News + Thought
How does advertising work on LLMs?
AI-powered platforms are no longer just tools for research - they are rapidly becoming environments where decisions are made. With the introduction of advertising into platforms like ChatGPT, LLM advertising has moved from theory to reality.
For brands, this marks the beginning of a new kind of media channel - one that sits closer to recommendation than interruption, and one that has the potential to capture demand at the exact moment of intent.
In early February this year, OpenAI confirmed it would begin testing advertising within ChatGPT responses for users in the United States.
With over 800 million weekly users, many of whom are on free or lower-cost tiers, this represents a significant opportunity to reach large, engaged audiences within a conversational environment.
Ads appear in a clearly defined format, placed separately from the AI-generated response as “sponsored” content. What makes this format different is not just where ads appear, but when they appear.
Unlike traditional digital advertising, which often competes for attention in feeds or search results, LLM ads are embedded within high-intent moments. They sit alongside a user’s query as it is being explored, often at the point where options are being evaluated or decisions are forming. This creates a fundamentally different dynamic. Rather than interrupting a journey, these ads exist within it - positioned closer to a recommendation than a distraction.
While ChatGPT is currently leading the way, it is unlikely to remain the only platform to introduce advertising. Gemini is taking a more cautious approach, focusing on integrating ads into its AI-powered search experience through features such as AI Overviews, rather than placing ads directly within the chatbot itself.
Are paid ads coming to Gemini or Claude?
Although Google has stated there are currently no plans for ads within Gemini, industry reports suggest this may change as the product matures. I personally would be surprised if Gemini remains entirely ad-free by this time next year.
Conversely, Claude has publicly rejected advertising within its conversational interface, with Anthropic stating that conversations are not an appropriate place for ads.
However, even here, monetisation is expected to emerge through alternative models such as integrations, APIs and trusted data partnerships, where brands may influence visibility by supplying structured, reliable data rather than buying placements.
As with any emerging channel, pricing for LLM advertising is still developing. Early indications suggest that it will command a premium, particularly in its initial phase, but will also benefit from relatively low competition.
What kind of budget should you set aside for advertising on LLMs?
I should absolutely qualify this with the standard "YMMV" disclaimer before I give an answer here, but indicative ranges we're currently seeing in clients we work with are CPCs between £2 and £10, CPMs between £15 and £60, and CPAs from £30 upwards.
We are also seeing new metrics for assessing spend surface - such as cost per recommendation (CPR), which reflects how often a brand is surfaced within AI-generated responses.
While these figures will evolve over time, they point to a familiar pattern seen in other digital channels: early-stage inefficiencies can create opportunities for those willing to test and learn.
The competitive dynamic is also changing dramatically. In traditional search, multiple brands can coexist on a results page. In an AI-generated response, the number of recommendations is far more limited.
This creates a more concentrated form of competition, where being included - or excluded - has a significantly greater impact. At the same time, this presents a clear opportunity.
As with the early days of search and paid media, the landscape is still forming. Brands that engage early will benefit from lower competition, faster learning cycles and the ability to shape how they are represented within these systems.
There are, however, important considerations. Some in the industry have raised concerns around trust and authenticity, questioning whether the introduction of advertising could affect how users perceive AI-generated responses.
Similar concerns were raised in the early days of search advertising, which ultimately evolved to balance relevance with commercial content.
We also strongly feel that the effectiveness of LLM advertising will also depend on how well platforms match ads to user intent, an area where companies with mature advertising infrastructure may have an advantage - while AI is becoming increasingly influential in shaping decisions, it is not yet replacing the transaction itself.
The discovery phase may be becoming increasingly embedded into AI tech, but traditional ecommerce is still where the conversions happen: users overwhelmingly still tend to complete purchases and enquiries directly with brands, and I would be surprised to see this change in the short to medium-term.
LLM advertising is still in its early stages, but its direction is clear. AI platforms are becoming central to how people discover, evaluate and choose between options.
The way I always frame it to clients is that the opportunity is not just to appear in front of users, but to be present at the moment decisions are being shaped - and in a landscape where AI is increasingly determining what gets recommended, that distinction is what really matters.
Please get in contact with us via the form below if you'd like help with your AI search strategy.
Making you visible on AI Platforms
Need help becoming visible on AI platforms? From ChatGPT to Claude to Gemini, our experts have a wealth of experience building and optimising websites and content for AI-led discovery. Get in touch today to see how we can help your brand improve visibility.
Speak to a specialist
