Credit: Levart Photographer Via Unsplash
ChatGPT ads have arrived in Australia, prompting agencies to question how quickly brands should move in a channel still defining its role, measurement and economics.
The rollout went live on April 17, making Australia one of the first markets outside the US pilot.
The US test launched in February at a $US200,000 minimum spend before dropping to $US50,000 within weeks, alongside a self-serve manager introduced in mid-April.
The move is OpenAI’s early effort to turn ChatGPT from a product-led growth tool into a scalable advertising business, testing commercial demand and whether conversational AI can support intent-led media at scale.
WPP Media head of search Dan Aguirre-Davies told AdNews the speed of expansion signals a shift away from legacy platform rollouts.
“The speed of expansion is worth noting, Australian advertisers were live within eight weeks of the US pilot. It shows how quickly this space is evolving and the urgency to validate the commercial model at scale,” he said.
He said early demand spans multiple categories, with retail and B2B particularly well positioned where queries are already high intent or shoppable formats are being tested.
Holdco partnerships are key to early validation, with structured global clients best placed to test the system before wider rollout, he said.
Laura Prieto, senior digital director at UM Australia, said prioritising English-speaking markets makes sense and positions Australia as a strong early testing ground.
“Prioritising English-speaking markets makes sense, and Australia is a strong testing ground for early adopters,” she told AdNews.
Zenith Australia head of search Jack Telford said high local AI adoption and early US signals are shaping expansion into additional contained markets.
“High local AI adoption and early US pilot signals are driving expansion into relatively contained additional markets,” he told AdNews.
Deepali Kumari, senior performance manager at Enigma, said the Australian rollout reflects OpenAI's broader ambitions rather than a measured market-by-market strategy.
"From a timing perspective, it feels like OpenAI are using Australia as an early test-and-learn market before it gets scaled out globally," she told AdNews.
She said high-intent categories are best placed to move early, pointing to strong existing signals in travel and local search.
"We're already seeing strong signals in travel and local intent, where users are using ChatGPT for research and then clicking through to platforms like Google Maps or websites to complete actions," she said.
"Brands that prioritise being present within the answer journey, rather than relying purely on last-click efficiency, will likely be best positioned in this evolving landscape."
AdNews understands L’Oreal was one of the first brands to advertise on ChatGPT in Australia.
Media Words founder Elise Hedley-Dale said the US pilot has evolved quickly, with entry thresholds shifting within weeks.
“The pilot launched in the US in February at a $200k spend minimum, and OpenAI has already dropped that floor to $50k and opened a self-serve manager,” she told AdNews.
“Ten weeks in and the barrier is being lowered quickly,” she said.
Agencies broadly agree ChatGPT ads will initially concentrate in high-consideration categories including travel, retail, automotive and technology.
Yango media director Luke Creeley said performance will depend on integration into answers rather than interruption.
“The brands that will succeed are those that act as an extension of the AI answer, rather than a sales offer within the middle of the content,” he told AdNews.
Bench Media COO Shai Luft said the shift is driven by intent, not format.
“This is less about a new ad unit and more about where intent is starting to shift. People are using ChatGPT to explore options, compare things and ask for recommendations,” he told AdNews.
Search Works/BCM managing director Lukas Temple said brands must adapt to earlier-stage decision-making environments.
“Playing upstream means you need to play by upstream rules. Highly targeted paid advertising, as powerful as it can be, can never solve that wholly,” he said.
Measurement remains limited to impressions and clicks, with no conversion tracking and lower click-through rates in early US tests.
Kumari said the instinct to judge the channel by traditional performance metrics misses the point.
"While early signals may show lower CTR and inconsistent measurement, that's not unusual, display and video often follow the same pattern, with lower clicks but strong assisted conversions and meaningful influence on decision-making," she said.
Hedley-Dale said the issue runs deeper than measurement itself.
“This is a utility environment. People are there to complete a task, which makes it closer to search than to something that builds a brand,” she said.
“A visible cohort of commentators are positioning this as a brand-building channel. It can’t be both, and without agreed measurement there’s no way to settle that.”
“You’re not buying access to ChatGPT’s full user base. Paid placements are limited to free and lower-tier users, while subscribers, the most intensive users, don’t see ads,” she said.
Nick Hayes, head of digital at The Media Store, said the model remains commercially unresolved.
“For a company putting a lot of eggs in the ad revenue basket, all of this is ironic at best and completely tone deaf at worst,” he told AdNews.
He said marketers should rely on lift studies and marketing mix modelling rather than click attribution.
Telford said early metrics reflect behaviour differences, not weaker performance.
“We shouldn’t assume they will follow similar patterns to traditional search ads. The experience and intent is different,” he said.
Hedley-Dale said reporting standards remain unclear, particularly around how claims are being measured.
“OpenAI’s own line on this is telling: ‘we’re seeing no impact on consumer trust metrics, low dismissal rates of ads.’ Measured by whom, against what baseline, and being reported by the same company selling the inventory?” she told AdNews.
She said current structures limit perceived value.
“That’s not premium inventory. That’s a test-and-learn budget dressed up as a media buy,” she said.
The rollout began with global holding groups including WPP, Omnicom and dentsu, giving early access to test scale and measurement frameworks.
Hayes said the approach mirrors earlier platform launches.
“The approach is similar to Netflix’s ad rollout, holdcos first, high entry costs, limited measurement and a platform evolving its commercial model in real time,” he said.
“It’s a controlled environment to prove demand before broader access.”
Luft said access is already broadening beyond initial partners.
“The initial rollout through global holding groups is pretty typical. It is important to be clear that access is not exclusive to them. This is already accessible more broadly across the market,” he said.
Creeley said US CPMs have shifted sharply, dropping from US60 to US25 in under ten weeks.
Hedley-Dale said lower entry costs may redistribute rather than reduce risk.
“Reportedly around 80% of advertiser interest is coming from small and midsized businesses, which are also the least equipped to evaluate what they’re buying,” she said.
“A lower barrier to entry isn’t democratisation if the people entering the market can’t properly pressure test the product.”
She said operational support remains unresolved.
“Ads without someone to pick up the phone is half a product,” she said.
Shah Ghaffurian, CEO of Magic, said the rollout will expand as capability matures.
“The shift is as much about capability as buying,” he told AdNews.
“What’s more interesting is the skillset demand it accelerates, understanding high-intent query environments and conversion architecture, combining search expertise with broader audience planning capability,” he said.
Prieto argued that broader access only heightens the importance of thoughtful adoption. As AI tools become more widely available, the real challenge shifts from access to decision-making.
“When anyone can buy it, the discipline of deciding whether to buy it matters more, not less,” she explained.
Telford agreed on the need for caution, warning that some brands may rush in without a clearly defined approach. He noted that momentum alone isn’t a strategy.
“It’s easy to get swept up in the fear-of-missing-out before nailing down the strategic play.
Looking at the bigger picture, Creeley positioned ChatGPT as part of a wider ecosystem rather than a standalone solution. He emphasised integration over replacement.
“An individual ChatGPT strategy will fit alongside traditional paid search, SEO as well as social search, each playing a critical role in brand discovery,” he said.
Hayes said the format is not structurally new.
“It is about as status quo as you get in digital media,” he said.
Ghaffurian compared the shift to early-stage platform adoption cycles.
“It’s similar to when TikTok emerged as a new channel. Clients still rely on agencies to recommend where it fits, structure testing and integrate it into the broader plan,” he said.
Prieto said AI accelerates execution but not judgement.
“The talent in the team can unpack client-specific nuances to tailor those outputs and drive business outcomes for our clients,” she said.
Hedley-Dale said advisory capability will separate agencies.
“The agencies that struggle will be the ones whose value was built on volume and process. The ones that hold up will be ones where a senior strategist is still in the room, telling clients the hard things,” she said.
Agencies agree the bigger shift is not paid inventory but organic visibility in AI-generated answers.
Prieto said brands are already behind.
“Many clients are thinking about AI readiness, but it’s already time to be AI relevant,” she said.
Creeley said organic presence is the deeper structural change.
“The organic side is a far greater shift in our industry and offers the most opportunity for brands. Paid ChatGPT ads are just the tip of the iceberg,” he said.
Luft said credibility now determines visibility.
“It is less about keywords and more about how your brand shows up across the web, how credible your content is, and whether these models see you as a useful source,” he said.
Hayes said the implications extend beyond advertising.
“These products are having a huge impact on automation, personalisation, privacy and analytics. This is a great opportunity for agencies to move beyond their classical role and evolve from solving for advertising, to solving for marketing,” he said.
Ghaffurian said the market is still forming a clear position.
“Clients are starting to ask. Most agencies don’t have a clean answer yet. That’s the space to watch,” he said.
Have something to say on this? Share your views in the comments section below. Or if you have a news story or tip-off, drop us a line at adnews@yaffa.com.au
Sign up to the AdNews newsletter, like us on Facebook or follow us on Twitter for breaking stories and campaigns throughout the day.