ChatGPT Ads. Reach people while they're still describing the problem.

Two separate things happened, and most articles blur them together. On 24 August 2026, ChatGPT Ads went live for users across 31 European markets, the Netherlands included. Dutch people on the free plan are seeing sponsored cards in ChatGPT right now.

Buying them is a different story. There's no self-serve Ads Manager in the Netherlands yet — you cannot open your own account and launch a campaign today. In Europe, access currently runs through OpenAI's Ads Solutions team and its agency and technology partners, with self-serve rolling out after.

A ChatGPT ad is one clearly labelled sponsored card that can appear underneath a relevant ChatGPT response. It carries your name and logo, a headline of 3–50 characters, up to 100 characters of body copy, one image and a link to your page. No video. No carousel. No interactive formats. One card, one message, very little room.

And the part everyone gets wrong: you cannot buy your way into the answer. No budget makes ChatGPT recommend you inside its response. Ads sit next to the answer, never in it, and OpenAI is explicit that advertising doesn't influence what ChatGPT says. If someone offers to get your brand mentioned by ChatGPT for a fee, that's not advertising — that's a different conversation (and usually a worse one).

Health and financial services: United States only

OpenAI opened both categories in July 2026 — consumer medical devices, dental, hospitals and urgent care, health insurance, medical testing, vision; auto loans, credit cards, deposit accounts, insurance, investments, mortgages, payments — but only in the US, with manual review and licence verification. Outside the US, health and financial services advertising is generally not allowed. For a Dutch advertiser, that means both categories are effectively closed for now.

We'll say the obvious thing: healthcare and fintech are two of the sectors we work in most, so this one stings for us too. If that's you, GEO AI Visibility is currently the better route to AI visibility — it has no category restrictions, and your buyers are reading AI answers whether or not anyone can advertise next to them.

And even when your category is fine

Ads don't appear in sensitive conversations at all: mental health, personal health, emotionally reliant contexts, and anything OpenAI's brand-safety rules flag as unsuitable. That puts a hard edge on the whole premise of this channel. The more personal someone's problem is, the less likely there's any advertising next to it. Which is probably right — but it does mean "people describe their problems to ChatGPT" isn't the same as "you can be there for all of them."

Two rules that catch people out on creative

• Your landing page has to match your ad. Approved creative pointing at a page that introduces different content gets pulled.
• Your ad may not look like part of ChatGPT. No imitating the interface, no styling it as if it were the answer.

From search intent to conversational intent

Say you sell software that helps companies organise their sales process. On Google, someone searches "CRM software small business". They already know their need and they've already translated it into a product category. Strong commercial intent, and expensive because everyone else knows that too.

The same person in ChatGPT might open with: "We've got six people in sales and we're basically running everything out of a spreadsheet. Leads slip through, and I have no idea which quotes are still open. How should we organise this?" They never say the word CRM. They may not know the category exists. But the problem is dead-centre relevant to you — and nobody is bidding on it, because it isn't a keyword. That's conversational intent, and it's the actual product here.

This is the one to understand before you spend anything. A highly relevant conversation does not mean a ready buyer. Someone might ask ChatGPT: "Why is our living room still cold when the heating's been on all day?" Genuinely relevant if you sell insulation. But it is a completely different person from the one googling "cavity wall insulation quote". The second one is close to converting. The first one is forming an opinion.

That's the trade. ChatGPT Ads buys you relevance earlier in the decision — while someone is researching the problem, discovering that solutions exist and deciding what "good" looks like. Sometimes that's more valuable than a late-funnel click. Often it converts more slowly, or needs a different landing page and a softer next step than a quote request. So we don't treat ChatGPT Ads as a cheaper Google. It's a new moment in the journey: problem → solution → comparison → decision. Google Ads mostly owns the last two. This is a way into the first two.

The most common mistake is treating ChatGPT Ads like Google Search with a different interface. You don't buy keywords. You write context hints — descriptions of the conversations, needs and situations where your product is a reasonable next step. They're not exact-match, and they don't guarantee delivery.

Weak hint

"CRM software"

Strong hint

"Sales teams of five to fifteen people still running their pipeline in spreadsheets, losing follow-ups, unsure which quotes are still open, and looking for a way to get visibility without a six-month implementation."

The difference isn't length. It's that the second one describes a person in a situation, so the system can recognise the dozen different ways someone might describe that same problem.

You don't win this auction by outspending

ChatGPT Ads runs a relevance-weighted, second-price auction. Your bid is multiplied by a relevance score built from four things together: your context hints, your ad title, your ad copy and your landing page. A lower bid at high relevance beats a higher bid at low relevance. Which means the work is in the specificity, not the budget. OpenAI's own starting guidance sits around $3–5 CPC or $60 CPM. If you're consistently clearing well above that, the usual cause isn't a bid that's too low — it's hints that are too broad.

So a ChatGPT Ads strategy doesn't start with keyword research. It starts with mapping:

• What problems do your customers actually have?
• How do they describe those problems before they know the jargon?
• What are they trying to achieve?
• Which alternatives are they weighing?
• What trade-offs are they making?
• And at which moment are you a sensible next step?

A scaled e-commerce test

Roughly $60k spend over 15 days, 1.49x blended ROAS, ~$1.72 CPC, 2.35% conversion rate — with daily ROAS swinging between 0.2x and 2.9x.

A SaaS tool test

97,000+ impressions, 1.30% CTR, very few sign-ups.

A B2B agency test

8,940 impressions, 0.65% CTR, 58 clicks, zero qualified leads.

Likely a good fit when

• your product solves a clearly describable problem
• customers need explanation or advice before they buy
• several solutions are possible and get compared
• the right answer depends on the customer's situation
• there's real consideration before purchase
• a lead or customer is worth enough to justify testing a young channel

We see the most interesting potential in SaaS and tech, business services, business software, e-commerce with a consideration phase, training and education, energy and sustainability, automotive, travel, and more complex consumer products.

Likely a weak fit when

• purchases are impulse-driven or very low value
• almost no research happens before buying
• your buyers are overwhelmingly on paid ChatGPT plans
• your category isn't allowed to advertise at all — legal services, recruitment and housing listings, gambling, alcohol and adult content are out everywhere, and health and financial services are US-only

But industry isn't really the deciding factor. The better question is: how likely is it that your customer talks their problem through with ChatGPT before they buy?

A short assessment of conversational intent, problem–solution fit, consideration depth, customer value and whether your site is ready to be part of the targeting.

More and more companies ask us the same question: "How do we get ChatGPT to mention us?" With GEO (Generative Engine Optimization) we improve the odds that AI systems understand who you are, what you offer and when you're a relevant source. But there's no guarantee ChatGPT names you organically — and that hits smaller and lesser-known brands hardest. You can be the perfect answer to someone's question and still lack the authority to surface on your own.

ChatGPT Ads creates a second route. GEO grows your organic visibility in AI; ChatGPT Ads buys visibility in relevant conversations right now. You don't have to be the best-known company in your market to be relevant to the problem someone is describing. That's why we don't treat GEO and ChatGPT advertising as competing disciplines. They're two halves of one AI visibility strategy.