ChatGPT Ads in Spain: Access, Cost and Measurement

ChatGPT Ads is already rolling out in Spain. OpenAI has opened European access initially through its Ads Solutions team, partner agencies and technology partners, while it prepares self-serve access. The new channel lets advertisers create impression, click and conversion campaigns inside ChatGPT, with delivery based on relevance to the conversation's context.

The novelty isn't placing a conventional ad next to a chatbot. The ad can appear below the response while a person is explaining what they need, comparing alternatives, narrowing down constraints and getting closer to a decision. The brand doesn't buy the response and can't influence it: it competes for a separate advertising slot through a combination of bid, relevance, creative, landing page and conversational intent.

In this guide we cover what a business needs to know before investing: how ChatGPT Ads works, how much it can cost, what context hints are, how to set up OpenAI Pixel and Conversions API, what metrics Ads Manager offers, and how to separate paid advertising from organic visibility in AI engines.

The essentials in 30 seconds

• Availability: OpenAI included Spain in its expansion to 31 European markets during the week of August 24, 2026.
• Access: the initial European rollout runs through OpenAI Ads Solutions, agencies and technology partners; self-serve access will follow later.
• Placement: ads appear below the response, clearly labeled and separated from organic content.
• Delivery: the system weighs context, intent, ad, landing page, context hints and the expected probability of achieving the result.
• Buying: OpenAI supports CPM and CPC, plus conversion optimization with cost-per-click billing.
• Measurement: OpenAI Pixel, Conversions API, oppref, UTM parameters, web analytics and CRM can be combined.
• Fundamental limit: paying for ChatGPT Ads does not get a brand mentioned or recommended in the response.
• Recommendation: before scaling, validate access, consent, instrumentation, lead quality and real economic value.

What is ChatGPT Ads?

ChatGPT Ads is OpenAI's advertising product for showing relevant ads inside ChatGPT. According to OpenAI's basic ads documentation, the format appears below a response and includes the advertiser's name, logo or favicon, a headline, a description, an image and a landing page.

OpenAI Ads Manager is the tool used to create and manage accounts, campaigns, ad groups, creatives, budgets, bids and conversions. It's worth distinguishing the two concepts: ChatGPT Ads is the inventory and the advertising experience; Ads Manager is the operating platform.

Advertising runs through a system separate from the model that generates the response. OpenAI states that advertisers cannot alter, rank or condition what ChatGPT answers. Seeing an ad also does not mean OpenAI recommends or endorses the advertised product.

This separation matters. A company can get paid impressions without being organically mentioned by ChatGPT, and can appear in organic responses even without advertising. Mixing the two would produce false attribution.

Can you already advertise on ChatGPT from Spain?

OpenAI announced on August 18, 2026 the expansion of ChatGPT Ads to 31 European markets, including Spain, France, Germany, Italy, the Netherlands, Sweden, Norway, Denmark and Austria. The company said that initial access for European advertisers would run through its Ads Solutions team, partner agencies and technology partners, and that self-serve access via Ads Manager would open later during the summer.

So the short answer is yes: the rollout in Spain is underway. Actual onboarding availability, inventory and features may vary during the beta depending on the account, the market and the type of advertiser. The safest route is checking eligibility on the official OpenAI Ads portal or working with a partner that already has operational access.

Ads are shown on the Free and Go plans. Plus, Pro, Business, Enterprise and Edu subscriptions remain ad-free. OpenAI also does not show ads to accounts identified as belonging to users under 18, and limits delivery in sensitive contexts or contexts incompatible with its brand safety policies.

The European expansion turns what were initial US tests into a channel that Spanish advertisers should now at least add to their experimentation agenda. That doesn't mean every business should invest immediately: availability and some features may still change during the beta.

If a business already has a concrete opportunity and needs to validate access, instrumentation and launch, it can check our ChatGPT Ads management service. This guide, by contrast, keeps an informational intent: understanding the channel and deciding with good judgment whether it's worth testing.

Demand exists before the tools can measure it well

The channel's arrival has created a search category almost from scratch. At Elevam we ran a demand study on August 21, 2026 covering 49 queries related to ChatGPT Ads: 45 in Spanish and four in English. They included "agencia ChatGPT Ads", "agencia ChatGPT Ads España", "contratar agencia ChatGPT Ads", "OpenAI Ads España" and "publicidad en ChatGPT para empresas".

The term "ChatGPT Ads" did show the most visible comparative signal in Google Trends, although it wasn't included as an exact term in the Keyword Planner batch. The conclusion isn't that demand doesn't exist, but that the category is too recent for traditional tools to have enough history.

How advertising works in ChatGPT

ChatGPT Ads' delivery system selects ads based on expected relevance and results. Among the signals OpenAI states it uses are:

• The context and intent of the current conversation.
• The ad's headline, description and image.
• The landing page.
• The context hints provided by the advertiser.
• Certain signals from the user's broader experience when ad personalization is enabled.
• The estimated probability of reaching the campaign result.

Eligible ads compete in a relevance-weighted second-price auction. The bid matters, but it doesn't act alone: an ad with an offer loosely related to the conversation, a generic page or an imprecise message can lose competitiveness even if the advertiser is willing to pay more.

This logic brings ChatGPT Ads closer to intent-based marketing, but conversational intent carries more information than a short query. A person can explain their goal, budget, prior experience, constraints, location and decision criteria all at once. For the advertiser, that richness increases the opportunity for relevance — and also the cost of a superficial strategy.

What context hints are, and why they aren't keywords

Context hints are notes the advertiser adds at the ad-group level to describe conversations, topics or situations where its product could be useful. The official guide to creating ad groups clarifies that they help the system interpret how well the offer fits, but they are not exact-match keywords, closed audience rules, or instructions to appear against a specific question.

A weak context hint stays at the category level: "accounting software." A more useful one adds need, audience and situation: "finance leads at SMBs comparing accounting software to automate invoicing and bank reconciliation." The second formulation offers more context without pretending to control the user's exact words.

To build them, the brand needs to turn its commercial research into observable scenarios:

• What the customer is trying to solve.
• What alternatives they're comparing.
• What constraints shape the purchase.
• What objections come up before deciding.
• What feature is a dealbreaker.
• What situation triggers the need.
• What outcome they expect.

Each ad group should focus on one product category, need or intent area. If two contexts require different messages or landing pages, they should be split. Mixing unrelated uses makes it impossible to know which hypothesis is driving results.

How to create ads that are useful for a conversation

The standard ad combines brand, favicon, headline, text, image and destination URL. In the ad creation documentation, OpenAI recommends these ranges:

The space constraint forces prioritization. The headline needs to explain what the product offers; the description needs to give a concrete reason to consider it. Generic formulas — "transform your business," "the solution you need" — waste the context and make it harder for the user to understand what the ad is for.

The landing page is part of the relevance signal and of the post-click experience. Sending every campaign to the corporate homepage tends to break continuity. A conversation about a specific need should lead to a page that answers that same need, keeps the ad's message and makes the intended action easy.

Creative should be tested as a system, not as a single piece. For each intent area, vary at least the core promise, how specific the headline is, the visual approach, the proof or trust signal and the landing page. The goal isn't to produce dozens of indistinguishable ads, but to identify which combination of context, message and destination drives a business outcome.

How much does it cost to advertise on ChatGPT?

There is no universal fixed rate for ChatGPT Ads in Spain. Cost depends on the objective, the bid, relevance, competition, market, inventory and the campaign's ability to generate the expected result.

OpenAI supports campaigns bought on CPM to optimize for impressions and campaigns on CPC to optimize for clicks. For click campaigns it recommends starting with a maximum bid of between 3 and 5 dollars. That figure is bid guidance, not a guaranteed CPC, a mandatory minimum, or a forecast of Spanish costs.

The right question isn't just how much a click costs, but how much value it produces. A high CPC can be profitable if the conversation reveals advanced intent and the campaign generates customers with enough margin. A low CPC can be expensive if it drives traffic that doesn't fit, junk form submissions, or non-incremental sales.

Campaign objectives: impressions, clicks and conversions

The objective chosen should match the maturity of the measurement setup:

• Impressions or reach: suited for testing delivery, awareness and contextual fit when there isn't yet enough conversion signal.
• Clicks: useful for validating messages, landing pages and the ability to generate qualified traffic.
• Conversion optimization: geared toward a configured standard action, with cost-per-click billing and optimization toward the selected event.

Conversion-optimized campaigns need an event with enough volume and a healthy tracking setup. If the chosen conversion happens too rarely, is poorly implemented or represents a shallow action, the algorithm will have a weak signal to work with. In an early phase it may be smarter to validate delivery and traffic quality first, rather than demanding a stable cost per acquisition from the system without enough data.

How to measure ChatGPT Ads conversions

Ads Manager lets advertisers record post-click actions such as a purchase, a lead or a sign-up. OpenAI's conversion measurement documentation states that setup starts from a data source and can send events via OpenAI Pixel, Conversions API, or both.

OpenAI appends a click identifier called oppref to the destination URL. The Pixel can capture it and store it in a first-party cookie to link a later conversion back to the ad. Where available, the same identifier can be sent via Conversions API.

The most resilient implementation combines browser and server. If a conversion is sent via both Pixel and API, both events need to share the same event identifier so OpenAI can deduplicate them. It's also necessary to preserve oppref through redirects and internal navigation; losing it reduces attribution capability.

Ads Manager currently reports impressions, clicks, spend, CTR, average CPC, average CPM and conversions. Data can be viewed by campaign, ad group and ad, in tables, charts and CSV exports. OpenAI notes that attributed conversions can take between 24 and 48 hours to appear.

OpenAI's conversion figure doesn't have to match GA4, the CRM or another advertising platform exactly. Differences can come from attribution windows and models, time zones, consent, storage, deduplication and modeled conversions. A discrepancy alone doesn't prove the implementation is wrong.

The five-layer dashboard

At Elevam we split how we read ChatGPT Ads into five layers, to keep platform metrics from replacing the business outcome:

The fifth layer should not be attributed to the ad campaign. Its job is to keep a brand from confusing paid exposure with organic authority.

Privacy, consent and measurement quality

OpenAI supports advanced matching using normalized, encrypted first-party information, and offers an automatic mode where the Pixel detects certain compatible data in forms, normalizes it and applies SHA-256 before sending it. The documentation states that unencrypted data is not transmitted through that feature.

This doesn't remove the advertiser's obligations. OpenAI requires that conversion data only be shared where permitted, that users receive clear information, and that the necessary consents are obtained. In Spain and the EU, implementation needs to be coordinated with the consent management system, the privacy policy and the advertiser's legal judgment.

Measurement should be designed before launch, not rebuilt after money has already been spent. At minimum, a business should validate:

• Which events will be sent and under what name.
• Which action counts as the primary conversion.
• How Pixel and Conversions API will be deduplicated.
• How oppref will be preserved.
• What UTM parameters will distinguish campaign, group and creative.
• How forms, e-commerce and CRM will be connected.
• How each lead's quality will be classified.
• What attribution window and model will be used to decide.

ChatGPT Ads vs. Google Ads, Meta Ads and GEO

ChatGPT Ads shares elements with other advertising systems, but occupies a different place in the user experience.

GEO stands for Generative Engine Optimization: the work aimed at increasing the probability that a brand gets retrieved, mentioned, recommended or cited by AI engines. It doesn't buy inventory and can't promise a specific response.

To observe that dimension, Antropus lets you track mentions, recommendations, citations and URLs separately. It's an SEO and GEO measurement platform built by Elevam and is therefore first-party information, not an independent auditor of the agency.

Asier López Ruiz also created the CREF framework to align brand authority with the signals AI models interpret. At the replicable-measurement layer, the HSA Protocol — Human · Search · AI — was developed by Asier López Cabañas at Elevam Labs, with strategic review and business validation from Asier López Ruiz. The operational distinction is simple: Ads Manager measures advertising; a GEO methodology measures organic presence. Only afterward can both layers be read together within a single acquisition strategy.

You can also read how to correct what ChatGPT says about your company, the organic half of this same conversation.

How to launch a ChatGPT Ads campaign, step by step

1. 01

   Define the business outcome

   The campaign needs to start with an action and a value: purchase, demo, qualified lead, sign-up or intent-driven visit. "Try ChatGPT Ads" is not a measurable objective.

2. 02

   Verify access, policies and billing

   Before producing creative, confirm the account and country are eligible, complete business verification when required, and review the policies that apply to the sector and the landing page.

3. 03

   Set up measurement

   Configure Pixel, Conversions API, events, UTM, oppref, analytics and CRM. Test every event before turning on spend.

4. 04

   Research intent areas

   Analyze searches, sales conversations, sales calls, objections, support tickets and comparisons. The output should be a map of needs, not a list imported from a keyword tool.

5. 05

   Structure campaigns and groups

   Separate objectives, countries, platforms and intent areas. Each group should bring together ads that can share a coherent message and landing page.

6. 06

   Write context hints

   Naturally describe who needs the product, in what situation and for what outcome. Avoid isolated terms, exaggerations and contexts the product can't actually solve.

7. 07

   Create ads and landings

   Prepare several combinations of headline, description and image. Route each ad to the most relevant page, with a verifiable proposal and a clear action.

8. 08

   Launch a bounded experiment

   Define budget, minimum duration, hypothesis and decision criteria. Don't change context, ad, offer and landing at the same time if you later want to understand what drove the variation.

9. 09

   Evaluate quality, not just volume

   Trace every lead or sale back to its source, check commercial quality, and review attribution discrepancies before scaling.

10. 10

    Document what you learned

    Record what was tested, what happened, what limitations the sample had and what changes for the next iteration. On a beta platform, operational memory will be a more defensible advantage than a promise of results.

Which companies should try ChatGPT Ads now

The channel can be especially interesting when an intent the user can explain, an easy-to-understand offer and a measurable conversion all come together. Some profiles with an early fit include:

• E-commerce with a structured catalog and known margins. OpenAI already supports campaigns based on product feeds and states they have been among the best-performing formats in its program, though there isn't yet a sufficient independent benchmark.
• B2B software and services with a clearly measurable demo, trial or request.
• Education, travel, home, technology and comparable products, always within advertising policies.
• Brands that already have dedicated landing pages, analytics and a CRM capable of reporting quality and revenue.
• Companies able to experiment without demanding stable profitability from week one.

Companies that can't measure conversions, don't have a differentiated offer, rely on a generic homepage, operate in categories incompatible with the policies, or need immediate volume the beta still can't guarantee should wait or build the infrastructure first.

How we manage ChatGPT Ads at Elevam

At Elevam we treat ChatGPT Ads as an acquisition system, not an isolated inventory purchase. The work starts before opening Ads Manager and ends after the conversion the platform reports. The operation connects five disciplines: paid media (account structure, objectives, budget, bidding and optimization), conversational intent (research into problems, comparisons, constraints and decision situations), creative and landing pages (continuity between context, ad, offer and destination page), measurement (OpenAI Pixel, Conversions API, oppref, UTM, analytics and CRM), and business and GEO (lead quality, sales, margin and a separate read on organic presence).

To evaluate our work, or any other agency's, a business should demand evidence in seven areas: real access to Ads Manager and campaign experience; Pixel, Conversions API and analytics installation; intent research and command of context hints; creative and landing-page optimization capability; integration with Google Ads, Meta Ads, CRM and sales; separation between paid results and organic visibility; and reporting that connects spend, customers, revenue and margin.

On the organic side, MarketingDirecto placed Elevam first in Spain's first GEO agency ranking of 2026, with 86 out of 100 points and a methodology based on baselines across four AI models. The recognition doesn't guarantee ChatGPT Ads performance, but it does document the ability to tell paid media and generative visibility apart.

Our experience with US accounts, and the results from those campaigns, remain first-party information until we publish an anonymized case with a period, sample and methodology. We don't hide that: if a business requests an evaluation, the right test is to review access, setup, anonymized evidence and the limits of what the data can actually support.

Want to assess whether the channel fits your business? Check Elevam's ChatGPT Ads management or request an initial review of access, measurement, intent and viability.

A first-90-days plan

Conclusion

ChatGPT Ads introduces an advertising surface inside a moment of exploration and decision-making. Its potential comes from the context the user provides, but that same context penalizes generic campaigns: the offer, the ad, the context hints and the landing page all need to answer a shared need.

The Spanish opportunity is early and the platform is still in beta. Getting in early can speed up learning, but it doesn't replace instrumentation, judgment or statistical patience. The priority should be building a system that can measure from impression through to revenue, that separates advertising from organic visibility, and that scales only when the data supports the decision.

If you already have a measurable conversion, a clear proposal and the ability to experiment, the next step is to prepare a controlled test. On Elevam's ChatGPT Ads page we explain how we approach access, instrumentation, the campaign and optimization.

Frequently asked questions about ChatGPT Ads

Asier López Ruiz is CEO of Elevam and Antropus, a Spanish agency specializing in digital acquisition, paid media, SEO and Generative Engine Optimization. He created the CREF framework to align brand authority with the signals AI models interpret.