If you use the free version of ChatGPT, you might have already noticed small “Sponsored” cards appearing under some of its answers. That started in Australia back in April. What’s less well known is that Australia is one of only seven countries (at the time of writing this) in the world where businesses can actually buy those ads right now.
We’ve just set up our first campaign, and this post is about why we think that’s worth your attention, even if you never spend a dollar on it.
Why this matters
Increasingly, when someone’s website breaks or they need a new accountant, they don’t Google it, they ask an AI chatbot. We’ve been banging on about this shift for a while with Answer Engine Optimisation (AEO). ChatGPT ads are the paid version of the same idea: your business appears in front of someone mid-conversation, right at the moment they’re describing a problem you solve.
Clicks cost more than Google (roughly $3 to $5 USD), but competition is thin, and the people seeing your ad have just described their problem in detail rather than typing three words into a search box.
How to get started
If you’ve ever poked around Google Ads or Meta’s Ads Manager, you’ll recognise almost everything here. Same campaign, ad group and ad structure, mostly the same jargon. Here’s the path, with the things we tripped over noted along the way.
Sign up and set up the account. Head to ads.openai.com. You’ll need an OpenAI account, and before any ads can serve you’ll need to complete billing and upload an account logo. Take care on the account details: your country, currency and time zone can’t be changed once the account is created. OpenAI’s account setup guide covers this step.
Create a campaign. You’ll choose an objective, either Reach (paying for impressions) or Clicks (paying per click). A Conversions objective is listed as “coming soon”. For most small businesses, Clicks is the one. Campaign type is Standard (you write the ad) or Product Feed, which generates ads from a product catalogue and is aimed at online stores. Budgets work like everywhere else: a total cap or a daily one.
Audiences (you can probably skip this). There’s a Create Audience tool for uploading a list of your customers’ emails or phone numbers, the same idea as Meta’s Custom Audiences, used to filter who sees your campaigns. Useful if you’re a store with a big customer list. A services business with a modest client base won’t get much from it yet, and we skipped it.
Conversion tracking (optional, but worth doing). Under Tools and then Conversions, you create a “data source”, which generates a tracking pixel: a script you paste into your website’s header (you can put this into the Elementor > Settings > Code section) so it loads on every page. Then you create a conversion event. Pick the standard event closest to your goal, something like “Lead created”, and it gives you a second, smaller piece of code. That one goes only on the page that confirms the action happened, like the thank you page after your contact form, not on every page, otherwise every visitor counts as a conversion. One quirk: the field that says “choose how long after an ad click to count conversions” can’t actually be changed, it’s fixed at 30 days for everyone. Presumably that gets fixed down the track.
Ad group and ad. Here you set a maximum cost per click (the platform warns you in real time if your bid is too low to deliver consistently), your destination URL (add UTM tags to it so your analytics can see what the ad traffic actually did), and your context hints, which deserve their own section below. The ad itself is one 50-character headline, one 100-character description and a square image. That’s 150 characters to make your case, so keep it plain and put the important words first.
Then you submit for review and wait. OpenAI’s Ads Manager overview in their help centre fills in anything we’ve skimmed here.
The new bit: context hints
Instead of keywords, you write “context hints”, plain English descriptions of the conversations where your business is relevant. You’re not programming a keyword filter. You’re briefing an AI on when you belong in the conversation, and it makes the call.
For our WordPress maintenance campaign, we didn’t write “wordpress maintenance melbourne”. We wrote things like: “Business owners whose web designer has disappeared or stopped responding and who need someone to take over their existing WordPress site.” That’s a situation, not a keyword. Someone mid-panic because their site broke isn’t going to type a tidy search phrase, but they will describe exactly that to ChatGPT.
When you hit that blank context hints box, might be a good idea to ask an AI chatbot to help fill it in. Describe your business and ask for six to eight hints describing situations your customers are in when they’d be chatting with an AI. Yes, using an AI to write targeting for an AI ad platform. That’s the new paradigm in one sentence.
What we’re hearing
We’re dabblers here too, one campaign in, so alongside our own experience it’s worth passing on what the SEO and tech community has been reporting from the first couple of months.
Targeting is coarse. There’s no suburb or radius targeting like Google has, so a Geelong plumber’s ad can show to someone in Cairns, and your ad copy has to do the qualifying. Reporting is thin: you never see which conversations triggered your ad, so if a campaign flops, diagnosing why is guesswork. Results are all over the shop, with some advertisers reporting a handful of impressions after days of waiting and others landing signed clients within a fortnight. And there are the usual beta rough edges, including occasional outages and a temporary authorisation hold on your card when you set up billing.
The counterpoint: the platform is moving quickly. Gaps advertisers were complaining about two months ago had already been filled by the time we went through setup. What’s missing today probably isn’t missing next quarter.
Should you try it?
Our take: if you’re a small business with a bit of experimental budget, this is worth a modest test. Not because we can promise results (we can’t, and neither can anyone else right now), but because the cost of finding out is low and the cost of being late to a channel is usually higher. The businesses that learned Google Ads in 2003 and Facebook ads in 2010 had years of cheap clicks before everyone else piled in. Nobody knows yet whether ChatGPT ads will follow that pattern, but the entry fee for finding out is now zero minimum spend and a bit of your time.
Treat it as research and development. Set a small budget, tag your links so your analytics can tell you what the traffic actually did, keep your expectations modest, and see what comes back. Worst case, you’ve learned how the next generation of advertising works before your competitors have heard of it.
Our own campaign has just gone live. We’ll report back with real numbers, what the clicks cost, what the traffic did, and whether any of it turned into actual enquiries, in a follow up post. No cherry picking, whatever the result.
And if you’d rather someone else set it up, that’s a thing we do. The tracking pixel, the context hints, a landing page that greets your AI-referred visitors properly, and the analytics to tell you honestly whether it’s working.
Get in touch and we’ll have a chat about whether it makes sense for your business (Note: only available to our WordPress Maintenance Customers).
