Something has changed in how people use Google. Not an algorithm update. Not a new feature appearing above the results. The search behavior itself has shifted.
Queries have gotten longer. More detailed. More exploratory. People no longer type "Nike running shoes" to buy. They type "what running shoes are best for asphalt if I have sensitive knees." And Google AI Mode answers with a full paragraph, not ten blue links.
The problem? Most Google Ads campaigns are built for the first type of query. Not the second. And the gap between them isn't just about format. It's about intent, funnel stage, and ultimately ROI.
When queries move, everything moves
Google has publicly confirmed what many of us have observed for months: AI Mode is changing the nature of search. Shivani Mohan, Google VP, said it directly: people are asking "much longer, much more detailed questions." According to Google's AI Mode Report from May 2026 (U.S. data), AI Mode queries are three times longer than traditional searches.
This isn't a cosmetic change. It's an intent shift. And it has implications for every dollar you spend on paid search.
When someone types a long, nuanced question, they're not in buying mode. They're researching. They're comparing options, weighing trade-offs, forming an opinion before they ever click an ad. They're top-of-funnel, not bottom-of-funnel. And the ads that show up in this context have a completely different job than what they were optimized to do.
Most PPC campaigns, everywhere in the world, are optimized for direct conversions: CPA, ROAS, cost per lead. Bottom-of-funnel metrics applied to a context that is gradually but unmistakably becoming about research and exploration. That's where the disconnect starts, and few advertisers have noticed it yet.
The numbers few people read correctly
Philipp Schindler, Alphabet's Chief Business Officer, explicitly linked slower revenue growth to AI Mode. He cited "the lower monetization we were seeing in AI Mode" as a direct factor. The Q3 2026 numbers confirm the trend: Search revenue grew 12% year-over-year, down from 19% in Q1 and 17% in Q2.
Paid clicks grew just 8% in Q3, compared to 13% in each of the two prior quarters. Cost-per-click still rose 4%. In business terms: you're paying more per click, getting fewer clicks, and those clicks come from people who may not be ready to buy.
This isn't a crisis. But it's a signal you can't ignore if you're running Google Ads campaigns with meaningful budgets.
And the detail that matters most: Google hasn't published AI Mode-specific ad performance data. We don't know what the actual CTR, conversion rate, or CPA looks like in AI Mode. Nobody does. Not even Google seems entirely sure. Which means every advertiser is navigating blind through an environment that's changing underneath them.
Why your PPC metrics are misleading (and it's not their fault)
If the only way you evaluate campaign performance is through blended CPA and overall ROAS, you have a problem. These metrics mix performance from purchase-intent queries with performance from research queries. And if the share of the latter is growing (it is), your CPA will look worse without you doing anything wrong.
It's a more subtle version of the problem we discussed in the reach vs. attention debate: old metrics applied to new contexts produce false conclusions. A higher CPA from AI Mode doesn't necessarily mean a weak campaign. It means a different intent that requires a different approach.
The issue is that most agencies and marketing departments continue to report on the same dashboards, with the same KPIs, as they did in 2024. Meanwhile, user behavior has moved. Google has confirmed the move. Only the campaigns stayed put.
It's like driving with last year's GPS. The roads have changed, but you're following the same directions. You'll arrive somewhere, sure. Probably not where you need to be.
Three budget adjustments you can make today
You don't need to pause campaigns or panic. But you need to adapt concretely, not just "monitor the situation."
Separate your AI Mode budget. Treat reach from AI Mode and AI Overviews as a separate experiment with its own success metrics. Don't blend these results into your overall CPA. Allocate 10-15% of your search budget as a dedicated test and measure it independently. This will show you clearly what works in the new context and what doesn't.
Adapt messaging for research intent. If queries are longer and more exploratory, your ad copy needs to answer questions, not sell directly. "Discover your options" works better than "Buy now at 20% off" when someone is still researching. Landing pages need to match: educational, with useful content, not product pages with a single buy button.
Measure mid-funnel, not just final conversions. Add engagement metrics: time on site, pages visited, micro-conversions (newsletter signup, guide download, chat initiated). If your only KPI is direct sale, you'll conclude that AI Mode doesn't work. But maybe it works exactly as it should, just earlier in the funnel than you expected. That's not failure. That's a different kind of success that your current reporting doesn't capture.
And don't ignore what's happening in organic at the same time. Declining organic traffic and paid search moving up-funnel are symptoms of the same shift. AI is rewriting how people interact with search. Paid and organic are both affected, just from different angles. If you're only watching one side, you're missing half the picture.
Google will need to resolve the tension between the AI Mode experience (great for users) and its monetization (bad for revenue). We'll likely see new ad formats, new placements, and new billing models designed for research queries. But that could take 6-12 months.
In the meantime, companies that understand the funnel has shifted will have a real advantage. It's not about spending more. It's about spending differently, in a context that no longer looks like it did two years ago. And those that keep optimizing for a funnel that no longer exists? They'll pay more and more for weaker and weaker results, without understanding why.
