Why You Should Care About These 5 Changes
Google has just made five updates to its AI Mode interface. While these might not seem like groundbreaking, flashy updates, they’re important because they reveal what Google thinks is valuable content. After all, AI Overviews plummeted position-one click-through rates by 58%. Starting in 2025, Google referral traffic fell 33% across 2,500+ news sites. When AI Overviews were featured, DMG Media’s clicks nosedived by 89%. Clearly, these changes go beyond design. But what does it all mean?
While some of these changes might seem like small shifts, digging into these will give us high signal on where the industry is headed directionally, especially given all the noise.
That there is Iyer, the founder of ACME.BOT. The next five moves he makes are more than just interface polish– they are directional bets. Knowing a bit about each could mean the difference between adapting fast & getting left behind.
Change 1: Explore new angles
When AI responses end, Google is including a “Further Exploration” block—it links out to deep dive articles and analyses on different angles of a topic that you wouldn’t see in a simple summary. Searching for urban greening? In the block, you’ll see case studies on NYC’s High Line or Seoul’s Cheonggyecheon Stream Restoration. These aren’t the simple “what is urban greening” content that you might expect. Instead, Google is surfacing case studies that it seemingly commissioned when nobody else bothered to.
Here’s what that tells us:
- Depth trumps breadth. Think: case studies, original analysis, and contrarian takes. Things like “10 ways to green your city” won’t rise to the surface.
- Subtopic specificity matters. The more narrow your subtopic, the better your odds. Which is the exact opposite of how most content teams work today. When you write highly granular, angle-specific content, your content has more surface area for this block.
- Content ideation on steroids. Remember how we used to look at the search suggestions as a tool to brainstorm blog posts? Similar to that—although now, we’re seeing examples of what Google might actually reward if people continue to dig in.
‘Seeing results from new angles is interesting — it might be a goldmine for content ideas, same way search suggestions has been.’ — Iyer, Founder of ACME.BOT
The takeaway: Publishers who hid their best content behind paywalls or over multiple pages now have serious incentives to do standalone, highly specific analyses. The bet’s on standalone reports. But whether the audience agrees is TBD.
Change 2: Get Advice from People Who’ve Been There
Here’s what we’re seeing: Google appears to be rewarding non-commodity content and punishing commodity. Commodity content: recycled how-tos, listicles, and generic content that really anyone could have written. Non-commodity content is different. Google Search’s director Danny Sullivan defines non-commodity content by identifying three qualities: they are unique (brings something others don’t), specific (talks about an actual instance, not generic rules), and authentic (shows firsthand knowledge). Content that someone who wasn’t there couldn’t have written.
“Clear visibility for actual quotes is very welcome. Doubles down on the non-commodity content.”
Google’s basically telling you: real beats generated. Act accordingly.
Change 3: Easily Access News Subscriptions
Google again just gave publishers a lifeline—but only if you’ve already built a direct relationship with readers. Now, when an AI response references a link in your active news subscription, Google calls it out with a “Subscribed” label. Sounds helpful. Problem is, in 2025, Google search referrals dropped 33%. So even a visibility boost needs teeth to be impactful.
Early testing? People significantly clicked labeled links more. But here’s where it gets real:
| Metric | Linked Subscribers | Unlinked Subscribers |
|---|---|---|
| 3-Month Page Views | +34% | +9% |
Searches on AI are now highlighting brand loyalty as a way for discovery. But, it’s not a byproduct — it’s the goal. Publishers that have been building subscription offerings can fill out Google’s subscription linking form and will start showing up on the AI platforms already being searched by users.
Putting so much emphasis on subs, web search is becoming a little more like social media.
Not everyone’s being rewarded by Google. It’s actually rewarding the publishers who its readers have already chosen to pay for. That’s a meaningful shift–and the next change shows how content structure itself decides whether you get seen at all.
Change 4: See links right where you need them
Google is now embedding links inline — right where the text matters or at the points of interest, not just at the bottom of the page. For example, in a query about a bike trip in California, links about the Pacific Coast touring guide appear right next to the analysis of the terrain, while links to training tips appear next to the note about the mileage. Context is now right where you’re reading.
Here’s what changes:
- Being specific beats being broad. Rather than lots of text, well-structured, specific content with individual sub-sections get into the SERP inline, while a wall of text won’t.
- Mapping to a topic matters. Content that answers a sub-question will show up inline where searchers are reading about that question rather than simply being buried.
- Your content structure matters for rankings. Content that has clean headers, sections that look modular, and directly answers questions about discrete sub-topics all work toward visibility.
Sometimes, thin and interchangeable content will be hidden. Alternatively, well-organized, specific answers will be prominently placed exactly where they need to live. Essentially, if your content can’t provide a relevant answer to a particular question, Google indicates that it shouldn’t have a high spot in the answer.
Really like the clear visibility into the actual quotes. Doubles down on the non-commodity content.
Iyer, Founder of ACME.BOT. As you build structures for AI search, focus on specificity and clarity—that’s where the real opportunity lives.
Change 5: Linked Website Context on Desktop
One of the most needed features in generative search is context. It took Google a while to admit that users wouldn’t click blindly on suggested links. But finally, it shipped desktop hover previews. In this feature, when you hover over an inline linked website in an AI response, you’ll see the website name or page title — allowing you to get more context on the page before actually clicking into it. For example, if you search for how to renew your passport, you might see “Renew Your Passport Online – Travel” from the U.S. Department of State.
Just one percent.
…only 1% of users clicked a cited link in an AI Overview.
Meaning, without hover clarity, links don’t just underperform, they disappear. Your page title is not just an SEO checkbox. It’s a CTR lever inside AI Mode. And a recognized domain name is a competitive advantage you can’t ignore.
This is the place where you physically wire your title clarity, structured content, and domain authority signals into your content and SEO strategy. With tools like ACME.BOT, you can start surfacing these signals where it has the most impact—ensuring your pages not only get linked, but get clicked.
“At this point, the context on websites shows up as just a quick way to see the domain a link is from. Maybe over time this will get richer with entity information from the Knowledge Graph.”
The Bigger Picture: One Consistent Signal Across All Five Changes
The painful truth: AI search cares nothing for generic content. Just look at how Google uses these five changes to address an embarrassing number. Only 1% of users click cited links in AI Overviews. That’s the problem. They’re not just arbitrary features – they’re Google’s response to it. The five changes force differentiation. Original angles, expert quotes, subscription signals, inline links, and hover previews all reward the exact same thing: content that’s unique, specific, and authentically yours. Commodity content? No angles, no quotes, no inline links. It just disappears. To get traffic and visibility, you need original perspective, recognized brand identity, and structured data working in unison.