Index
What is ChatGPT Ads and where do the ads appear
Why it is relevant for a used car dealership
How targeting works: context hints instead of keywords
Available ad formats for dealerships
How to create an account in OpenAI Ads Manager
How to upload stock as a product feed
How to structure your first campaign
How to measure results: pixel, conversions, and attribution
How it differs from Google Vehicle Ads
What to expect in the first 30 days
Dealcar and ChatGPT Ads
Frequently Asked Questions
What is ChatGPT Ads and where do the ads appear
ChatGPT Ads is OpenAI's advertising system integrated within ChatGPT. The ads appear below the model's response, in a separate block labeled as sponsored, without mixing with the text generated by the AI.
There are three points that are worth clarifying from the start because they often cause confusion. First: the ad is not part of the response. The AI responds the same whether there is advertising or not. Nobody can pay to have ChatGPT change what it says. Second: only adult users on the Free and Go plans see them. Paid plans (Plus, Pro, Business, Enterprise, Education) do not show ads. Third: during the current beta, stock feed products are only eligible for ads, they do not appear in organic ChatGPT conversations.
The channel arrived in Spain and 30 other European markets in August 2026, with self-service access from OpenAI Ads Manager (ads.openai.com). Spain has over 8.5 million weekly ChatGPT users according to OpenAI data, and the percentage using the free plan is the one that can see the ads.
Why it is relevant for a used car dealership
A used car buyer comes to ChatGPT at a different stage than when they go to Google. On Google, they type "used SUV Madrid under 20000 euros". On ChatGPT, they explain their situation: they need a car to take two kids to school, have an 18,000 euro budget, don't want petrol, and are concerned about long-term reliability. All of that information comes in the same conversation.
Read also how car search is changing with ChatGPT in 2026.
This changes the value of each impression. An ad that appears when someone is describing exactly the car they are looking for intercepts intent at a moment of active consideration, not in a generic exploratory search. Early advertisers in other industries sharing data report CTRs between 1.5% and 3.2%, well above the 0.35% average for conventional display.
For a dealership with used stock, the opportunity is specific: a buyer who tells ChatGPT "I'm looking for a diesel family car with under 80,000 km and under 15,000 euros in Valencia" can see, right below the model's response, an ad with the exact car that dealership has in stock.
How targeting works: context hints instead of keywords
The most important technical difference between ChatGPT Ads and Google Ads is the targeting system. Google Ads works through keywords: the advertiser bids on keywords that users type into the search engine. ChatGPT Ads works through context hints: natural language descriptions of the potential buyer's situations, needs, and use cases.
A context hint for a used car dealership could be: "someone looking for a second-hand family car with a good maintenance history and under 100,000 km for daily city use." It is not a keyword; it is a description of the buyer. OpenAI's system matches that description with the context of active conversations and determines in which conversations to display the ad.
Read also what ChatGPT Ads context hints are and how to use them to sell cars.
This has two practical consequences. First, there is no need to research keywords: you must understand who your dealership's buyers are and describe their situations accurately. Second, the targeting is more contextual and less mechanical: you cannot control exactly which searches your ad appears in, but you can describe the desired buyer profile in great detail.
Available ad formats for dealerships
There are two main formats available in the current beta.
Standard text ads include the advertiser's name, a headline, a description, and a link to the landing page. It is the simplest format and works well for brand campaigns or general lead generation.
Product ads generated from a feed show direct vehicle information: car image, title (make, model, version), price, and a link to the vehicle details page. This is the most relevant format for dealerships because it allows each car in stock to appear as its own ad without needing to create each ad manually.
The product format is the one that best fits the operations of a dealership: stock changes constantly, prices update, and cars come and go. A feed-based system synchronises these changes automatically in the ads.
How to create an account in OpenAI Ads Manager
Access to OpenAI Ads Manager is available at ads.openai.com. The registration process requires an OpenAI corporate account (different from a personal ChatGPT account), a valid payment method, and passing account validation, which can take one to two business days.
Some industries have additional eligibility restrictions: financial services and health have specific requirements. Used car dealerships are not among the restricted categories that OpenAI has published to date.
Once the account is validated, the Ads Manager has a structure similar to Meta Ads: campaign level (objective and budget), ad group level (targeting by context hints), and ad level (creatives or product feed).
How to upload stock as a product feed
The product feed is the file containing the dealership's inventory in a structured format. OpenAI reads this file and automatically generates product ads for each vehicle.
Check the car product feed technical guide for ChatGPT Ads.
There are three upload methods available depending on technical level and volume. Manual CSV or TXT is the entry point for testing the channel with a small number of cars: you prepare the file, upload it directly in Ads Manager, and that's it. The drawback is that products expire after two weeks if not updated, requiring you to repeat the process frequently if your stock changes.
The hosted URL is the recommended method for most dealerships: the stock is published on a public URL in a structured format and OpenAI reads it periodically. When a car is sold or the price changes, the URL updates and the feed automatically reflects the change.
The SFTP connection is for technical teams who want to automate server-to-server transfers. The host is sftp.commerce.openai.com on port 443.
The minimum fields required for each vehicle in the feed are: item_id (the car's unique identifier, usually the VIN or registration plate), title (make, model, version, and year), description (main features), url (link to the vehicle details page on the dealership's website), image_url (main photo of the car), price (price with currency), availability (in stock or out of stock), seller_name, and seller_url.
The availability field is the most critical: if a car is sold and the feed is not updated, the ad remains active, and buyers arrive at a page for a car that no longer exists. That is why the hosted URL method synchronised with the dealership's management system is the most robust option.
How to structure your first campaign
For a dealership starting with ChatGPT Ads, the simplest and easiest structure to evaluate is a product feed campaign with the dealership's full stock and a single ad group with a broad context hint.
The campaign objective can be clicks (CPC) or conversions (oCPC if the pixel is installed). For the first campaign, the click objective is easier to manage: you pay for each click received and can analyse which cars generate the most engagement without needing to set up conversion events from day one.
The minimum daily budget to get useful data in a local market like a Spanish city is between 10 and 20 euros a day. Below that, the volume of impressions may be insufficient to optimise.
The context hints for the first ad group should describe the profile of the dealership's most frequent buyer. If the dealership mainly sells budget city cars and hatchbacks, the context hint should describe someone looking for an affordable car for daily city use. If it sells SUVs and family cars, the context hint describes a family looking for space and reliability.
How to measure results: pixel, conversions, and attribution
OpenAI Ads Manager offers native metrics: impressions, clicks, CTR, CPC, CPM, and conversions. Conversions require installing the OpenAI pixel on the dealership's website.
The pixel is installed with a snippet of JavaScript code in the head of all pages on the website. Once installed, you set up the events you want to measure: visits to the vehicle details page, contact form submissions, and clicks on the WhatsApp button are the three most relevant events for a dealership.
To measure traffic coming from ChatGPT Ads without needing the pixel, you can use the parameter oppref=chatgpt in destination URLs, which allows you to identify visits originating from ChatGPT Ads in Google Analytics.
Check also how to measure leads and sales from ChatGPT Ads in a dealership.
Attributing a sale to a conversation initiated on ChatGPT is the most complex challenge of the channel: the buyer may have seen the ad on ChatGPT, visited the website days later from Google, and called to close the sale. None of those three interactions has direct attribution to ChatGPT Ads in a conventional last-click model. To measure the real impact of the channel, the most practical method is to ask the buyer at the time of their visit how they found the dealership.
How it differs from Google Vehicle Ads
Google Vehicle Ads (VLA) and ChatGPT Ads are complementary channels, not substitutes. The main difference lies in the stage of the buying process at which they capture the buyer.
Check also the complete Google VLA guide for dealerships.
Google VLA intercepts active search: the buyer has already decided they want a 2022 Toyota RAV4 with under 80,000 km and searches for it directly. ChatGPT Ads intercepts the consideration phase: the buyer is still defining what type of car they need, and ChatGPT helps them decide.
In terms of intent, Google VLA captures high and specific intent. ChatGPT Ads captures contextual and conversational intent, which can be very high if the buyer is in the decision phase, or lower if they are in the exploratory phase.
Google VLA targeting is by keywords and vehicle model. ChatGPT Ads targeting is by context hints and buyer profile. Both systems can work in parallel without cannibalising each other.
For a dealership with a limited budget that is already using Google VLA with good results, ChatGPT Ads is an additional channel to test, not a replacement. For one that does not use VLA, it is advisable to activate VLA first (where intent is more direct) and add ChatGPT Ads as a second layer.
Check also Dealcar Boost, the digital visibility channel for dealerships.
What to expect in the first 30 days
The channel is in beta, and results vary significantly depending on the sector, geographical area, and stock profile. The first 30 days should be understood as a learning phase, not a scaling phase.
What usually happens in the first week: the feed is validated, ads start serving, and the first impressions appear. The initial volume may be low depending on the budget and competition in the area.
What can be analysed at the end of the first month: which cars in the feed generate the most clicks, which context hints work best, what the average CTR of the ads is, and whether visits generated by ChatGPT Ads show different behaviour on the website (time on page, page views) compared to traffic from other channels.
What cannot be concluded yet in 30 days: whether the channel consistently generates sales. The buying cycle for a used car can be several weeks, and attribution is complex. The most useful metric in the first month is whether leads from ChatGPT Ads are of sufficient quality to justify the cost per click.
Dealcar and ChatGPT Ads
Dealcar is working on integrating stock into the product feed format compatible with ChatGPT Ads, so that cars published on Dealcar can be served as ads on ChatGPT without needing to prepare the feed manually.
If you want to see how it works and be among the first dealerships to activate the channel, visit dealcar.io/dealcar-boost/chatgpt-para-concesionarios.
Frequently Asked Questions
Is ChatGPT Ads currently available in Spain?
Yes. OpenAI included Spain in its expansion to 31 European markets in August 2026. The self-service Ads Manager has been available at ads.openai.com for Spanish advertisers since late August 2026, although the platform is still in beta.
How much does it cost to advertise on ChatGPT Ads?
The available buying models are CPM (cost per thousand impressions, from approximately $25), CPC (cost per click, with a recommended bid of $3 to $5), and oCPC (optimisation for conversions, available since August 2026). There is no stated minimum spend for the current self-service model.
Do cars from the feed also appear in ChatGPT's organic responses?
No. During the current beta, feed products are only eligible for ads. They do not appear in ChatGPT's organic conversations. OpenAI has mentioned that it might offer this capability in the future.
Do I need my own website to advertise on ChatGPT Ads?
Yes. Each ad directs to a landing URL which must be a page on the dealership's website: the vehicle details page, the contact page, or the stock page. A dealership without its own website can use the URL of their profile on a portal, but they lose control over the landing experience and the ability to install the tracking pixel.
How many cars do I need in the feed to start?
According to published independent tests, a feed with just six products works without errors. OpenAI's documentation suggests a minimum of 1,000 products for scaled feed campaigns, but early experiments with small catalogues show that the system accepts smaller feeds in the current beta.
Index
What is ChatGPT Ads and where do the ads appear
Why it is relevant for a used car dealership
How targeting works: context hints instead of keywords
Available ad formats for dealerships
How to create an account in OpenAI Ads Manager
How to upload stock as a product feed
How to structure your first campaign
How to measure results: pixel, conversions, and attribution
How it differs from Google Vehicle Ads
What to expect in the first 30 days
Dealcar and ChatGPT Ads
Frequently Asked Questions
What is ChatGPT Ads and where do the ads appear
ChatGPT Ads is OpenAI's advertising system integrated within ChatGPT. The ads appear below the model's response, in a separate block labeled as sponsored, without mixing with the text generated by the AI.
There are three points that are worth clarifying from the start because they often cause confusion. First: the ad is not part of the response. The AI responds the same whether there is advertising or not. Nobody can pay to have ChatGPT change what it says. Second: only adult users on the Free and Go plans see them. Paid plans (Plus, Pro, Business, Enterprise, Education) do not show ads. Third: during the current beta, stock feed products are only eligible for ads, they do not appear in organic ChatGPT conversations.
The channel arrived in Spain and 30 other European markets in August 2026, with self-service access from OpenAI Ads Manager (ads.openai.com). Spain has over 8.5 million weekly ChatGPT users according to OpenAI data, and the percentage using the free plan is the one that can see the ads.
Why it is relevant for a used car dealership
A used car buyer comes to ChatGPT at a different stage than when they go to Google. On Google, they type "used SUV Madrid under 20000 euros". On ChatGPT, they explain their situation: they need a car to take two kids to school, have an 18,000 euro budget, don't want petrol, and are concerned about long-term reliability. All of that information comes in the same conversation.
Read also how car search is changing with ChatGPT in 2026.
This changes the value of each impression. An ad that appears when someone is describing exactly the car they are looking for intercepts intent at a moment of active consideration, not in a generic exploratory search. Early advertisers in other industries sharing data report CTRs between 1.5% and 3.2%, well above the 0.35% average for conventional display.
For a dealership with used stock, the opportunity is specific: a buyer who tells ChatGPT "I'm looking for a diesel family car with under 80,000 km and under 15,000 euros in Valencia" can see, right below the model's response, an ad with the exact car that dealership has in stock.
How targeting works: context hints instead of keywords
The most important technical difference between ChatGPT Ads and Google Ads is the targeting system. Google Ads works through keywords: the advertiser bids on keywords that users type into the search engine. ChatGPT Ads works through context hints: natural language descriptions of the potential buyer's situations, needs, and use cases.
A context hint for a used car dealership could be: "someone looking for a second-hand family car with a good maintenance history and under 100,000 km for daily city use." It is not a keyword; it is a description of the buyer. OpenAI's system matches that description with the context of active conversations and determines in which conversations to display the ad.
Read also what ChatGPT Ads context hints are and how to use them to sell cars.
This has two practical consequences. First, there is no need to research keywords: you must understand who your dealership's buyers are and describe their situations accurately. Second, the targeting is more contextual and less mechanical: you cannot control exactly which searches your ad appears in, but you can describe the desired buyer profile in great detail.
Available ad formats for dealerships
There are two main formats available in the current beta.
Standard text ads include the advertiser's name, a headline, a description, and a link to the landing page. It is the simplest format and works well for brand campaigns or general lead generation.
Product ads generated from a feed show direct vehicle information: car image, title (make, model, version), price, and a link to the vehicle details page. This is the most relevant format for dealerships because it allows each car in stock to appear as its own ad without needing to create each ad manually.
The product format is the one that best fits the operations of a dealership: stock changes constantly, prices update, and cars come and go. A feed-based system synchronises these changes automatically in the ads.
How to create an account in OpenAI Ads Manager
Access to OpenAI Ads Manager is available at ads.openai.com. The registration process requires an OpenAI corporate account (different from a personal ChatGPT account), a valid payment method, and passing account validation, which can take one to two business days.
Some industries have additional eligibility restrictions: financial services and health have specific requirements. Used car dealerships are not among the restricted categories that OpenAI has published to date.
Once the account is validated, the Ads Manager has a structure similar to Meta Ads: campaign level (objective and budget), ad group level (targeting by context hints), and ad level (creatives or product feed).
How to upload stock as a product feed
The product feed is the file containing the dealership's inventory in a structured format. OpenAI reads this file and automatically generates product ads for each vehicle.
Check the car product feed technical guide for ChatGPT Ads.
There are three upload methods available depending on technical level and volume. Manual CSV or TXT is the entry point for testing the channel with a small number of cars: you prepare the file, upload it directly in Ads Manager, and that's it. The drawback is that products expire after two weeks if not updated, requiring you to repeat the process frequently if your stock changes.
The hosted URL is the recommended method for most dealerships: the stock is published on a public URL in a structured format and OpenAI reads it periodically. When a car is sold or the price changes, the URL updates and the feed automatically reflects the change.
The SFTP connection is for technical teams who want to automate server-to-server transfers. The host is sftp.commerce.openai.com on port 443.
The minimum fields required for each vehicle in the feed are: item_id (the car's unique identifier, usually the VIN or registration plate), title (make, model, version, and year), description (main features), url (link to the vehicle details page on the dealership's website), image_url (main photo of the car), price (price with currency), availability (in stock or out of stock), seller_name, and seller_url.
The availability field is the most critical: if a car is sold and the feed is not updated, the ad remains active, and buyers arrive at a page for a car that no longer exists. That is why the hosted URL method synchronised with the dealership's management system is the most robust option.
How to structure your first campaign
For a dealership starting with ChatGPT Ads, the simplest and easiest structure to evaluate is a product feed campaign with the dealership's full stock and a single ad group with a broad context hint.
The campaign objective can be clicks (CPC) or conversions (oCPC if the pixel is installed). For the first campaign, the click objective is easier to manage: you pay for each click received and can analyse which cars generate the most engagement without needing to set up conversion events from day one.
The minimum daily budget to get useful data in a local market like a Spanish city is between 10 and 20 euros a day. Below that, the volume of impressions may be insufficient to optimise.
The context hints for the first ad group should describe the profile of the dealership's most frequent buyer. If the dealership mainly sells budget city cars and hatchbacks, the context hint should describe someone looking for an affordable car for daily city use. If it sells SUVs and family cars, the context hint describes a family looking for space and reliability.
How to measure results: pixel, conversions, and attribution
OpenAI Ads Manager offers native metrics: impressions, clicks, CTR, CPC, CPM, and conversions. Conversions require installing the OpenAI pixel on the dealership's website.
The pixel is installed with a snippet of JavaScript code in the head of all pages on the website. Once installed, you set up the events you want to measure: visits to the vehicle details page, contact form submissions, and clicks on the WhatsApp button are the three most relevant events for a dealership.
To measure traffic coming from ChatGPT Ads without needing the pixel, you can use the parameter oppref=chatgpt in destination URLs, which allows you to identify visits originating from ChatGPT Ads in Google Analytics.
Check also how to measure leads and sales from ChatGPT Ads in a dealership.
Attributing a sale to a conversation initiated on ChatGPT is the most complex challenge of the channel: the buyer may have seen the ad on ChatGPT, visited the website days later from Google, and called to close the sale. None of those three interactions has direct attribution to ChatGPT Ads in a conventional last-click model. To measure the real impact of the channel, the most practical method is to ask the buyer at the time of their visit how they found the dealership.
How it differs from Google Vehicle Ads
Google Vehicle Ads (VLA) and ChatGPT Ads are complementary channels, not substitutes. The main difference lies in the stage of the buying process at which they capture the buyer.
Check also the complete Google VLA guide for dealerships.
Google VLA intercepts active search: the buyer has already decided they want a 2022 Toyota RAV4 with under 80,000 km and searches for it directly. ChatGPT Ads intercepts the consideration phase: the buyer is still defining what type of car they need, and ChatGPT helps them decide.
In terms of intent, Google VLA captures high and specific intent. ChatGPT Ads captures contextual and conversational intent, which can be very high if the buyer is in the decision phase, or lower if they are in the exploratory phase.
Google VLA targeting is by keywords and vehicle model. ChatGPT Ads targeting is by context hints and buyer profile. Both systems can work in parallel without cannibalising each other.
For a dealership with a limited budget that is already using Google VLA with good results, ChatGPT Ads is an additional channel to test, not a replacement. For one that does not use VLA, it is advisable to activate VLA first (where intent is more direct) and add ChatGPT Ads as a second layer.
Check also Dealcar Boost, the digital visibility channel for dealerships.
What to expect in the first 30 days
The channel is in beta, and results vary significantly depending on the sector, geographical area, and stock profile. The first 30 days should be understood as a learning phase, not a scaling phase.
What usually happens in the first week: the feed is validated, ads start serving, and the first impressions appear. The initial volume may be low depending on the budget and competition in the area.
What can be analysed at the end of the first month: which cars in the feed generate the most clicks, which context hints work best, what the average CTR of the ads is, and whether visits generated by ChatGPT Ads show different behaviour on the website (time on page, page views) compared to traffic from other channels.
What cannot be concluded yet in 30 days: whether the channel consistently generates sales. The buying cycle for a used car can be several weeks, and attribution is complex. The most useful metric in the first month is whether leads from ChatGPT Ads are of sufficient quality to justify the cost per click.
Dealcar and ChatGPT Ads
Dealcar is working on integrating stock into the product feed format compatible with ChatGPT Ads, so that cars published on Dealcar can be served as ads on ChatGPT without needing to prepare the feed manually.
If you want to see how it works and be among the first dealerships to activate the channel, visit dealcar.io/dealcar-boost/chatgpt-para-concesionarios.
Frequently Asked Questions
Is ChatGPT Ads currently available in Spain?
Yes. OpenAI included Spain in its expansion to 31 European markets in August 2026. The self-service Ads Manager has been available at ads.openai.com for Spanish advertisers since late August 2026, although the platform is still in beta.
How much does it cost to advertise on ChatGPT Ads?
The available buying models are CPM (cost per thousand impressions, from approximately $25), CPC (cost per click, with a recommended bid of $3 to $5), and oCPC (optimisation for conversions, available since August 2026). There is no stated minimum spend for the current self-service model.
Do cars from the feed also appear in ChatGPT's organic responses?
No. During the current beta, feed products are only eligible for ads. They do not appear in ChatGPT's organic conversations. OpenAI has mentioned that it might offer this capability in the future.
Do I need my own website to advertise on ChatGPT Ads?
Yes. Each ad directs to a landing URL which must be a page on the dealership's website: the vehicle details page, the contact page, or the stock page. A dealership without its own website can use the URL of their profile on a portal, but they lose control over the landing experience and the ability to install the tracking pixel.
How many cars do I need in the feed to start?
According to published independent tests, a feed with just six products works without errors. OpenAI's documentation suggests a minimum of 1,000 products for scaled feed campaigns, but early experiments with small catalogues show that the system accepts smaller feeds in the current beta.
