Are Ads Coming to AI Search?

Most paid search teams are optimizing for a platform their users are quietly leaving. Not in large numbers - not yet. But the shift is measurable, directional, and accelerating in ways that matter to anyone managing a search budget right now.

The specific tension: ChatGPT is already pulling high-intent users away from Google at scale, yet almost no paid search program accounts for it. Campaigns are still built entirely around Google. Budget allocation, copy testing, conversion tracking - all of it assumes Google is where the search happens. For a growing slice of queries, it isn't.

This piece is a synthesis of what we know from current data, what OpenAI is doing with advertising, and what it means operationally for teams running paid search programs today.

AI Search Is Becoming an Ad Platform - Whether You're Ready or Not

The clearest signal comes from click-flow data. Per Ahrefs research reported by Search Engine Land, ChatGPT is the sixth most-clicked destination from Google - behind YouTube, Google's own properties, Reddit, Facebook, and Wikipedia. That ranking alone says something. ChatGPT isn't a niche research tool anymore; it's sitting inside the same tier as platforms that have had dedicated ad strategies for years.

What makes this operationally significant is the interception rate. Only 11.1% of clicks to ChatGPT are intercepted by Google's zero-click features - the lowest rate among major sites in that tier. Google's answer boxes, knowledge panels, and featured snippets tend to intercept a much higher share of clicks to most destinations. ChatGPT is getting through. Users are searching on Google and then landing on ChatGPT, which means the intent signal starts on Google but the engagement happens somewhere else.

For a performance marketer, that's a workflow problem. You bid on a keyword. A user searches that term. They click a result that takes them to ChatGPT, where they ask a follow-up question. Your ad impression happened. Your click didn't. Your conversion definitely didn't. The budget spent, the result didn't follow.

OpenAI has recognized the commercial opening this creates. Ads on ChatGPT are now expanding across 31 European countries, which signals that the ad model is past the experimental phase and moving toward standard rollout. This isn't a test anymore - it's a product.

The Paid Search Landscape Is Fragmenting, Not Collapsing

The data doesn't support panic, but it does support action. Paid clicks on Google have remained stable over the past 15 months; it's organic clicks that declined 2.8 percentage points over the same period. That distinction matters. If you run paid search, your immediate exposure is lower than it might feel reading headlines. But if you depend on organic for reach and use paid to close - which most B2B content programs do - the organic decline is already affecting your pipeline math. Understanding how B2B teams miscalculate content ROI becomes even more critical when the platforms driving that pipeline are shifting underneath you.

The more important dynamic is reallocation. Users aren't abandoning search; they're distributing it differently. A user who would have run four Google queries to research a software purchase might now run one Google query, follow a link to ChatGPT, and complete the research there. The total search intent is the same. The platform capturing it has changed.

One figure worth sitting with: per the same Ahrefs research, 86% of clicks from Google go to destinations not explicitly searched for. Google's algorithm is already routing users through a complex discovery path before they land anywhere. The addition of AI tools like ChatGPT as a common waypoint is less a rupture than an extension of something already happening. The funnel was already nonlinear. Now it has a new node.

For performance marketers, this is a new channel worth testing. For content strategists, it's a signal about where user intent is spending time. For leadership, it's a budget question that's arriving earlier than most roadmaps anticipated.

Three Patterns from Teams That Are Responding Well

We see teams handling this in three distinct ways, and the sequence matters: measure first, test second, scale third.

Audit and measure before you move budget

The first pattern is teams that mapped their high-intent keywords and tracked where users were actually landing. One approach: take your top 20 paid keywords by spend, run them manually across Google, ChatGPT, and Perplexity, and document where you appear organically and where you don't. Some teams doing this discovered that 18% of their highest-intent queries were generating significant ChatGPT activity - queries for which they had no presence on the AI platform at all.

This audit doesn't require a developer or a new tool. It requires a few hours and a structured spreadsheet. The output isn't a strategy - it's a baseline. You can't measure a shift you haven't documented. Teams that have already built robust content measurement frameworks using tools like GA4 and Google Search Console for content intelligence will find this kind of cross-platform auditing far easier to layer on top of existing workflows.

Run a contained test before restructuring the whole program

The second pattern is teams that started narrow: one product category, one AI search platform, one month. They ran ads on ChatGPT alongside their existing Google campaigns, tracked conversion rates against a matched Google cohort, and treated the whole thing as a learning exercise rather than a performance bet.

The value isn't the immediate conversion data - sample sizes are too small for statistical confidence in a one-month window. The value is operational: you learn what AI search ad copy needs to do differently, what landing pages work, and what the approval workflow looks like for a new platform. Those lessons compound. Teams that skip this step and wait for AI search to become "proven" will spend 2026 paying to catch up.

Rethink the category, not just the channel

The third pattern is teams that stopped framing their work as "Google ads" and started framing it as "paid search across platforms." This sounds like semantics, but it changes how budgets get proposed, how copy gets briefed, and how results get reported. When the category is Google, every other platform is a distraction. When the category is paid search, other platforms are part of the portfolio.

These teams adjusted their messaging for how AI search users actually phrase queries. Users on ChatGPT are more likely to ask "how do I choose between X and Y?" than "buy X." That's a different intent state, which means a different ad copy brief and a different landing page expectation. The teams that recognized this early are building copy variants and landing page tests now, while volume is low enough to learn cheaply.

What to Do with Your Paid Search Budget Right Now

The practical question most teams are sitting with: should budget move from Google to ChatGPT? The honest answer is no - not yet, and not in bulk. Google still drives the majority of paid clicks by a wide margin. Pulling significant budget before you have conversion benchmarks on the new platform is a fast way to damage a quarter's numbers.

The better framing is: allocate a test budget. Something in the range of 5-10% of a single campaign's monthly spend, running on one AI platform, with clean conversion tracking. The goal isn't performance parity with Google - it's data you don't currently have. Every month you don't have that data is a month your competitors who are testing might be building an advantage.

There's also a messaging implication that doesn't require any ad spend to address. AI search users are in a different cognitive mode than Google users. They're synthesizing, comparing, and reasoning - not scanning for the best-ranked link to click. Your ad copy and landing pages built for Google's intent patterns may underperform in that context, not because the platform is worse, but because the user is doing something different. Reviewing your top-performing Google landing pages through that lens - and stress-testing whether your content stack is quietly costing you deals by failing to meet users where their intent actually lives - is a zero-cost starting point that any team can action this week.