ChatGPT Ads arrives in Italy: how advertising shifts when it enters conversations

by Alberto Narenti

As of last Monday, 24 August 2026, ChatGPT Ads is officially available in Italy. The launch of OpenAI’s ad platform unlocks a new placement within the digital media mix, allowing brands to establish a presence precisely when consumers are actively seeking solutions to a problem, comparing alternatives, or reaching a decision.

It presents a compelling opportunity that is ripe for exploration: the platform is still in its infancy, competition, as well as performance benchmarks, remains limited, and the core focus will not merely be buying clicks, but understanding the precise role ChatGPT plays within consumer decision journeys.

According to the study How People Use ChatGPT, published by OpenAI in 2025, approximately 70% of consumer conversations are centred on personal needs, such as seeking information, getting advice, planning activities, or solving everyday problems, rather than work-related tasks. While these data points do not yet demonstrate a direct impact on bottom-funnel conversions (lead generation, sales, bookings, etc.), they clearly indicate that people are already relying on ChatGPT throughout their decision-making process. Consequently, this is a touchpoint where establishing a strategic brand presence is becoming vital.

What will we cover in this article:

• ChatGPT Ads officially launches in Italy
• From search query to conversational dialogue
• Where ads appear and who sees them
• A platform already more mature than early beta tests
• What early adopters and initial test partners are saying
• Decoding user intent becomes essential
• From conversation to landing page
• Yes, but what is the actual ROI?
• Will ChatGPT Ads replace Google Ads, Meta, and established channels?
• How to approach ChatGPT Ads during this initial phase

ChatGPT Ads officially launches in Italy

Following initial testing launched in the United States in February 2026 and subsequent rollout across other markets, ChatGPT Ads has been virtually available in Italy since 24 August: the infrastructure is ready, though launching live campaigns is not yet enabled. Ads will go live shortly, over the coming weeks during Q3 2026, giving marketers clear lead time to prepare.

This is undoubtedly major news for performance marketers and media planners, and it would be overly simplistic to view it merely as yet another ad channel to stack alongside Google, Meta, TikTok, or existing placements within digital media plans.

The true point of distinction for ChatGPT Ads lies in the precise moment interaction occurs between consumer and brand. OpenAI itself frames ChatGPT as “an environment where people do not merely input a few keywords, but explain what they are trying to achieve, which criteria matter most, and what constraints they face.” This creates a fundamental point of differentiation compared to most advertising environments we have operated in over the past two decades.

From search query to conversational dialogue

Search has accustomed us to optimising for intent expressed through a search query, but even intent is evolving in a more nuanced direction.

When a user searches for “SME CRM software”, those four words represent an extremely valuable signal: they tell us there is interest, perhaps even a need, but they reveal little else. The linguistic filter of summarisation has already taken effect: the user does not explicitly detail their requirements, but compresses them into a concise phrase legible to a search engine. And as with any compression, critical information is lost regarding context, underlying motivations, and where the individual is in their buyer’s journey.

By contrast, a conversation with artificial intelligence can unfold in a far more detailed, structured manner, helping marketers capture that crucial layer of insight.

For instance, that same user might explain to the AI that they manage a 40-employee company, currently use a specific management software, are dissatisfied with its available integrations, need to connect the CRM to their marketing automation platform, and are benchmarking three prospective solutions within a set budget.

In another example closer to everyday consumer behaviour, a user might explain that they want to refresh their wardrobe with more sustainable clothing, prefer recycled or certified materials, have a fixed budget, and are looking for brands with transparent supply chains. Or they might explain that they are furnishing a small living room, have a predefined style, and need durable materials without exceeding a specific price point.

In all these cases, the interaction yields far deeper data than a generic search. We are no longer working with an isolated search query (a compressed phrase entered into a search bar), but with the user’s entire decision-making context. This is arguably one of the most compelling features of advertising within ChatGPT.

Indeed, OpenAI explains that ad selection considers the conversation topic, previous ad interactions, and, where enabled by user settings, additional signals for ad personalisation. At the same time, privacy is strictly maintained: conversations are never shared with advertisers, who receive only aggregated, anonymised performance metrics.

For marketers, this represents a fundamental shift. For years, we have worked to reconstruct user intent through incomplete proxy signals (keywords, pageviews, declared interests, audience segments, past behavior). Here, context is communicated directly by the user through natural language.

Naturally, this does not automatically guarantee higher performance. However, it provides a new, contextually rich placement where demand, information retrieval, and brand evaluation coexist simultaneously within the same interaction.

Where ads appear and who sees them

Ads are displayed as sponsored content, visually separated from ChatGPT’s organic responses, and, crucially, do not influence the output generated by the model.

This separation is fundamental to maintaining system credibility, much like Google’s long-standing policy that ad spend has zero impact on organic SEO. Advertising therefore exists alongside the conversational stream, drawing context from it while remaining distinctly labelled to prevent user confusion.

Currently, ads are served exclusively to adult users on the Free and Go plans, whereas Plus, Pro, Business, Enterprise, and Education tiers remain entirely ad-free. This provides a compelling insight into audience reach: free and entry-level subscription tiers represent the primary touchpoints for broad consumer demographics.

As always, this necessitates a thorough evaluation of strategic objectives and audience profiling. There is no universally effective media channel in isolation; rather, effectiveness depends on how closely a placement aligns with a specific target audience, business model, and stage within the decision-making process.

A platform already more mature than early beta tests

The initial phase, which we observed over recent months, was deliberately kept straightforward. In a short space of time, however, the platform has transitioned from CPM-based bidding to introducing CPC and subsequently oCPC (an optimised cost-per-click system geared toward conversions).

Furthermore, advanced targeting and reporting features have been rolled out, including geo-targeting, custom audience segments, the OpenAI Pixel, Conversions API (CAPI), and third-party attribution and measurement integrations.

ChatGPT Ads does not appear destined to remain a basic sponsored format tied solely to conversational context. Instead, it is progressively taking on the full characteristics of an enterprise advertising platform, with bidding, measurement, and optimisation framework increasingly aligned with the mechanics of major digital ad ecosystems. That said, it remains an emerging platform in its early stages.

What early adopters and initial test partners are saying

Markets where ChatGPT Ads has been active for several weeks are already providing initial feedback. While volume remains too limited to establish definitive performance benchmarks, several key dynamics are beginning to emerge.

The first involves media costs. CPC and CPM vary significantly depending on the vertical and the level of intent targeted, mirroring performance patterns observed in paid search. In highly competitive sectors, a single click carries a fundamentally different value than in broader, generalist categories; declaring a single “average ChatGPT Ads CPC” to build baseline media forecasts is therefore unhelpful at this stage.

Metrics such as CTR and conversion rates must also be interpreted with caution. The platform remains in an iterative phase, with ad formats, placements, bidding mechanics, and targeting options continually being refined. Because the ad engine is still in its infancy, any benchmark established today will likely have a short shelf life.

However, this dynamic is precisely what makes early testing and establishing a platform footprint so vital before competitors gain ground.

Decoding user intent becomes essential

ChatGPT shifts the evolution of digital advertising into a distinctly different environment: here, the buying signal originates not merely from a user’s historical online behaviour, but directly from the substance of the live interaction.

For brands, the key strategic question becomes not just whom to target, but within which specific conversations establishing a presence adds real value:

• A medical centre might aim to appear while a patient is actively benchmarking specialist expertise, pricing, and location proximity.
• A furniture brand could integrate into a dialogue surrounding kitchen design or furnishing a new property.
• A B2B software vendor might seek exposure while an enterprise evaluates tools to automate a specific business workflow.
• An automotive dealership could target a consumer determining the ideal model based on family requirements, annual mileage, and budget constraints.

This logic converges contextual and behavioural targeting in a novel way. Context is no longer defined merely by the webpage being viewed; it is defined by the specific problem the user is striving to solve.

From conversation to landing page

What happens when the user leaves the AI conversation to jump over to our website? Here, another distinction emerges compared to other advertising channels, perhaps less obvious, but operationally critical.

Imagine someone who has just spent a few minutes explaining to ChatGPT what they are looking for. They have outlined their preferences, needs, and budget, ruled out certain options, and asked for comparisons between different alternatives.

At that point, they encounter a relevant ad, click through, and arrive on a standard corporate homepage designed to speak to every audience all at once.

The risk of a disconnect is obvious.

The richer the pre-click context becomes, the more aligned the post-click experience needs to be. This could make dedicated landing pages, tailored messaging, and an experience design that seamlessly guides the user, without forcing them to start again from scratch, even more vital.

Campaign quality, therefore, won’t simply depend on purchasing media space. Creative assets, landing pages, value propositions, and performance measurement will continue to play a fundamental role in driving results.

A prime example of this evolution comes from the e-commerce landscape, specifically Shopify. Through Shopify Catalog and Agentic Storefronts, eligible merchants’ products can be discovered directly within ChatGPT whenever they prove relevant to a user’s conversation. Should the user decide to buy, the path to checkout is remarkably seamless: particularly on mobile devices, the Shopify store’s full checkout can be rendered directly inside ChatGPT’s in-app browser, maintaining the store’s branding, payment options, and customisations.

This is a significant shift because it eliminates much of the friction that traditionally exists between discovery and conversion. A user can begin with a highly detailed query, such as sourcing a product with specific features, evaluating options, and progressively narrowing down their selection, and seamlessly reach the merchant’s checkout without having to restart their buyer journey via a generic search.

Yes, but what is the actual ROI?

How do we measure the true return on our ad spend? How should we manage sales attribution for this channel? OpenAI has already introduced the OpenAI Pixel, Conversions API (CAPI), and more advanced measurement capabilities than those available during the initial project phase. This represents a crucial milestone, enabling us as advertisers to link ad traffic directly to bottom-line business outcomes. However, evaluating this channel solely through a last-click conversion lens would be a mistake, as indeed it is for any channel, irrespective of ChatGPT.

This is why I remain convinced that the future of marketing measurement lies less in searching for a single, perfect model and far more in adopting complementary frameworks: standard analytics and attribution for tactical insights; incrementality testing where feasible; and Marketing Mix Modelling (MMM) to evaluate overall channel contribution when data and volume permit.

Will ChatGPT Ads replace Google Ads, Meta, and established channels?

Most likely not, and it is arguably not even a particularly helpful question to ask.

Google remains an extraordinarily effective tool for capturing existing intent. Meta has developed an immense capacity to unlock latent demand, even when a user is not actively searching for a product. YouTube drives reach, brand recall, familiarity, and demand generation. ChatGPT, on the other hand, occupies a distinct stage: the phase where a consumer is actively trying to figure out what step to take next.

It sits in a space positioned between discovery, evaluation, and decision-making, where boundaries are naturally far more fluid than the linear models used to depict the customer journey for years – no offence to funnel purists.

This does not mean ChatGPT will automatically become a central fixture in every media plan. Rather, it means we now have access to an ad placement built around a rapidly growing consumer behaviour: leveraging artificial intelligence to gather information, compare options, evaluate trade-offs, and make informed choices.

How to approach ChatGPT Ads during this initial phase

At SAY, we believe the right approach to onboarding onto ChatGPT Ads is grounded in rigorous analysis and controlled experimentation, capturing early-mover advantages on the platform in direct alignment with our mission to guide clients with an uncompromising focus on measurable outcomes.

Being among the first to leverage a new channel offers a clear competitive edge, particularly while auction pressure remains lower than in mature ad ecosystems. However, being an early adopter does not mean allocating spend blindly.

The foundational question must focus on core use cases: within which specific user conversations can your product or service deliver genuine, context-aware relevance?

From there, we can architect an integrated execution model bridging targeting, creative assets, landing page experience, and measurement, defining upfront which KPIs to track and what strategic role to assign to the channel.

At such an early stage, we advise against immediately reallocating substantial budget away from established performance platforms. A far more prudent strategy is to establish a dedicated experimental test budget, gather baseline data, evaluate traffic quality, and progressively assess whether performance metrics justify scaling your presence.

Our objective is not simply to prove whether ChatGPT Ads works, but to discover how to deploy it most effectively to drive bottom-line business outcomes for our clients.

The key open question remains how advertising will integrate into this environment without undermining user trust or blurring the boundary between generative AI output and commercial messaging. We cannot yet predict the precise share of wallet ChatGPT Ads will capture in future media plans, how ad inventory will scale, how formats will evolve, or what density ad auctions will reach. It is too early to have all the answers, but it is unequivocally the right time to start uncovering them, combining strategic agility with a commitment to measuring the results that truly matter.

Looking to evaluate whether ChatGPT Ads has a role to play in your media mix?

SAY can help you structure initial pilot tests, integrate them seamlessly into your overall advertising strategy, and build a robust measurement framework designed to evaluate their true impact on bottom-line business outcomes.

Get in touch for a tailored consultation!

Sources

OpenAI, ChatGPT Ads expands across Europe, 18 agosto 2026
https://openai.com/it-IT/index/chatgpt-ads-expands-across-europe/

OpenAI, Testing ads in ChatGPT
https://openai.com/it-IT/index/testing-ads-in-chatgpt/

Aaron Chatterji, Thomas Cunningham, David J. Deming, Zoe Hitzig, Christopher Ong, Carl Yan Shan, Kevin Wadman, How People Use ChatGPT, NBER Working Paper 34255, settembre 2025
https://www.nber.org/papers/w34255

Shopify, Vendita su ChatGPT
https://help.shopify.com/it/manual/online-sales-channels/agentic-storefronts/chatgpt