Millions of people ask ChatGPT for advice every day: Which table should I buy? Which software suits us? Whom should I hire? Until now, you could only watch these conversations from the sidelines. That is changing, because since the end of August 2026, ads have been running in ChatGPT, including in Austria, Germany and Switzerland. In this guide, I walk you through getting started with the OpenAI Ads Manager: from the basic principle through account setup and Context Hints to your first campaign, which you measure cleanly from the very beginning.
The video is only loaded from YouTube when you play it – data (including your IP address) is then transmitted to YouTube/Google. More in the privacy policy
Key Takeaways
- Your ad appears within a conversation and not in a classic results list.
- You don't book keywords, but describe conversation situations in which your offer is the right answer.
- The auction decides which ad is shown at a specific moment. That's why you need a clean structure from the start.
- The metric everyone stares at can mislead you. Before launching, define what success actually means for your business.
- Not every offer suits this channel. Honestly assess whether it makes sense for you.
Step 1: Understand conversational advertising instead of results lists
On Google, users see a list of links and choose for themselves. In ChatGPT, your ad is part of the conversation. The person has just described specifically what they need, such as a table for a small office or software for a team of ten people. It is precisely this context that makes advertising relevant, but it also requires sensitivity. An ad that doesn't fit the question feels intrusive and gets you nothing. So first think about what situation your customer is in before you configure anything in the system.
Step 2: Replace keywords with conversation descriptions
The most important change in thinking concerns targeting. You don't book search terms, you describe conversations. Ask yourself what questions people ask before they even start googling. For your offer, collect five to ten typical questions that your customers actually ask in sales conversations. These give rise to the conversation topics you target. This way you reach people in an early decision phase, that is, exactly when they are still open to recommendations.
Step 3: Create an account and build the first campaign
Start by creating your account in the OpenAI Ads Manager. Set up the ad account and keep the basic structure clear: one account, within it a campaign with a clear goal, within it the ad with your offer. In the first campaign, work with exactly one offer instead of several at the same time. Only that way can you clearly see later what worked. The course guides you through this setup using a continuous example, so that you can then transfer the steps to your own product.
Step 4: Set Context Hints and understand the principle of the auction
Context Hints are cues with which you tell the system in which conversations your offer fits. Phrase them as specifically as possible: What problem do you solve, for whom, and in what situation? The more precise the description, the better it can be assessed whether your ad fits a conversation. The auction then decides which ad appears at a particular moment. Bid and relevance work together here. The course explains the details thoroughly. To start with, it is enough to internalize the basic principle: A high bid alone is not enough if the offer doesn't fit the conversation.
Step 5: Set up clean measurement
Before you release budget, define how you will measure success. The metric everyone stares at is often just an intermediate value. Clicks or impressions say little if they don't lead to inquiries or closed deals. Set a goal that fits your business, such as qualified inquiries or appointment bookings, and check it regularly. Set up measurement before the launch and not only afterwards. Retroactively, a clean comparison can hardly be established.
Step 6: Honestly assess whether the channel suits you
Not every offer is suitable for advertising in conversations. If your customers rarely ask an AI before buying, or your offer requires a great deal of explanation, this channel may be a poor fit. A clear warning sign is if you cannot describe the relevant conversations cleanly. You should also be alert if your measurement delivers no usable results after a reasonable test period. In this case, it is worth taking a pause and reassessing before you invest any more budget.
Summary – Running ChatGPT Ads: From Your First Account to a Measurable Campaign
Advertising in ChatGPT works differently from search engine advertising. Your ad appears within the conversation, you work with conversation descriptions instead of keywords, and the auction determines the moment of delivery. If you start with a clear account setup, formulate Context Hints precisely, and define your measurement before launch, you build a campaign that can be evaluated. The head start is currently large, but it only applies to those who have understood the mechanics and apply them cleanly.
FAQ
Where are ads displayed in ChatGPT?Your ad appears within a conversation and not in a classic results list like on Google.
Since when and where do ads run in ChatGPT?According to the video, they have been running since the end of August 2026, including in Austria, Germany, and Switzerland.
Do I need keywords for ChatGPT ads?No, you do not book search terms, but describe conversation situations in which your offer is the right answer.
What are Context Hints?Context Hints are cues with which you tell the system in which conversations your offer is relevant. The more specifically they are formulated, the better the delivery fits.
Which metric should I observe first?Do not rely on a single number. Set a goal before launch, such as qualified inquiries, and measure against it.
Should I test multiple offers at the same time?To start, it makes more sense to begin with a single offer. That way it remains clear which measure contributed to the result.
