worth it.