ChatGPT Ads arrives in Spain: what it means for a B2B company and how to measure it
OpenAI opened ChatGPT Ads for self-serve in Spain on 31 August 2026. What changes for a B2B company, what to expect, and how to measure it in sales.
On 31 August 2026, OpenAI opened ChatGPT Ads for self-serve in Spain and in 30 other European countries. Since then, any company can create an account and advertise inside ChatGPT conversations.
It is a new channel. It is not a revolution and it is not a fad. It is one more place where buyers are, with its own rules and its own traps. This article explains what it is, which kind of B2B company it makes sense for, and how to avoid the usual mistake: measuring it in clicks.
What it is and what it is not
ChatGPT Ads shows ads inside ChatGPT’s answers, based on what the user is asking about. If someone asks how to choose a logistics provider for Europe, a logistics operator can appear there.
What makes it different from Google:
- The intent is richer. On Google, the user types three words. In ChatGPT, they describe their situation. The platform knows more about the context.
- There is no results page. The ad competes with an answer, not with nine links.
- There is less competition. It is a new channel. Pricing, inventory and practices are still forming.
What it is not: it is not a replacement for Google or LinkedIn. It is one more channel, with its own place in how the budget is split.
Which B2B companies it makes sense for
Not all of them. It is worth being honest about that.
It makes sense when the buyer does research before getting in touch. Professional services, software, machinery, logistics, consulting. Purchases where someone spends days comparing options and asks long questions. In those sectors ChatGPT is already part of the research process, and being in the answer matters.
It makes less sense when the purchase is local and immediate, or when search volume is so low that no platform accumulates enough data to optimise.
The mistake that is about to be repeated
Every time a new channel appears, the same cycle repeats. An account is opened, a small budget is tested, cost per click and cost per lead are checked, compared against Google, and a decision is made.
And it is the wrong comparison. A new channel brings a different kind of user. Its leads may be more expensive and better, or cheaper and worse. Cost per lead will not tell you. Only the CRM will, weeks later, once you know how many qualified and how many bought.
If ChatGPT Ads is judged on cost per lead in the first two weeks, the decision will be close to random.
How to measure it properly from day one
The pieces are the same as for any other channel. The difference is that they have to be in place before the first euro is spent, not after.
1. Tag the traffic
Every ad in ChatGPT leads to a page with parameters identifying the channel, the campaign and the ad. Without this, the lead enters the CRM as “direct” or “unknown” and there is nothing to measure.
2. Store the origin in the CRM
The form captures those parameters and saves them with the lead. If the CRM has no field for it, create one before launching.
3. Apply the same definition of qualified
The lead from ChatGPT goes through the same criteria as the one from Google or Meta. Same salesperson, same process, same definition. If it is treated differently, the comparison is worthless.
4. Wait for the sales cycle
If the company’s sales cycle is six weeks, there is no conclusion to be drawn before six weeks. Measure by cohort: sales are assigned to the month their lead came in, not the month they closed.
5. Compare on cost per sale
In the end, the only valid comparison between ChatGPT Ads, Google and Meta is cost per qualified lead and cost per sale. Nexus measures all three channels on the same criteria and shows the comparison in a single dashboard.
A sensible test budget
For a company spending from €5,000 a month on advertising, a ChatGPT Ads test that actually means something:
| Element | Criterion |
|---|---|
| Budget | Enough for 30–50 leads, no fewer |
| Duration | A full sales cycle, plus two weeks |
| What is compared | Cost per qualified lead and per sale against Google and Meta |
| Decision | Keep, scale or close, based on the CRM number |
With fewer than 30 leads there is no sample. With less than one sales cycle there are no sales to count.
The other side: showing up without paying
ChatGPT Ads is the paid part. There is an organic part that matters just as much or more: getting ChatGPT and other assistants to cite your company when someone asks about your category. This is known as AI visibility, or GEO.
You cannot buy it. It depends on how your website is structured, on whether the information about your company is clear and consistent, and on whether there are sources that mention it. It is work similar to SEO, with rules that are still being written. It deserves the same caution: no promises about positions or results, and tracking of which questions bring your company back and which do not.
What is coming
Meta has said it intends for any advertiser to be able to create and target ads entirely with artificial intelligence during 2026. Google keeps pushing Performance Max, whose adoption has gone from 60% to 71%. OpenAI is entering advertising. All three platforms are heading the same way: the advertiser supplies the signal, the machine decides the rest.
In that context, the advantage for a B2B company is not in mastering every platform. It is in having a clean signal of which leads buy, and sending it to all of them.
Before you open the account
If your company is thinking about testing ChatGPT Ads, the question to answer first is not how much to spend. It is whether the CRM is ready to tell you, two months from now, how many sales it brought. The diagnostic checks that and tells you what is needed for the test to be measurable.