We’ve been running real ChatGPT ad campaigns for a full month since they launched in Australia. The ads reached a lot of people, sent a few hundred of them to the website, and turned four of those into phone enquiries for our client.
As a relatively cheap way to buy awareness and traffic on a platform millions of people now use every day, it did a great job and is a promising new channel.
For a business already running Google and Meta that wants to be early where people now start their research, that’s worth having. If you’re looking for a new direct lead or revenue driver, it’s likely best to stick to other proven channels, at least for the moment.
Content Overview
The one-month scorecard
We’ve been running ChatGPT Ads for a commercial furniture supplier in Perth that fits out cafes, restaurants and venues with chairs, tables, booths etc. We ran the trial for that account across one month, from mid-June to mid-July 2026, with real budget behind it. Here’s what the month delivered:
- Spend: $2,250
- Impressions: around 57,000. Plenty of people saw the ad.
- Landing page views: just over 400. A small share clicked through to the site.
- Conversions: four customer enquiries.
Line those three numbers up and the channel explains itself. Tens of thousands of impressions, a few hundred visits, a handful of enquiries. ChatGPT ads buy attention and site traffic very well as it did for our client, they don’t buy cost effective leads as such.
How ChatGPT ads work
A quick bit of plumbing, because it explains the numbers. ChatGPT ads show up as “Sponsored” results underneath the answer, and only Free and Go users see them. Anyone on Plus, Pro, Business, Enterprise or Education gets an ad-free experience, per OpenAI’s own help documentation. So a slice of the highest-intent, highest-spending users never lay eyes on your ad.
Targeting is contextual, not keyword-matched. You’re bidding to appear against the rough shape of a probable prompt, not an exact search term like Google, and geography only goes down to state or country. The creative is tiny too. You get a short headline, a short description and one square image. There’s not a lot of room to qualify a buyer before the click, you’re just throwing your hat in with the answers. If you want the fuller background on the platform, we covered it in our launch explainer on ChatGPT ads coming to Australia.
What worked
It scaled without breaking
The best sign from the whole campaign was that it held up as we spent more. We started small and lifted the daily budget several times over the month. The ads kept landing amongst the right conversations the whole way, rather than drifting off-topic as the budget grew. Plenty of channels get worse when you push more money through them so this is a promising sign.
That matches what other early testers are seeing. Seer Interactive’s early observations found ad matching improving fast, with clearly off-topic ads dropping from around a fifth to a twentieth inside two weeks. It’s a new system, but it did a decent job of showing our ads next to relevant conversations. And when we lifted the budget, it kept doing that, rather than dumping the ads in front of random people just to spend the daily budget (Google is horrendous for this at times, so that’s a plus for ChatGPT).
Specific offers beat vague ones
The clearest lesson sat in the creative. The ads that named a specific product and said it was in stock did the heavy lifting. The vague, aspirational, bring-your-vision-to-life style ads pulled clicks but almost no enquiries. If you run this, lead with your sharpest, most specific offer and skip the mood pieces.
Quick to launch
One underrated plus is how easy the channel is to get into and manage. You can go from a new account to a live campaign in under an hour once you’ve planned it out. Add cheap clicks and no minimum spend, and the barrier to trying it is low. For a business already running Google and Meta, that makes ChatGPT ads an easy, cheap way to be early on a platform millions of people now use daily, without committing much to find out whether it fits.
If you’re already spending confidently on Google and Meta and you want a cheap, low-risk way to show up on a major AI platform, this is where ChatGPT ads make sense. Run tight, product-specific creative, cut what doesn’t convert, and treat the visibility as the win.
If that sounds like your setup, our ChatGPT ads management team can get you live and measured.
What didn’t work
Not every part of the campaign pulled its weight. A couple of the product areas we targeted spent steadily and returned nothing, so we paused them early and pushed that budget behind the groups that were converting. That’s the single most important habit on a channel this raw. Cut the losers fast, because a new platform hands you no shortage of expensive dead ends.
Then there’s what to expect from it. The enquiries we got were phone taps rather than booked jobs, so on a considered purchase like a commercial fit-out they’re a soft signal, not a sale. Best to judge ChatGPT on awareness and traffic, and treat any direct enquiries as a bonus on top.
The reporting doesn’t help either. You get almost no detail on who saw the ad, the tracking is fiddly, and the setup docs can point you the wrong way. Getting live takes under an hour. Proving what the campaign did afterwards takes a lot longer. Go in expecting to make calls on thin data.
So, is it worth your budget?
Prices are low right now because few advertisers are bidding, and that won’t last. Opascope’s benchmark work makes the same point. Today’s cheap rates reflect an early market, and they’ll climb as more advertisers arrive. So the question isn’t whether ChatGPT ads work in the abstract. It’s whether they fit where you are right now.
If you’re an established business already running Google Ads and paid social that are carrying your performance, and you want early visibility on a platform your customers are starting to use, add ChatGPT ads as a small awareness and consideration line. Think of it alongside programmatic display, not alongside your lead-gen Search campaigns. Being early on a channel your competitors are ignoring is worth something.
If you’re a newer or smaller business still getting your core channels right, skip it for now. Your first dollars belong where the intent already is, on Google, where people are actively hunting for what you sell. Spending a limited budget chasing awareness on an unproven channel is the wrong order of operations. Get the fundamentals earning first, then experiment.
Ads aren’t the only way to show up in ChatGPT. Being cited in the answers themselves, through generative engine optimisation, is the organic side of the same coin, and for many businesses it’s the better long-term bet. Having cut what didn’t work, our next move is to push the ads further down the funnel, adding contextual targeting aimed at people closer to buying, to see if it pulls more enquiries through.
Where this channel goes from here is still open. If OpenAI’s new conversion objective starts to perform, and the tracking and reporting get fixed, ChatGPT ads could grow into something that drives leads, not just visibility, and being early would pay off. If those pieces don’t land, and prices climb as more advertisers arrive, it stays a top-of-funnel awareness play and the maths gets tighter. We don’t know yet which way it breaks, and anyone who tells you they do is guessing.
So we’re watching four things, and they will decide it:
- The conversion objective: whether the new conversion-based optimisation shifts lead quality and cost, or just re-labels clicks.
- Tracking: whether OpenAI fixes the finicky setup and misleading docs, so the data you get back is solid.
- Reporting depth: whether you ever get audience and query-level insight, instead of just spend, impressions and clicks.
- Targeting: whether geography tightens below state level, which would make local campaigns far sharper.
Any one of those landing changes the calculation. Until they do, we run it as a measured test, keep a close eye on cost per result, and tell you honestly which way the numbers are pointing.
ChatGPT ads: quick questions
What does a month of ChatGPT ads get you?
In our trial, a month bought a lot of impressions, a few hundred landing page views, and four phone enquiries. There’s no minimum spend on the self-serve platform, so you can test with a modest budget. Just go in knowing it buys awareness and clicks, not a stack of leads.
Did ChatGPT ads generate leads?
Sort of. We recorded four enquiries, all phone taps, which sit somewhere between a click and a booked job. For a considered B2B purchase, that’s a soft signal. Treat ChatGPT ads as an awareness and traffic channel, and judge them on visibility and site visits rather than direct pipeline.
Who sees ChatGPT ads?
Only Free and Go users. Everyone on Plus, Pro, Business, Enterprise or Education gets an ad-free experience, so a chunk of the highest-spending users never see your ad. That’s worth factoring in before you set expectations on quality of reach.
Should I run ChatGPT ads for my business?
If your Google and Meta campaigns are already humming and you want to be early on a top LLM, yes, as a small awareness add-on. If you’re still building your core channels, no, put that budget into Search first. It’s an add-on for the well-established, not a starting point for newcomers.
Thinking about a test of your own? We’ll set it up, run it honestly, and tell you straight whether it’s worth continuing. Have a look at our ChatGPT ads management service, or get in touch and we’ll map it against the channels you’re already running.
Zac Fitzpatrick
PPC Service Lead
