ChatGPT Ads in Austria: What companies need to know

ChatGPT Ads have been available through self-service in Austria since the end of August 2026. Alongside search advertising, social ads and traditional display advertising, this creates a new digital channel: companies can reach people while they research possibilities, compare offers and prepare decisions in ChatGPT.

At first glance, this may sound like “Google Ads for ChatGPT”. The reality is more nuanced. In a conversation, people often describe their needs in greater detail than they would in a search query. They explain their situation, requirements, uncertainties and alternatives. This gives advertisers richer context – and a greater responsibility to communicate in a relevant, transparent and useful way.

This article explains how ChatGPT Ads currently work, which businesses may benefit from an early test and which foundations should be in place before the first euro is spent.

Our ChatGPT Ads service explains how MOREMEDIA® combines suitability assessment, campaign architecture, context hints, ads, landing pages and measurement into one managed offering.

In brief: ChatGPT Ads are paid, clearly labelled advertisements that may appear below suitable ChatGPT responses. They are technically and visually separate from the answer. Advertising spend cannot buy a mention and does not influence what ChatGPT says.

What are ChatGPT Ads?

ChatGPT Ads are paid placements within the ChatGPT interface. They may appear after a response and are identified as sponsored content. An ad can include the advertiser’s name, favicon, title, descriptive copy, image and a link to the destination page.

OpenAI emphasises three principles: ads remain separate from answers, they must be clearly recognisable as advertising, and advertisers do not receive access to personal chats. This is central to the acceptance of the format because people also use ChatGPT for sensitive and personal questions. Advertising must not undermine that trust.

The offering is still at an early stage. Functions, eligible industries, markets, optimisation models and formats may change quickly. Platform-related statements therefore need a date and should be verified again before every campaign launch.

Why the channel may matter to businesses

A conventional search often begins with only a few words: “CRM for SMEs”, “web design agency Linz” or “best accounting software”. In a ChatGPT conversation, the same person might provide much more context:

“We are a mid-sized manufacturing company with several locations and need a CRM our sales team can use without extensive training. Which solutions should we consider?”

A short search phrase becomes a described decision situation. This is the core potential of ChatGPT Ads. A company can appear not merely around a term, but at a moment when requirements, alternatives and expected benefits have already been articulated.

Context does not replace strategy. The ad still needs to present a credible offer, suit the situation and lead to a page that fulfils its promise. Generic advertising language and an unsuitable landing page do not become effective simply because they appear in an AI product.

How ad relevance is determined

ChatGPT Ads do not rely exclusively on conventional keywords. Current information indicates that the system can consider several relevance signals, including:

• the context of the current conversation;
• the ad title and copy;
• the content and quality of the landing page;
• context hints;
• available geographic and audience settings;
• additional permitted signals when personalisation is enabled.

Context hints describe what the company offers, who it helps and when it may be useful. They are not exact-match keywords and do not guarantee delivery for a specific question. They provide the system with understandable relevance guidance.

Ad groups should therefore not be organised solely around product names. A stronger structure reflects specific offers, audiences, problems and decision situations.

ChatGPT Ads, Google Ads and social ads

ChatGPT Ads do not automatically replace established channels. They add another form of intent and context to the media mix.

Google Search Ads

Google Search Ads respond to an active query. Search terms provide a strong signal, but often little context. Advertisers work with keywords, search intent, copy, bidding and landing pages.

Social ads

Social ads reach people through audience characteristics, interests, behaviour and platform signals. Demand may not yet have been expressed. Strong creatives and immediate attention are particularly important.

ChatGPT Ads

ChatGPT Ads can appear during an existing research or decision process. The conversational context may be richer than a search term. This can increase relevance, but it also means the ad needs to match the expressed need very precisely.

The practical question is not simply “Google or ChatGPT?”. A useful test compares channels against the same commercial objective: What kind of demand is reached? How valuable are the visits and enquiries? What is the cost per qualified action? Which stage of the customer journey does each channel support?

ChatGPT Ads and GEO are not the same thing

ChatGPT Ads and Generative Engine Optimization are often confused, but they serve different purposes.

GEO improves content, data and digital presence so that AI systems can understand a brand, represent it correctly and potentially cite it as a source. This is organic visibility. A mention cannot be bought or guaranteed.

ChatGPT Ads are paid, clearly labelled advertisements. They appear separately from generated answers and do not influence their content.

The disciplines can complement each other:

• GEO strengthens organic clarity and credibility.
• ChatGPT Ads create additional paid visibility in suitable decision situations.
• A technically sound, understandable website supports both organic discoverability and the quality of advertising landing pages.

A ChatGPT Ads campaign never purchases a recommendation from ChatGPT.

Who can see ChatGPT Ads?

At present, ads may appear in the ad-supported Free and Go offerings. Higher paid plans such as Plus, Pro, Business, Enterprise and Edu remain ad-free. OpenAI does not show ads to accounts identified as belonging to people under 18.

This matters when estimating potential reach. The entire ChatGPT user base is not automatically available as advertising inventory. General user numbers should not be used as a reach forecast. The relevant figures are the users actually reachable in a market and the placements permitted for the advertiser’s category.

Not every industry and context is eligible

OpenAI excludes sensitive, harmful and brand-unsafe conversations from ad placements. Advertising content is also subject to restrictions. During the early phase, the focus is primarily on selected consumer categories, local services, travel and experiences, digital products and education. Regulated or sensitive categories may be prohibited or require individual review.

Every pilot therefore needs a policy review at three levels:

1. Advertiser: Is the business model eligible and the identity correctly verified?
2. Ad and landing page: Are claims, images, offer and destination consistent and compliant?
3. Placement context: Is advertising permitted in that type of conversation?

Claims must be truthful. Exaggerated outcomes, false certifications, misleading pricing and an invented affiliation with OpenAI are not acceptable.

Building a sensible campaign

A robust test starts with the business objective, not with Ads Manager.

Define a measurable action

Should the campaign produce a qualified enquiry, demo booking, registration, purchase or download? Awareness can be legitimate, but it is more difficult to evaluate than a clearly defined action.

Clarify the offer and decision situation

What exactly is being promoted? Who is it relevant to? In which situation does the need arise? What problem does it solve? One focused offer is usually more suitable for a pilot than an entire portfolio.

Structure ad groups around context

Different audiences and use cases belong in separate ad groups. Context hints, ads and landing pages can then be written more precisely.

Develop genuinely different ad variants

Variants should not merely exchange individual words. Each ad needs a distinct angle: benefit, use case, objection, audience or differentiator.

Provide an appropriate landing page

The destination must continue the ad’s promise immediately. It needs a clear headline, understandable arguments, trust signals, a visible next step and a frictionless mobile experience.

Set a learning budget and stopping rules

A pilot is a learning project. Budget, duration and the criteria for scaling, adapting or ending the test should be agreed before launch.

Why the landing page is decisive

New platforms draw attention to targeting and ad formats. Yet the bigger lever often lies after the click. A visitor arrives with a specific expectation. If the destination is only a generic service overview, they must make the connection to their own need themselves.

A strong ChatGPT Ads landing page should:

• immediately continue and deepen the ad’s main promise;
• make the intended audience recognisable;
• explain services or outcomes concretely;
• answer objections and important questions;
• provide relevant references or other credible proof;
• offer one clear primary action;
• load quickly and work flawlessly on mobile devices.

It also needs to be technically accessible. Redirects, security rules, consent mechanisms and crawler settings must not unintentionally prevent review or delivery.

What else matters in complex B2B decisions

Many discussions of new advertising channels stop at ad formats, targeting and tracking. For services that require explanation, that is where the real work begins. In established SMEs and hidden champions, decisions are rarely made by one person alone. Management, marketing, sales and – depending on the offer – IT, data protection or procurement assess different aspects of the same investment.

MOREMEDIA® therefore considers the entire decision chain rather than ad delivery alone. The following MOREMEDIA® working model is not a feature or official scoring method provided by OpenAI. It is a practical framework that helps companies make better decisions before and during a controlled launch.

Plan for the decision group, not only the target audience

A target audience describes whom a campaign should reach. A decision group also shows who must approve the investment, evaluate it professionally, enable it technically or work with it later. On a landing page, those roles do not all require the same information.

• Management needs a clear business rationale, a realistic view of risk and explicit ownership.
• Marketing needs a credible position, brand consistency and criteria for relevant rather than merely numerous leads.
• Sales needs to understand the problem behind an enquiry, the maturity of the decision and the most useful next step.
• IT, data protection or procurement, where involved, need clear information about data flows, integration, operation, approvals and follow-on costs.

A persuasive destination organises this evidence without becoming an overloaded company presentation. The main promise remains clear, while deeper information answers the questions of the people who will later take part in the decision. This makes it easier for the person who clicked the ad to explain and progress the offer internally.

Build a context library from real customer language

Strong context hints should not be invented in isolation. A more reliable foundation comes from the language customers already use – for example in anonymised enquiries, meeting notes, frequently asked questions, objections, on-site searches or existing search-term reports. This does not involve reading private ChatGPT conversations; advertisers do not receive access to them.

A short context card is created for every relevant decision situation:

1. role or company type;
2. specific research trigger;
3. desired outcome;
4. constraint or objection;
5. required evidence;
6. appropriate destination and next action.

Together, these cards form a version-controlled context library. Context hints, ad variants, landing-page sections and new content topics can then be derived consistently from it. Personal data does not belong in this library; source notes must be anonymised and reduced to professionally relevant patterns before use.

The work remains useful even if ChatGPT Ads are not continued after the first test. The same library can improve FAQs, SEO and GEO content, Google Ads structures, sales guidance and service copy. This reduces the risk of investing only in one young platform.

Maintain message match through to the sales handover

The connection must not break after the click. Each ad group therefore receives a documented chain:

Decision situation → context hint → ad → landing page → contact and sales handover

The wording need not be identical at every stage, but the promise, intended audience and next step must retain the same meaning. If an ad promises an evidence-based channel assessment, the destination must not suddenly become a generic agency presentation. The person handling the first response must also know which offer and context led to the enquiry.

This end-to-end message match reduces expectation gaps, supports internal buy-in and reveals the handover at which qualified demand is being lost.

Evaluate lead quality through consistent sales feedback

Platform metrics show what happens up to the conversion. Whether an enquiry is genuinely suitable often becomes clear only during the first contact. Sales or consulting should therefore assess each usable lead through a small, consistent framework:

• Does the company fit the defined target profile?
• Does the actual problem fit the promoted offer?
• Does the contact influence the decision or have access to the decision group?
• Is there an identifiable project horizon?
• Has a concrete next commercial step been agreed?

Each criterion can be assessed internally as `0 = unknown or unsuitable`, `1 = partial` and `2 = clearly suitable`. This value is not an automated judgement about people and is not an OpenAI metric. Its purpose is simply to give marketing and sales a common language for lead quality.

Insights are fed back to the campaign team regularly and only in sufficiently aggregated form. Optimisation can then focus on qualified conversations, genuine opportunities and traceable commercial potential rather than the highest possible number of form submissions.

Distinguish incremental demand from channel shifting

A new platform may generate new demand – or receive credit for contacts that would have arrived through Google, direct traffic or another campaign anyway. No single attribution model can answer that question completely. MOREMEDIA® therefore combines several indicators:

• consistent UTM and ad identifiers;
• duplicate checks in the CRM;
• separate records for first known contact, last interaction and qualified opportunity;
• a voluntary source question during the first conversation;
• comparison of branded and non-branded demand as well as direct traffic;
• time- or region-based comparison groups, but only where data volume is sufficient.

With small samples, these findings remain directional and must not be presented as causal proof. They are still valuable because they prevent the last click from receiving all the credit and simple channel migration from being sold as incremental growth.

The MOREMEDIA® decision-readiness check

Before launch, five questions are assessed as `0 = absent`, `1 = partly in place` or `2 = clearly in place and verified`:

1. Offer: Is one relevant offer clearly defined and commercially sensible?
2. Decision situations: Are roles, triggers, objections and desired outcomes documented?
3. Evidence and message match: Does the landing page answer the decisive questions and continue the ad promise consistently?
4. Measurement and sales feedback: Do consent, conversion measurement, CRM attribution and lead-quality feedback work reliably?
5. Ownership and response: Is one person responsible for budget, approvals, data quality and timely handling of enquiries?

The total is guidance, not a performance forecast:

• 8–10 points: The foundations for a limited, controlled launch are in place.
• 5–7 points: Close the clearly identified gaps before reassessing.
• 0–4 points: Do not invest media budget yet; first make the offer, destination or measurement dependable.

This process gives management a transparent investment decision, creates a shared view of quality for marketing and sales, and provides technical stakeholders with clear requirements. The value does not come from launching ads as early as possible, but from learning the right things early.

Measuring ChatGPT Ads

Ads Manager currently provides metrics such as impressions, clicks, spend, click-through rate, average CPC, average CPM and conversions. Depending on the objective, buying or optimisation models such as CPM, CPC and conversion-optimised click campaigns may be available.

Conversion measurement can combine:

• URL and UTM parameters;
• existing web analytics;
• OpenAI Pixel;
• Conversions API;
• unique event IDs to avoid duplicate conversions.

The technical setup is only half the task. Teams first need to decide which events have commercial value. A button click is not yet a qualified enquiry. Meaningful reporting links platform data with completed forms, booked meetings, qualified leads and actual sales outcomes.

Privacy and consent

OpenAI states that advertisers cannot access users’ chats, chat history, memories or personal details. Companies receive aggregated campaign data.

Standard data-protection obligations still apply on the advertiser’s own landing page. Pixel-based, analytical or server-side conversion measurement requires a review of data flows, legal basis, consent, privacy notices and contractual relationships. Technical implementation does not replace legal advice.

At minimum, testing should cover consent not yet given, optional tracking declined, optional tracking accepted, successful and unsuccessful form submissions, and conversions reported through both browser and server.

When an early pilot may be appropriate

A ChatGPT Ads pilot may be especially relevant when:

• the offer requires explanation;
• people compare several options before deciding;
• the product or service provides a clear practical benefit;
• a focused landing page exists or can be created;
• a measurable action is available;
• the company has sufficient learning budget;
• the category and advertising claims are permitted.

Starting immediately is less useful when the offer is unclear, no suitable destination exists, enquiries cannot be measured, current campaigns already fail for fundamental reasons, the sole motivation is to “do something with AI”, or a guaranteed return is required before any test.

The most important pilot metrics

CTR and CPC are useful, but insufficient. A pilot should assess delivery, reaction, website behaviour and business outcomes.

• Delivery: impressions, CPM and consistency of delivery.
• Reaction: clicks, CTR, CPC and differences between variants.
• Website behaviour: engaged sessions, relevant content interaction, form starts and technical errors.
• Business results: completed conversions, qualified enquiries, cost per qualified lead and resulting sales opportunities.

Qualitative learning is equally valuable during an early platform phase: Which messages work? Which situations fit? Which audiences respond? What expectations do visitors bring to the landing page?

Common mistakes

• copying a Google Ads structure without adapting it to conversational context;
• using only one ad per offer;
• linking to a generic homepage;
• implementing measurement only after launch;
• presenting a paid placement as a ChatGPT recommendation;
• confusing curiosity about a new platform with a business strategy.

The right first question is not “How quickly can we launch?” but “Which testable hypothesis do we want to examine?”

A controlled pilot in six steps

1. Assess suitability: Review offer, category, policies, market and technical foundations.
2. Formulate a hypothesis: Define audience, decision situation, value and conversion goal.
3. Prepare destination and tracking: Test message match, mobile UX, consent and events.
4. Build the campaign: Create ad groups, context hints and several distinct variants.
5. Launch with limits: Use the agreed learning budget and monitor data quality.
6. Make an evidence-based decision: Scale, adapt or stop according to predefined criteria.

Conclusion: Test early, but do not start blindly

ChatGPT Ads offer another way to reach people who are researching, comparing and preparing decisions. Conversational context may make these encounters valuable because it contains more information than a short search phrase.

The channel does not remove the fundamentals of effective advertising. A clear offer, useful ads, suitable creatives, a convincing landing page and reliable measurement remain essential. New requirements are added: context hints, platform policies, the strict separation between advertising and answers, and careful communication around trust and privacy.

For SMEs and hidden champions, the sensible approach is therefore a controlled pilot with a clear hypothesis. Companies that prepare the foundations properly can learn early whether ChatGPT Ads deserve a role in their marketing system.