If there’s one theme running through the latest search and paid updates, it’s this: forecasting and performance planning are getting more complex (but not impossible).
AI systems are becoming more influential, but less predictable. Ad platforms are opening up new inventory, while simultaneously reducing the amount of data advertisers receive back. Visibility is spreading across more surfaces, but becoming harder to measure cleanly. And many of the metrics teams have relied on for years are quietly losing their usefulness.
None of this is happening overnight. There’s no single switch being flipped. Instead, it’s a steady reshaping of how discovery, trust, and performance work across search, paid media, and social.
Below are the most important changes, what they actually mean, and how teams should be thinking about them.
New research from Rand Fishkin puts hard data behind a growing concern: AI brand and product recommendations are not stable enough to be treated like rankings.
The study ran nearly 3,000 identical prompts across ChatGPT, Claude, and Google AI. Each time the same question was asked, the answers changed. Different brands appeared, lists came back in different orders, and even the number of recommendations varied wildly.
Statistically, the chance of receiving the same list twice was less than 1 in 100. The chance of seeing the same list in the same order was closer to 1 in 1,000.
This is not a bug. It’s a consequence of how generative systems work. They are probabilistic, not deterministic. They do not “rank” brands in the way search engines rank pages.
The one signal that did hold up was visibility over time. When prompts were run repeatedly, certain brands appeared across a high percentage of responses, even though their position moved constantly.
“Ranking position in AI” is not a meaningful metric
AI visibility needs to be measured probabilistically, not positionally
Tools claiming precise AI rankings should be treated cautiously
AI optimisation is about entity relevance and association, not list placement
Read more
Google has clarified how Search Console handles impressions when a URL appears both inside an AI Overview and as a traditional organic result.
The answer is simple, but important: it counts as one impression.
AI Overviews are treated as a single position in the search results. All links within that Overview share that position, and if the same URL also appears as a blue link, Search Console aggregates those appearances rather than counting them separately.
This behaviour is consistent with how Google handles other SERP features, such as knowledge panels. Repeated exposure on the same page does not generate additional impressions.
AI Overview visibility does not artificially inflate impression data
Search Console impressions represent query exposure, not placement frequency
Appearing in both AI Overviews and blue links still matters for trust and credibility
Search Console has clear limits as a tool for measuring AI-driven visibility
As search continues to shift toward AI-generated answers, YouTube is no longer a supporting channel. It is core infrastructure.
AI Overviews increasingly pull from video content, especially where explanation, demonstration, or visual proof matters. Tutorials, walkthroughs, reviews, and product demos are being cited directly, often instead of traditional webpages.
The scale reinforces the shift. YouTube is now the second most-visited site globally. TV is the primary viewing device by watch time in the US. Viewers watch over 1 billion hours of YouTube content on TVs every day. And up to 29.5% of AI Overviews cite YouTube, making it the most referenced domain.
This means AI Overviews are not just summarising articles. They are synthesising content across formats.
YouTube should be treated as a core SEO channel for relevant brands
Video content directly influences AI Overviews, not just YouTube traffic
Structured, intentional video outperforms sporadic uploads
SEO, content, and video need to operate as one system
Google has pushed back on a growing trend of restructuring content purely to appeal to AI systems.
Speaking on the Search Off the Record podcast, Danny Sullivan made Google’s position clear. Content should not be engineered specifically for LLM consumption, and publishers should not create separate versions of content for AI systems versus human users.
While chunked or highly modular content may produce short-term gains in some cases, Google’s view is that those gains are unlikely to last. As ranking systems evolve, they will continue to prioritise content written primarily for people.
This is consistent with Google’s long-standing guidance, but the timing matters. As AI search becomes more prominent, the temptation to “optimise for the model” is growing. Google is signalling that this is the wrong long-term bet.
Creating content purely for AI or LLMs is not a sustainable strategy
Short-term wins from content chunking may disappear over time
Google continues to prioritise clarity, usefulness, and human-first writing
Producing multiple versions of the same content increases long-term risk
Google has begun rolling out AI Overviews that replace the traditional local 3-pack, and early data suggests the impact on local visibility is significant.
Local SEO expert Joy Hawkins reports that some businesses are seeing drops of 50 percent or more in Google Business Profile interactions when AI Overviews appear. In these cases, the familiar map-based pack disappears entirely, replaced by an AI-generated summary.
This fundamentally changes how users interact with local results. Fewer map interactions, fewer profile clicks, and fewer direct calls. Ads, however, appear to benefit from increased prominence in these layouts.
According to Places Scout data, AI-powered local packs already appear on around 8 percent of local keywords, and adoption is increasing.
Google Business Profile interactions may decline without ranking changes
Local SEO performance will be harder to diagnose using traditional signals
Discovery is shifting from maps to AI summaries
Local strategies will need to adapt beyond pack rankings
OpenAI has begun pitching paid ad placements inside ChatGPT, positioning them as a high-attention, premium environment.
The trade-offs are clear. CPMs are high, around $60, roughly three times Meta. Reporting is limited to impressions and clicks. There is no visibility into conversions or purchases. Ads live inside conversations rather than simply pushing users out to a site.
OpenAI’s strong stance on privacy limits targeting and attribution. The bet is that context, attention, and brand safety compensate for the lack of granular data.
AI conversations are emerging as a new ad inventory type
Early tests are about learning and exposure, not efficiency
Strict CPA optimisation is unrealistic in the short term
Context and trust are the primary value drivers
Treat ChatGPT ads as upper-funnel, test-and-learn activity
Use selectively where brand safety and innovation signalling matter
Monitor how measurement evolves before scaling
Meta is experimenting with paid subscriptions across Instagram, Facebook, and WhatsApp, alongside premium AI-powered creation tools.
The core experience remains free, but advanced features, deeper insights, stealth viewing, and expanded AI capabilities may sit behind paywalls. Over time, this could reshape how users create content, manage privacy, and engage with platforms.
Different subscription tiers could also change the balance between organic and paid visibility, as well as the signals advertisers rely on for targeting and measurement.
User behaviour on Meta platforms may fragment
Premium AI tools could shift content volume and quality
Organic and paid dynamics may change by subscription tier
Measurement assumptions may need revisiting
Watch closely how subscriptions affect engagement and content signals
You need to factor potential shifts into long-term planning and avoid over-reliance on historical benchmarks
Google is testing the inclusion of third-party publisher endorsements directly within search ads.
These endorsements appear below ad copy and include a publisher name, logo or favicon, and a short quote. They are visually separated from advertiser-written text, making them feel closer to editorial validation than an ad extension.
If this expands, it could materially change how users interpret paid results, especially in competitive categories where trust plays a major role.
Search ads may start to feel more like recommendations
Third-party credibility could influence performance more than copy
Trust signals may become a competitive advantage in auctions
Auditing existing PR, reviews, and third-party coverage is key
You need to factor external validation into the search strategy
Prepare for a shift toward trust-led paid visibility
The Facebook ads performing best right now are not winning through novelty. They are winning through clarity.
Problem-solution framing, clear differentiation, customer testimonials, and simple checklists consistently outperform trend-led or overly polished creative. In a fast, exhausting scroll environment, reducing thinking effort matters more than visual flair.
Brands that make the value obvious in seconds are outperforming those trying to impress.
Reduced cognitive load improves performance
Clear structure beats inspiration
Proof and specificity outperform vague claims
Taken together, these updates paint a picture that is far less chaotic than it might first appear. While the surface-level changes can feel frequent and sometimes disorienting, the underlying direction is actually becoming clearer: discovery is increasingly shaped by AI systems, paid platforms are leaning harder into automation, and visibility is spreading across more surfaces than traditional search alone ever covered.
What this really means is not that marketers have lost control, but that control has moved. It now sits upstream, in how clearly intent is defined, how well content is structured, how credible and consistent brand signals are, and how cleanly performance data feeds into automated systems. Teams that focus on these fundamentals are finding that they are not reacting to every update, but benefiting from them.
This is where many businesses feel the gap, because making sense of how they interact across search, paid media, and social takes time, context, and experience. A change that looks like a threat in isolation can often become an advantage when viewed as part of a broader system, especially when competitors are still chasing surface-level tactics or last month’s metrics.
At Verkeer, this is the work we focus on every day. We help businesses interpret these shifts in practical terms, understand what genuinely needs attention, and translate constant platform change into clearer strategy, stronger performance, and more confident decision-making. Not by chasing every new feature, but by building foundations that hold up regardless of how search results, ad formats, or AI systems continue to evolve.