ChatGPT ads: should a local business advertise inside ChatGPT?

What ChatGPT ads are and where they appear

A ChatGPT ad is a paid placement that sits underneath ChatGPT's reply, labelled as sponsored and kept visually apart from the answer. It's the AI version of the ads above and below Google results, and it shares their basic rule: the paid placement is kept apart from the unpaid answer.

OpenAI started testing ads in the United States on 9 February 2026. Its help article on ads in ChatGPT says you may see an ad below a response, that you'll see one ad unit at a time (which can hold one or several images), and that it's also testing a multi-image ad format that shows while ChatGPT is generating an image. Every ad is labelled and shown separately from the response.

Not everyone sees them. Ads may appear for people on the Free and Go plans. Plus, Pro, Business, Enterprise and Edu accounts don't get ads, and nor do accounts OpenAI identifies as belonging to someone under 18. Temporary Chats don't show ads either, and Free users can switch to an ads-free mode in exchange for lower message limits and fewer tools.

That matters for a trade business. The people who see your ad are only a slice of ChatGPT users, so the audience is smaller than the headline user numbers suggest. Some of the homeowners and office managers who'd book you will be on a paid plan, or using ChatGPT through a work account, where no ads appear.

OpenAI also keeps ads away from certain chats. Its help article says ads aren't eligible to appear near sensitive topics such as personal health, mental health or politics, and political advertising isn't allowed at all. Health and financial services advertisers may be eligible under stricter criteria.

• Where: below a ChatGPT response, one ad unit at a time, plus a visual format being tested during image generation.
• Label: marked as sponsored and visually separated from the answer.
• Who can see them: logged-in adults on Free and Go plans in regions where ads run, plus logged-out users with ads suited to all audiences.
• Who can't: Plus, Pro, Business, Enterprise and Edu accounts, accounts identified as under 18, and Temporary Chats.
• What users control: hide or report an ad, see why it was shown, turn off ad personalisation, and clear ad data.

Are ChatGPT ads available in Australia?

Yes, for buying. At the time of writing (October 2026), OpenAI's Ads Manager availability table lists Australia as available for self-service, which means a business whose legal entity is based here can open an account and run campaigns. Whether a particular Australian user sees an ad still depends on their plan and eligibility.

The Australian rollout came in steps. OpenAI's post on testing ads in ChatGPT was first published on 9 February 2026 for the US test. An update dated 26 March 2026 said the next phase would expand beyond the US, starting with pilots in Canada, Australia and New Zealand. An August 2026 update added the United Kingdom, Mexico, Brazil, Japan and South Korea. OpenAI's ads FAQ describes a focused pilot from February through April, followed by self-serve onboarding in a beta Ads Manager.

What OpenAI hasn't published is just as useful to know. We found no Australian start date on its own pages, no Australian audience figures and no performance benchmarks. Its advertiser FAQ says plainly that, as an early platform, ChatGPT Ads doesn't yet have performance benchmarks across advertisers, industries or campaign types. So any agency or article quoting a typical return for Australian trades is guessing.

Don't assume a US article describes what you'll get here. Several guides ranking for this topic describe targeting options and buying conditions from the US launch. Check the availability table and the campaign settings in your own account before you plan anything, because the beta changes often and OpenAI says delivery, inventory, formats and buying options will keep changing.

Do ChatGPT ads change the answers?

No, according to OpenAI. Its ads help says ads don't influence ChatGPT's answers, ads run on separate systems from the chat model, and advertisers have no ability to shape, rank or alter ChatGPT's responses. It also says seeing an ad doesn't mean OpenAI endorses or recommends the advertiser.

For a local business, that cuts both ways. You can't buy your way into the answer. If someone asks ChatGPT which electricians near them handle switchboard upgrades, paying for an ad won't put your name in the list ChatGPT writes. The ad may show underneath, but the answer is still built from what ChatGPT finds and trusts. That's the job of getting recommended in ChatGPT, and it's a separate piece of work.

It also means a competitor can't buy their way above you in the answer. If ChatGPT already names your business when people ask about your kind of job, a rival's ad sits below that answer, labelled as sponsored.

How does OpenAI pick which ad to show? It says the system weighs expected relevance and outcomes. When several ads are eligible, it considers a mix of factors, including relevance and how much each advertiser bids. The signals it lists are below.

• The context and intent of the current conversation.
• The ad's landing page, title and copy.
• Context hints the advertiser writes to describe when the ad is relevant.
• The advertiser's targeting selections, such as location and platform.
• When the user has ad personalisation turned on, select signals from their wider ChatGPT use, such as past chats and memory.

How buying works: campaigns, objectives and location targeting

You buy ChatGPT ads in OpenAI's Ads Manager Beta. You sign in with an OpenAI account tied to a work email, set up billing, complete account verification, then build campaigns. Each campaign holds ad groups, and each ad group holds ads with a title, description, image and landing page. Every ad is reviewed against OpenAI's ad policies before it can run.

The objective you choose sets how you're charged. OpenAI's guide to creating campaigns for ChatGPT Ads lists three objectives: Views, Clicks and Conversions. Conversions campaigns optimise delivery toward an event such as a lead submission, but you're still billed per click or per thousand impressions, not per lead. You set either a campaign-total budget or a daily budget, and OpenAI explains that a daily budget is an average across a week, so some days can run up to twice the daily amount.

Targeting starts with country and can go narrower. The campaign settings are countries, platforms (Android app, Android web, desktop web, iOS app or iOS web) and optional custom audiences. At the time of writing (October 2026), the same help article says campaigns that don't use a product feed can also target supported locations such as states or regions, cities, markets and postal codes, where available. Availability varies by country, so the only way to confirm the Australian options is to search for locations during campaign setup in your own account. New product-feed campaigns stay at country level, and OpenAI documents no radius option. OpenAI also says some new self-serve accounts can only advertise in their home country at first, until identity verification is done and they've reached a set level of spend there.

Instead of keywords, ad groups use context hints: short phrases that describe what you offer, who it helps and when it's useful. OpenAI's own example is a plumbing business splitting emergency repairs and preventive maintenance into separate ad groups. It's clear that hints aren't exact-match keywords or targeting rules, and they don't guarantee delivery for particular conversations.

Location settings won't do all the sorting for a trade. A hypothetical Brisbane plumber may be able to pick Brisbane or a set of postcodes, if those options appear for Australia in their account. If they don't, a country-wide campaign can show the ad to someone asking about a blocked drain in Perth. Even a city or postcode setting rarely matches a real service area, and there's no radius to draw around the depot. The landing page has to do the sorting, by naming the area served at the top so out-of-area people leave instead of filling in the form.

The fit test: is a ChatGPT ads test worth it for your trade?

Run a test only if you can answer yes to all four questions below. If any answer is no, put the money and time into channels you can already measure, and come back when the beta matures.

Start with the first question, because most trades skip it. People ask ChatGPT for help with jobs they need to understand before they buy. Choosing a solar battery, planning a bathroom renovation or working out what a switchboard upgrade involves are research-heavy. A burst pipe at 10pm isn't. Someone in that spot usually rings a plumber they find on a map or in search, and a ChatGPT conversation never starts.

Then think about what one job is worth to you. ChatGPT Ads has no published benchmarks, so your first campaign is a learning cost. A trade with large, considered jobs can absorb that while it works out whether the channel produces enquiries. A trade with small, frequent, low-margin call-outs usually can't.

Take two hypothetical businesses. A solar and battery installer on the Sunshine Coast sells jobs that buyers research for weeks, the kind of decision someone might ask an AI tool to help them think through. That's a reasonable test case, provided the landing page states the service area clearly. A 24-hour locksmith in the same region sells urgent jobs to people who aren't chatting to an AI. Its money is usually better spent on Google Business Profile work and Google Ads.

1. Is the job researched in ChatGPT? Ask your recent customers how they found you and what they looked up first. If nobody mentions AI tools, wait.
2. Is one job worth enough to fund learning? A test with no benchmarks needs room to fail without hurting cash flow.
3. Can the location targeting open to you hold your service area? Search your area in Ads Manager to see which Australian cities or postcodes you can pick. If the closest fit still wastes many out-of-area clicks, you need a landing page that filters hard, or this isn't the channel yet.
4. Can you trace a lead to the ad? You need the OpenAI Pixel or Conversions API, tagged landing URLs and call tracking in place before you spend anything.

Paid placement versus organic citation in ChatGPT

Buying a ChatGPT ad doesn't earn a citation in ChatGPT's answers, and being cited doesn't get you ad space. They're separate systems with separate crawlers, and OpenAI's own documentation shows where the line sits.

OpenAI's overview of its crawlers lists two that matter here. OAI-SearchBot surfaces websites in ChatGPT's search features, and sites that block it won't be shown in ChatGPT search answers. OAI-AdsBot only visits pages submitted as ads, checks them against OpenAI's policies and may use their content to judge when the ad is relevant. Its data isn't used to train OpenAI's foundation models.

So an ad's landing page gets read for ad purposes only. It doesn't feed the answer. The organic route works the other way. When ChatGPT searches the web, its responses can include citations to sources, and it can use an approximate location to find local results. Whether your site is one of those sources depends on whether it can be crawled and on what your pages and other sites say about you. Our guide to AI recommendations versus mentions covers the difference between being named and being suggested.

Neither route is a promise. Nobody can guarantee a citation, and an ad can be outbid or fail review. The sensible order for most local businesses is organic first, because it keeps working when you stop paying, then a paid test once the basics are measured. That mirrors the choice between SEO and Google Ads, with the extra caution that this ad platform is still in beta.

How to measure a ChatGPT ads test before you spend

Set up measurement first, then launch. A test you can't measure tells you nothing, and with no published benchmarks, your own numbers are the only ones that count.

OpenAI supports conversion tracking through a JavaScript Pixel, a server-side Conversions API, or both. A Conversions campaign needs at least one supported standard event, such as a lead submission, and custom events aren't supported for that objective yet. Ads Manager reports impressions, clicks, spend, click-through rate, average cost per click, average cost per thousand impressions and conversions once tracking works. OpenAI notes that spend can show up to 7 to 8 hours late.

Keep ad clicks apart from free ChatGPT visits. Our guide to measuring ChatGPT referrals shows where organic ChatGPT visits land in GA4. Put your own UTM tags on every ad landing URL so paid visits land in a separate bucket. Then add call tracking, because many trade leads still ring rather than fill in a form.

Write your claims carefully. The ACCC says information a business gives about its services must be accurate, truthful and based on reasonable grounds, and that rule covers advertising and websites on any platform. A ChatGPT ad is no exception, and nor is the landing page it points to.

Decide in advance when you'll stop. Agree a budget you're happy to lose, a test period and the lead quality you need, then judge the test against those, not against the platform's own dashboard alone. Our guide to attributing leads across paid and organic shows how to compare channels on booked jobs instead of clicks.

1. Install the OpenAI Pixel or Conversions API and fire one test lead before launch.
2. Tag every ad landing URL with utm_source, utm_medium and utm_campaign so paid clicks don't mix with organic ChatGPT referrals.
3. Use a tracked phone number on the ad landing page and log which calls became quotes and booked jobs.
4. Name your service area at the top of the landing page so out-of-area visitors leave before enquiring.
5. Set a stop rule before launch: a fixed budget, a fixed period and a minimum number of qualified enquiries.
6. Review results against booked work in your job system, not only Ads Manager clicks.