ChatGPT Ads in Europe: how does Ads Manager work?

ChatGPT Ads is now available across Europe.

OpenAI announced the expansion into 31 European markets, later updating the announcement on 31 August 2026 to confirm the rollout of self-service access through Ads Manager.

For European advertisers, this changes one thing in particular: ChatGPT Ads moves into a phase where campaigns can be configured and managed directly by advertisers.

The advertising product itself has also become more structured.

From the early tests focused mainly on impressions, OpenAI has expanded the platform to support CPM, CPC and conversion-focused optimisation, alongside geographic targeting, custom audiences and proprietary measurement tools.

For anyone already working with Google Ads or Meta Ads, several elements will feel familiar.

The distinctive part is the context in which delivery happens: a conversation where users can explain needs, criteria and constraints in much greater detail than through a single keyword.

How ChatGPT Ads Manager is structured

Ads Manager follows a fairly recognisable hierarchy.

At campaign level, advertisers define the objective, budget, campaign period, markets and, where relevant, the conversion event.

Within each campaign are ad groups, which organise products, services and different areas of intent.

Each ad group then contains the ads themselves.

The structure is therefore familiar to performance marketers, with one particularly interesting addition: context hints.

These signals help the system understand the types of conversations, needs and situations where the offer may be relevant.

For teams coming from paid search, this is one of the most important concepts to understand.

Context hints: how they differ from keywords

Context hints can include topics and terms related to a product, although their role covers a broader layer of meaning than a traditional keyword.

They describe what the product offers, who may benefit from it and in which situations it could be useful.

For example, a generic hint such as “running shoes” provides relatively little context.

“Cushioned daily running shoes for someone preparing for their first 5K” describes a much more specific need and gives the system more information to interpret potential relevance.

OpenAI's guidance for ad groups explains how context hints help define the relevant conversational environment rather than operating as exact-match search terms.

Delivery can also consider signals such as conversation intent and context, landing page, headline, ad copy and targeting settings.

When ad personalisation is enabled, some signals from the user’s broader ChatGPT experience may also contribute to relevance.

This makes ChatGPT Ads particularly interesting for categories where people tend to articulate their needs before making a decision.

Software, travel, ecommerce, professional services and products that require comparison are natural examples.

CPM, CPC and oCPC: which buying models are available

The evolution of bidding shows how quickly OpenAI is building ChatGPT Ads into a more complete performance platform.

Ads Manager currently supports three main buying and optimisation models.

With CPM, the campaign focuses primarily on reach and visibility, with cost calculated on impressions.

With CPC, the objective shifts towards traffic and the advertiser pays for valid clicks.

With oCPC, or conversion-optimised cost per click, the system optimises delivery towards a specific post-click action measured by the advertiser.

For teams familiar with Google Ads or Meta, the progression is recognisable: delivery can increasingly be tied to a commercial outcome.

This also makes the quality of the conversion signal increasingly important.

How targeting works

OpenAI has progressively expanded the targeting options available within the platform.

Campaigns can be configured by country and, where available, more granular geographic areas.

Ads Manager also allows advertisers to select the platforms where ads can appear, including web, iOS and Android.

Advertisers can also use custom audiences, including them in or excluding them from campaigns.

The distinctive layer remains conversational matching.

ChatGPT can understand that two users are expressing similar needs even when they describe them in very different ways.

Ad group structure should therefore start from intent and use cases, separating needs that require different messages or landing pages.

An ecommerce brand, for example, could distinguish between users looking for a product for a specific need, users comparing alternatives and users already close to making a decision.

What do ads on ChatGPT look like?

Ads appear separately from ChatGPT responses and are clearly labelled as advertising.

Creative assets include familiar elements such as advertiser name, headline, description, landing page, logo and image.

The way those assets are written still requires some adjustment to the conversational environment.

OpenAI recommends specific, benefit-led copy that quickly explains what the product offers, who it is relevant for and when it can be useful.

Variation also matters.

Producing many almost identical ads gives the system relatively little opportunity to understand which message is best suited to different conversational contexts.

More meaningful tests can explore distinct angles: benefit, use case, problem solved, differentiation, offer or proof.

Creative and context hints therefore work together to provide the system with richer information about relevance.

Why conversational context changes media planning

A traditional search can be extremely concise.

A user might type "CRM for small business" and leave the platform to infer much of the intent.

Inside ChatGPT, the same person could say: "I need a simple CRM for a five-person sales team, with no developers in-house, email automation and a Shopify integration".

For advertisers, this context is particularly valuable.

It captures the user while they are defining the problem, evaluating alternatives and establishing decision criteria.

This gives ChatGPT Ads potential across both demand capture and consideration.

The creative message needs to reflect that level of detail.

Generic advertising may struggle to feel relevant when it appears alongside a highly specific conversation.

Pixel and Conversions API: how performance is measured

Measurement has also become considerably more developed compared with the first months of the pilot.

Ads Manager allows advertisers to create a data source and send events using OpenAI Pixel, Conversions API, or both.

OpenAI’s conversion measurement documentation explains how events can represent actions such as purchases, leads or registrations.

OpenAI can then associate eligible events with advertising interactions within the applicable attribution window.

This allows advertisers to move beyond impressions and clicks and start measuring conversions and downstream outcomes.

For oCPC campaigns, the setup becomes particularly important because the system uses the selected conversion event to guide optimisation.

For advertisers planning to test ChatGPT Ads as a performance channel, tracking should therefore be part of the setup before campaign launch.

Privacy: what does the advertiser actually see?

The conversational nature of ChatGPT makes privacy especially important.

OpenAI maintains a clear separation between advertising and user conversations.

Users’ chats are not shared with advertisers, and advertising does not influence ChatGPT's answers.

The system can use the context of the current conversation to determine advertising relevance and, when personalisation is enabled, may consider some broader signals from the user’s ChatGPT experience.

Those signals remain within ChatGPT. Advertisers receive performance and measurement data needed to manage campaigns.

OpenAI’s documentation on ads in ChatGPT also explains that ads are shown on eligible consumer plans, while paid and business-focused tiers remain ad-free.

Brand safety and sensitive categories also have specific rules

The advertising system includes restrictions around sensitive contexts and regulated categories.

OpenAI excludes advertising from conversations considered particularly sensitive or unsuitable from a brand-safety perspective and applies specific policies to regulated sectors.

This matters especially for areas such as finance, healthcare and legal services, where eligibility, creative messaging and context may require additional checks.

For brands, the initial assessment should therefore include the offer itself, the creative and compliance with the platform’s Ads Policies.

Is ChatGPT Ads really similar to Google Ads and Meta Ads?

Operationally, several similarities are now clear.

The platform includes campaigns, ad groups, ads, budgets, bidding, audiences, creative and conversion tracking.

That reduces the learning curve for teams already managing performance campaigns.

The matching logic introduces a different layer: the system works with conversational intent and context hints, with less dependence on matching an individual keyword to an ad.

Simply importing the structure of a Search campaign is therefore unlikely to capture the full opportunity.

A more useful starting point is to organise campaigns around problems, contexts, use cases and decisions.

What should European advertisers test first?

The arrival of self-service access gives European brands a practical opportunity to start with controlled tests.

The strongest starting point is products or services where ChatGPT can naturally enter the decision journey, with each ad group centred around a clear intent and several genuinely different creative variations.

The landing page deserves the same attention.

The system considers the destination page among the signals used to assess relevance, so the ad message, context hints and page content should tell a consistent story.

On the measurement side, Pixel and Conversions API allow advertisers to move quickly beyond CTR and evaluate conversion rate, CPA and the quality of generated actions.

Qualitative analysis will also matter in the early stages: understanding which products, messages and intents find traction inside a relatively new conversational advertising environment.

From pilot to a new performance channel

When ChatGPT Ads first started appearing in early 2026, the main questions concerned ad placement, available inventory and whether advertisers would be able to measure meaningful business outcomes.

Many of those components are now in place.

Ads Manager is self-service across Europe, bidding includes traffic and conversion-focused objectives, custom audiences are available, and the platform supports Pixel and Conversions API.

The ecosystem is still young and likely to evolve quickly.

The current level of maturity is already enough for brands to start treating ChatGPT Ads as a performance channel worth testing, with clear objectives, tracking and hypotheses established before budget is committed.

The useful question for European advertisers in September 2026 therefore concerns where conversational context can add value to the media mix and which signals can demonstrate that value.