News and Insights
What Google’s AI shift actually means for paid search
June 15, 2026
Google I/O and Google Marketing Live both ran in the same week this year, and together the two conferences made it pretty clear where search is headed. AI Mode is now where most of the activity is, and the way we buy is changing in response.
Search is moving from “tell Google what to bid on” to “tell Google what a good customer looks like.” Google Ads is not going away, but the accounts that perform well over the next 12 months will be the ones that have spent time getting their first-party data and conversion tracking in order.
This month’s Pulse is going all in on Google to look at the inner workings of AI Mode and how it’s rewriting the rules of paid search.
Trend 1: AI Mode is now the search experience
Google AI Mode passed 1 billion monthly users this year, with query volume more than doubling every quarter since launch. Gemini 3.5 Flash is the default model globally. The search box itself has been redesigned for the first time in 25 years and now accepts text, images, files, videos and Chrome tabs as input.
Three new ad formats are being tested in the U.S.: Direct Offers, Conversational Discovery and Highlighted Answers. They sit inside the AI-generated answer rather than next to it, which is a meaningful change for how ads actually appear to users.
Google is also rolling out information and booking agents that handle more of the journey before someone reaches your website. This summer the U.S. is getting 24/7 monitoring agents for Ultra subscribers, and agentic booking is expanding into local services like home repair, beauty and pet care.
💓 The Pulse
- Expect less website traffic: As more of the journey happens inside Google, retargeting audiences will get smaller. Softer website conversions such as newsletter sign-ups or content downloads may be affected.
- Your landing pages are part of the ad: Final URL expansion lets Google rewrite the path through your site, so the quality of your pages directly affects ad performance.
- Ads inside answers need different creative thinking: They appear inside the generated response, not alongside it, which changes the standards for an effective ad.
OUR TAKE
It is tempting to treat AI Mode as another channel to add into the plan, but it is closer to a new default. The work to perform well in AI Mode starts before the campaign goes live. First, look at the conversion model, the landing pages and the audience signals. Get those right and the campaign work becomes much easier.
Trend 2: Paid search is being rebuilt around signals
AI Max for Search came out of beta in April. The most useful addition is AI Brief, a natural language layer that lets you instruct the AI in plain English. You can tell it to focus on a particular product range, to avoid certain query types or to never mention pricing. It is the first natural language layer in the targeting and creative stack that is actually useful.
The thread running through AI Brief, AI Max, Performance Max, Broad Match and AI Mode ads is the same. They all optimize against the conversion data you feed them. If your conversion tagging, enhanced conversions, offline imports or value rules are weak, the AI will optimize toward the cheapest, easiest conversions rather than the most valuable ones.
It’s also worth being honest about why Google is pushing these products so hard right now. The less you control manually, the more freedom Google has to optimize on its terms.
💓 The Pulse
- Audit your signal layer before anything else: Conversion tagging, enhanced conversions, offline imports, customer lists and value rules. Until these are clean, nothing else will work properly.
- Test AI Brief like you would brief an agency creative: Try different prompts to see what the AI picks up and what it ignores.
- Hold off on a full move to AI Max or PMax: Keep manual campaigns running until your conversion tracking has been live and validated for 30 days, and where compliance, brand control or low conversion volume mean the AI is not yet ready to take over.
OUR TAKE
There is no one-size-fits-all answer here, and the right approach varies by account. Manual bid strategies and tightly controlled keyword sets still outperform automation in some categories, particularly for high-intent terms or compliance-sensitive briefs.
The most sensible position is a portfolio approach. AI-led where the signal layer supports it, manual where it does not, and a structured test plan to keep checking which is which as accounts evolve.
____________________________________________
About the author
Lee Faulkner leads the Performance side of FINN Paid Media, our global team of experts who live and breathe today’s fast-paced — and fast-changing — media landscape.
The team delivers a full suite of services, including omnichannel planning and buying, performance media strategy and management, and comprehensive measurement, resulting in award-winning campaigns that drive client success.
POSTED BY: Lee Faulkner
