# Are AI Bots Changing Google Ads Performance? What Marketers Need to Know in 2026 > Are AI agents clicking Google Ads, affecting CTR or triggering remarketing? Explore 2026 bot traffic research and what PPC marketers should monitor. Author: Mariliis Uibomäe, Brand Promotion Published: 2026-10-21 Last updated: 2026-10-21 Web version: https://brandpromotion.ee/en/marketing-blog/ai-bots-google-ads-2026 ## In brief AI-agent web traffic is growing, but this does not prove an impact on Google Ads performance. Review invalid clicks, CTR, network segments and conversion quality before drawing conclusions. ## Are AI bots quietly changing your Google Ads performance? Imagine asking an AI assistant to find the best laptop, compare insurance providers or research business loans. Instead of opening ten browser tabs yourself, the assistant does the research for you. It visits websites, compares information and might even help complete a purchase. Sounds convenient. But here's a question for marketers: **What happens when an AI agent encounters a Google Ad?** Does it click? Ignore it? Trigger tracking pixels? Accept cookies? Could its activity influence impressions, CTR or remarketing audiences? We're entering a period where not every website visitor is necessarily human. And not every automated visitor is necessarily bad. Let's look at what the research actually tells us. ## AI agents entered the chat. Big time - **+41.5%** AI agent traffic in Q3 2026. - **4.7×** growth in agentic browser requests in just one month. - **1,700%+** year over year growth in daily AI agent requests, according to Cloudflare. But here's the million dollar PPC question: are these agents also interacting with our ads? These are automated web requests, not confirmed Google Ads impressions or clicks. **Sources:** [DataDome, October 2026](https://datadome.co/threat-research/ai-traffic-report-q3-2026/); [Cloudflare, September 2026](https://blog.cloudflare.com/agentic-web/) [[TOC]] ## How much is AI bot traffic growing in 2026? Automated traffic has existed for decades, but AI agents are changing how it behaves. According to DataDome's AI Traffic Report Q3 2026: - Nearly **25 billion AI-agent requests** were recorded in Q3 2026, up 41.5% from Q2. - Agentic browser requests increased **4.7 times** from August to September 2026. - ChatGPT's cloud browser generated 27.4 million requests in September, compared with 6.1 million in August. - Claude's browser extension traffic almost tripled during the same period. These figures represent traffic observed across DataDome's network, not the entire internet. An earlier DataDome report found that AI-related requests increased 82.3% between July 2025 and June 2026, while their share of measured web requests rose from 1.2% to 1.7%. That's a useful distinction: AI traffic is growing quickly, but it still represents a relatively small share of all measured website requests. **Source:** [DataDome, AI Traffic Report Q3 2026](https://datadome.co/threat-research/ai-traffic-report-q3-2026/) [[RESEARCH:traffic]] [[RESEARCH:browsers]] ### The longer trend: 12 months of AI traffic DataDome's 2026 State of Bot & Agent Security Report shows the bigger picture: - From July 2025 to June 2026, DataDome detected **52.7 billion** AI agent and LLM crawler requests. - AI traffic grew **82.3%**, and its share of all requests rose from **1.2%** to **1.7%**. - June 2026 was the record month with **6.6 billion** requests. The sharpest growth came from March to June. - The largest AI traffic sources were Meta (**46.3%**) and OpenAI (**34.6%**). - **ChatGPT.com** drove **83.4%** of all AI referred traffic to websites. - AI agents targeting login pages increased **735.8%** in the first half of 2026. - At the current pace, AI traffic could exceed **79.2 billion** requests a year by 2027. [[RESEARCH:monthly]] **Source:** [DataDome AI Bot & Agent Traffic Trends 2026](https://datadome.co/resources/bot-and-agent-security-report/ai-traffic-trends/) ## Wait, aren't AI agents just bots? Technically, yes. But not all bots do the same things. **Search engine crawlers** visit websites to discover and index content. **Monitoring bots** check prices, stock availability, uptime and other information. **Malicious bots** may generate spam, scrape sensitive information or create fraudulent traffic. **AI crawlers** collect or retrieve content for AI-powered systems. **AI agents** can perform multi-step tasks, including researching products, navigating websites, filling in forms and sometimes helping users complete transactions. The distinction matters for advertisers. A malicious bot generating fake clicks is very different from an AI shopping assistant researching products for a genuine customer. Both may create automated website activity, but their intentions and business value are different. [[RESEARCH:bots]] ## Can AI agents actually click Google Ads? The short answer: technically, yes. But it depends on how the agent accesses the internet. An agent using a browser may encounter advertisements displayed on websites or search results. Depending on its capabilities and instructions, it may also interact with clickable elements. An agent retrieving information through an API may never receive advertisements at all. And an agent visiting a product page directly doesn't necessarily need to interact with a Google Ad. There is another complication. Google Ads filters invalid traffic, including certain automated clicks. However, standard Google Ads reports do not identify a separate category for purchases or clicks made by AI agents acting on behalf of real users. That leaves advertisers with an interesting measurement problem. **An AI agent could theoretically contribute to a legitimate customer journey without being clearly identifiable as an AI agent in the advertising reports.** There is currently insufficient public evidence to conclude that AI agents are causing widespread Google Ads CTR declines or increases in invalid clicks. ## Could AI agents influence remarketing audiences? Here's where things become particularly interesting. Imagine asking an AI assistant to research running shoes. The assistant visits several ecommerce websites, opens product pages and compares prices. One of those websites uses Google Ads remarketing. Could the agent's visit trigger a remarketing tag? Technically, a page visit can trigger tracking logic if the necessary conditions are met, including the relevant consent and browser environment. But that doesn't automatically mean you'll start seeing remarketing ads yourself. Several factors matter: - Whether the agent uses your own browser or an isolated environment. - Whether cookies and identifiers are shared. - Whether marketing consent exists. - Whether the advertising platform can associate the visit with an eligible user or audience. An agent operating in a separate cloud browser may create a completely different tracking context from the human who requested the research. And importantly, **entering a remarketing audience does not prove that an advertisement was clicked**. A direct website visit may also qualify. ## Can an AI agent accept cookie consent on your behalf? Consider a website that blocks further browsing until a cookie banner has been addressed. What should an AI agent do? Reject optional cookies? Accept everything? Ask the user? A browser-capable agent may be able to interact with consent buttons, but the resulting choice depends on the agent's instructions, capabilities and the design of the consent interface. It also raises privacy questions. If an agent clicks "Accept all", does that reflect the user's actual preferences? And if consent is recorded in the agent's browser, does it have any connection to the user's personal advertising profile? These questions deserve attention from marketers, developers and privacy teams. Research into cookie banner design already shows that the way consent choices are presented can significantly affect human decisions. A 2026 experiment involving 1,142 German internet users found acceptance rates ranging from approximately 67% to 95% depending on the interface design. That study measured human participants, not AI agents. It should not be treated as evidence of how frequently AI agents accept cookies. **Source:** [Privacy Concerns and Choices under Design Friction, 2026](https://www.sciencedirect.com/science/article/pii/S0167624526000259) [[RESEARCH:consent]] ## Are advertisers already noticing changes in CTR? PPC communities regularly discuss unusual increases in impressions, falling CTR, invalid clicks and unexpected traffic patterns. But a change in CTR alone tells us very little about the cause. For example, CTR may decline because of: - Broader keyword matching or changes in search terms. - Increased competition. - Different ad placements or search partner traffic. - Changes in campaign targeting. - Shifts in user behaviour. - Automated or invalid traffic. A sudden increase in impressions without a corresponding increase in clicks may be worth investigating. However, it is not proof that AI agents caused the change. The same applies to anecdotal reports from advertising communities. They can suggest useful questions, but they cannot establish causation without further investigation. ## What should PPC marketers check in Google Ads? Rather than assuming that AI agents are responsible for unusual performance, start with the data you already have. ### 1. Invalid clicks and invalid click rate Review the trend over several years, ideally using comparable campaigns and periods. An increase may justify further investigation, but Google does not label every invalid click as AI-agent activity. ### 2. Impressions and CTR Look for unusual changes in stable campaigns. Compare impressions, clicks and CTR together rather than treating CTR as an isolated metric. ### 3. Google Search vs Search Partners Segment your performance by network where available. Traffic patterns can differ considerably between Google Search and search partners. ### 4. Conversion quality Look beyond conversion counts. Are leads qualified? Are purchases genuine? Has the conversion rate changed without an obvious campaign explanation? ### 5. GA4 engagement and referral traffic Review engagement trends, landing pages and identifiable AI-platform referrals. Keep in mind that AI-agent activity may not appear as a normal human session, and referral reports will not capture every automated interaction. ### 6. Change history Before blaming bots, check what changed in the account. Keyword matching, bidding, budgets, ad assets and landing pages can all influence performance. **The most useful evidence would be a consistent pattern across multiple comparable campaigns or accounts, supported by traffic-quality data.** ## What if some AI-agent traffic is actually valuable? Not every automated visit is wasted traffic. Imagine an AI agent researching business software for a company. It compares pricing, reads documentation and evaluates providers. The human decision-maker may never personally visit every website that influenced the final recommendation. That creates a new challenge for marketers. How do we measure the value of a website interaction when part of the customer journey happens through an AI assistant? DataDome's Q3 2026 research found that some agentic browser traffic reached forms, account creation, shopping carts and payment-related endpoints. That does not prove completed purchases, but it suggests that agents are moving beyond simply reading content. For businesses, this makes clear product information, accessible website content and reliable conversion tracking increasingly useful. ## What happens next for Google Ads? We don't yet have enough evidence to say that AI agents are materially changing Google Ads CTR or conversion rates across the industry. But we do know that automated browsing is growing and that AI agents can perform more sophisticated website interactions than traditional crawlers. For PPC specialists, there are two questions worth following: **Could AI agents introduce new forms of advertising measurement noise?** And perhaps more interestingly: **Could they become a valuable part of the customer journey?** Both possibilities deserve investigation. ## Frequently asked questions ### Can AI agents click Google Ads? Browser-based agents may technically interact with ads, depending on their capabilities and instructions. There is no reliable public industry-wide figure showing how frequently this happens. ### Does Google Ads charge for bot clicks? Google filters clicks it identifies as invalid and applies its invalid-traffic policies. Not every automated interaction is necessarily classified in the same way. ### Can AI agents trigger remarketing pixels? Potentially, if they load a page where the tag fires and the necessary tracking conditions are met. That does not automatically associate the visit with the human user's advertising profile. ### Can AI agents accept cookies? Some browser-based agents can interact with cookie banners. Whether the choice is appropriate, legally valid or associated with a human user's browser depends on the circumstances. ### How can I identify AI-agent traffic in Google Ads? There is no dedicated AI-agent segment in standard Google Ads reporting. Advertisers can investigate invalid clicks, CTR, network segments and conversion quality, but these metrics alone cannot confirm AI-agent activity. ## Share your experience **Have you noticed unusual changes in your Google Ads performance?** I'm particularly interested in experiences involving invalid clicks, impression spikes, CTR changes and conversion quality. If you've noticed anything interesting, I'd love to compare notes. **Need help reviewing your Google Ads performance or conversion tracking?** Explore [Google Ads services](/en/marketing-services/google-ads), [analytics and tracking](/en/marketing-services/analytics-and-tracking) and [AI in marketing](/en/marketing-services/ai-in-marketing) to discuss your campaigns. [[CTA:buttons]] ### Research and references 1. [DataDome: AI Traffic Report Q3 2026](https://datadome.co/threat-research/ai-traffic-report-q3-2026/) 2. [DataDome: AI Traffic Trends 2025–2026](https://datadome.co/resources/bot-and-agent-security-report/ai-traffic-trends/) 3. [Privacy Concerns and Choices under Design Friction, 2026](https://www.sciencedirect.com/science/article/pii/S0167624526000259) 4. [Cloudflare: The Internet has a second audience, 2026](https://blog.cloudflare.com/agentic-web/) 5. [Google Ads Help: Invalid Traffic](https://support.google.com/google-ads/answer/42995)