Lead
On May 5, 2026, Google announced five changes to AI Mode and AI Overviews that are meant to give websites more visibility in AI search. The SEO industry is reacting with partial euphoria — understandable, after AI Overviews have been pulling substantial click volume out of classical search results for over a year.
But: not every one of the five changes is strategically equal in weight. Some are substantial visibility levers, others are cosmetic adjustments that don't solve the deeper click problem. Some benefit exactly those providers Google already favors — others open visibility windows for sites that use them strategically.
This article separates the levers from the PR. It places each of the five changes against three criteria: Who actually benefits? What needs to be done to use the lever? And how deeply does the change shift the structural logic of AI search — or does it stay cosmetic?
At a glance
- On May 5, 2026, Google announced five changes to AI Mode and AI Overviews — not all are strategically equal in weight.
- Two levers are structurally substantial: New perspectives (links to unique depth content) and inline links (links directly at the relevant text fragment in the AI answer).
- Two levers are niche solutions: Subscription Linking for news publishers with subscription models, first-hand perspectives from public discussions (Reddit, Quora, forums).
- One update is cosmetic: hover preview shows site name and page title on mouse hover over a link.
- The structural zero-click mechanic remains — the updates compensate part of the click loss since 2024, but don't lead back to pre-AI-Overview click levels.
- Strategic consequence: pillar content with own substance, snippet-suitable structure, and first-hand observations becomes more important, not less.
1. What Google actually announced — the five changes at a glance
In Google's official blog post, the five changes are described in a way that makes them look like one unified visibility package. They are, in fact, structurally very different. Here first the sober inventory, without evaluation:
First, new perspectives. Behind some AI answers, a suggestion area will appear that links to unique articles and deeper analyses on various aspects of the searched topic. Google itself describes this as "suggestions for further information".
Second, improved access to news subscriptions. Content from users' personal news subscriptions will be highlighted in AI Mode and AI Overviews. The feature runs through the "Subscription Linking" program, which publishers can enroll in via a Google form.
Third, first-hand experiences from other people. AI answers will include a preview of perspectives from public online discussions, social media, and other first-hand sources. Google supplements these links with context such as creator name or community name.
Fourth, links directly at the relevant text. AI answers will contain more inline links — placed directly next to the matching text fragment. In a bullet list, the link to a thematically fitting article can now appear right next to a specific bullet, not just as a generic source list at the end.
Fifth, more context for linked websites. On desktop, a quick preview of the webpage appears on mouse hover over a link — site name, page title, possibly additional meta information.
So much for the sober finding. Now to the placement.
2. Why this announcement comes now — the economic context
A technical announcement from Google is never just technical. It's always also a response to the economic and regulatory pressure the platform is currently under. These five updates don't come out of nowhere.
Three pressure lines have converged since mid-2024:
First, the click volume decline since AI Overviews. Several independent analyses — Pew Research with around 8 percent click-through rate decrease, Ahrefs with 34.5 percent CTR drop for top rankings with AI Overviews, the NiemanLab study with declines up to 25 percent on publisher domains — have documented what publishers and content sites have been experiencing since the introduction of AI Overviews. Classical organic clicks collapse, especially on informational queries.
Second, political pressure. EU regulators and US justice authorities have been examining for months whether Google is abusing its market power with AI Overviews at the expense of content providers. The ongoing antitrust case in the US has prominently placed the topic "fair treatment of third-party content".
Third, pressure from AI competition. ChatGPT with Search, Perplexity, Claude — all three AI search platforms link substantially more prominently to source sites than Google has so far. Perplexity, for instance, shows 6-10 visibly numbered source links right under the answer on every search result. Google is, in the public perception of "who treats web sources fairly", trailing behind.
Against this background, the announcement is a dual signal. To publishers and site operators: "We know you are suffering, and we are responding." To regulators and the AI competition: "Look, we now link more prominently."
That isn't meant cynically — the updates are real, and some of them are substantial. But the economic context explains why the communication is framed so PR-affine and why the substance behind the framing should be examined soberly.
3. Lever analysis — which of the five changes are substantial?
This is where it gets concrete. We evaluate each of the five changes against three questions: Who actually benefits? How high is the effort to use the lever? And how deeply does the change shift the structural click problem?
3.1 — New perspectives (★★★ substantial)
This is the most weighty of the five levers — and at the same time the one least perceived. "New perspectives" doesn't link to the obvious top source, but to unique articles and deeper analyses. That is exactly the visibility window that has been missing for pillar content in the past months.
Who benefits: Sites with real depth instead of surface content. A research-driven industry analysis, a technically substantial explanatory piece, an economic placement with own observations — these are the contents that gain visibility in this area.
Who loses: Sites that produce top-10 lists, generic how-tos, or copy-aggregated definitions. These contents end up in the main AI answer as a summarized statement, without the source being prominently linked. "New perspectives" is not the place for "best tips for X".
What to do: Shift content strategy toward depth — fewer, but more substantial. Every article should contain a statement that is not copy-aggregated from three competitor sources. Own observations, own data, own conclusions. Word count is secondary — substance is primary.
Structural impact: High. If this function is genuinely played out consistently, it opens a visibility slot that did not exist in classical search results in this form. Pillar sites with specialized expertise are preferred.
3.2 — Subscription Linking (★ niche, but relevant for news publishers)
This function is narrowly defined: it concerns content from users' personal news subscriptions. Those who are not news publishers don't benefit here directly.
Who benefits: News publishers with subscription models — FAZ, Süddeutsche, Spiegel, Welt, Zeit Online, the New York Times, Wall Street Journal, Washington Post, the Times, Financial Times and comparable. Those who activate Subscription Linking appear preferentially in AI Mode and AI Overviews when the searching user has a subscription with that publisher.
Who loses: No one directly. It is an additive function. But: for non-publisher sites, this is simply irrelevant.
What to do: News publishers should fill out the Google form, activate Subscription Linking, and verify technical prerequisites (Reader Revenue Manager, correct paywall markup, schema.org/CreativeWork with isAccessibleForFree property). Effort: 2-5 person-days depending on complexity.
Structural impact: Low, but relevant for the target group. It doesn't change the logic of AI search, but opens an additional visibility window for a very specific provider class.
3.3 — First-hand perspectives from public discussions (★★ relevant, with restriction)
Here, Google will link to Reddit, Quora, X threads, YouTube comments, specialist forums, and similar first-hand sources. This is a response to the widespread search behavior "query + reddit" — users want real opinions, not SEO texts.
Who benefits: Clearly Reddit, then Quora, YouTube, X. Those who have substantial user content on these platforms — or who run communities — gain visibility.
Who loses: Classical SEO sites that wanted to rebuild exactly that ("we'll do forum character too"). Artificially rebuilt forum content doesn't work in this category, because Google uses the first-hand source as a criterion — that is, actual user generation in a public discussion.
What to do: Here it becomes differentiated. For brands, it pays off to be substantially present in relevant Reddit, Quora, or specialist forum communities — not spammy, but as a professional voice. In B2B mandates, LinkedIn discussions and specialist communities (e.g. Stack Overflow, GitHub Discussions) are the equivalent lever. Effort: ongoing, hard to scale, impact cumulative over 6-12 months.
Structural impact: Medium. It is a new visibility slot, but it structurally belongs to communities, not to brand sites. Those who ignore the communities don't take place in this slot.
3.4 — Links directly at the relevant text (★★★ substantial)
Inline links instead of just a generic source list at the end — this is a fundamental change to linking logic. So far, Google would give an answer in AI Overviews, then below 2-3 sources — and click probability was low. If, in future, a fitting link stands directly next to the concrete statement fragment, click probability rises noticeably.
Who benefits: Sites with fragment-specifically citable content. Concrete definitions, clear how-to steps, concise explanations — that is, contents whose individual sentences function as standalone statements. These are preferentially embedded directly in the AI answer and provided with inline link.
Who loses: Sites with nested, narrative argumentation that only makes sense as a whole. If the individual sentence doesn't carry isolated, it won't be inline-linked either.
What to do: Check content structure for snippet suitability. Clear definitions at the start of sections, concise key statements, semantic HTML structure (H2/H3 clean, p-tags without unnecessary nesting). For longer articles, it pays off to set a "key takeaways" section at the top — that one is often fragment-cited.
Structural impact: High. If inline linking is rolled out consistently, click probability from AI Overviews changes substantially. This is exactly the lever that could partially compensate for the click loss of the past 18 months — if it comes consistently.
3.5 — Hover preview of linked websites (★ cosmetic)
On desktop, a preview appears on hover over a link with site name, page title, and possibly further meta information.
Who benefits: Sites with clear, trust-building site names and clean page titles. Established brand sites gain minimal additional click probability over lesser-known sources.
Who loses: No one directly. But: sites with weak page titles or unclear brand presence see that their weakness now becomes more visible.
What to do: Check page titles and site branding. That is, however, basic SEO that should be done anyway. Those with clean title-tag setup don't need to do anything additional here.
Structural impact: Low. This is cosmetic adjustment. It makes established brands a notch more click-attractive, but doesn't change the logic of AI search.
Summary — lever evaluation
Structurally substantial levers: New perspectives and inline links. Both change how AI search uses sites in answers — and both favor depth content over surface content.
Niche levers with clear target group: Subscription Linking for news publishers, first-hand perspectives for sites with community presence.
Cosmetics: Hover preview. Nice, but strategically negligible.
4. What content strategy follows from the updates
Those who take the lever analysis seriously arrive at three strategic conclusions for content work over the next 6-12 months. These are not new — but the Google updates amplify their effect.
First, depth beats breadth. Those who produce 30 mediocre articles a year gain less in AI search than those who produce 8 substantial pillar articles. "New perspectives" and inline links favor depth. Concretely this means: review content calendar, thin out topic pipeline, increase research budget per article.
Second, writing structure must become fragment-suitable. Every H2 section should have a clear entry sentence that defines the section. Definitions should be at the top, not at the end. Examples should be marked. Bullet lists should be cleanly structured. That is the writing-craft consequence of inline linking.
Third, own statements become harder differentiation. Those who summarize three competitor sources end up in the AI answer as a summarized statement without a prominent link. Those who hold an own thesis with own data or observations have a chance at a "new perspectives" slot. That is the strategic consequence: less aggregation, more own substance.
A fourth conclusion is platform-specific: news publishers should activate Subscription Linking within the next 4-8 weeks, otherwise they lose visibility in the transition phase against competing publishers who do it earlier. This is a classical window phenomenon.
5. What the updates don't solve — the structural problems remain
As substantial as the two strongest levers are: they don't solve the structural problems that AI search has created for content sites. Those who interpret the updates as a trend reversal overlook three structural facts.
First, zero-click mechanics remain the default. For many informational queries, the answer continues to take place in the AI overview. Even with "new perspectives" or inline links: if the answer stands complete in the AI overview, a non-trivial portion of users don't click further. Inline links raise click probability, but don't eliminate the phenomenon. The Pew Research study from July 2025 shows clearly: when AI Overviews appear, users click significantly less often on any link.
Second, source bias remains. Google favors established sources in AI Overviews — Wikipedia, large publishers, authoritative domains. New sites, even if they are more substantial in content, continue to need months or years to become prominent in source selection. The five updates don't change this. Those without domain authority won't be prominently inline-linked even with inline linking.
Third, classical SERP clicks continue to break away. AI Overviews continue to stand above classical organic results. Those who rank in positions 1-3 see in many topic areas continued lower click-through rates than before AI Overviews. This is the economic reality that performance marketing and SEO strategies must reckon with.
The honest placement therefore reads: the five updates are a step in the right direction — but they are not the reversal as which they are partly being communicated. They compensate part of the click loss. They open visibility windows for depth content sites. But they don't change the basic relationship between AI answer and web source.
Those who restructure strategy on the basis of these updates should do so with realism. Depth content pays off — but not because click volume from AI search will soon reach 2023 levels again. Rather because depth content is the only path that generates visibility at all in the new logic of AI search.
6. Concrete actions for the next 30, 60, and 90 days
Without an action plan, every analysis remains theory. Here the concrete steps, sorted by effort and impact.
Within the next 30 days
Inventory audit of top 20 articles for snippet suitability. Which articles have clear H2 sections with defining entry sentences? Which have bullet lists that are citable as fragments? Which are narratively nested and need structural rework? Effort: 1-3 days depending on site size.
Title-tag and site-branding check. Every main page (homepage, service pages, important pillar articles) should have a title tag that is trust-building in the hover preview. Brand name, clear content indication, no title-tag stuffing patterns. Effort: 0.5-1 day.
For news publishers: file Subscription Linking application. Fill out form, correctly technically mark subscription status of contents, verify Reader Revenue Manager. Effort: 2-5 days.
Within the next 60 days
Restructure content calendar for the next 6 months on pillar logic. Instead of 24 articles à 800 words, rather 8 pillar articles à 3,000-5,000 words plus 8 follow-up articles à 1,500-2,500 words. Per pillar article a clear own thesis with own substance. Effort: 1-2 days strategy work, then ongoing production.
Define community presence strategy. On which Reddit subs, Quora topics, specialist forums is the brand credibly positioned? Who takes care of it? Which tone? Caution: spam attempts get caught quickly — substantial contributions take time. Effort: 0.5 day strategy plus 1-2 hours/week ongoing.
Within the next 90 days
Produce first pillar articles in the new logic. Depth instead of breadth, own theses instead of aggregation, snippet-suitable structure instead of narrative nesting. After 2-3 months, check first indicators in Search Console and GA4: do the articles appear in "new perspectives" slots? Are they inline-linked? Effort: 5-10 days per pillar article.
Adapt tracking and measurement. Classical CTR tracking from Search Console is no longer enough. AI search visibility must be additionally measured — through manual SERP observation, AI Mode samples, brand mentions in Perplexity/ChatGPT/Claude. Effort: 1-2 days setup, then monthly observation.
Calvarius perspective — how we implement this in mandates
These five Google updates are an occasion to ask an overdue strategy question anew: how much effort do we invest in click visibility that is structurally declining further — and how much in the new logic of AI search visibility, which needs different levers?
Our mandate practice shows: those sites that have structurally shifted since 2024 to pillar logic, snippet suitability, and own substance currently gain noticeable visibility advantages in AI Mode and AI Overviews. Sites that have remained in the old logic (short SEO texts, keyword optimization as main steering, aggregation instead of own thesis) lose both classical clicks and AI visibility.
The five new updates amplify this dynamic. Those who have already shifted, benefit. Those who still work in the old logic see the gap widen.
For mandates in which we take over SEO/AI-SEO steering, this means concretely:
- Content audit for pillar suitability and snippet structure
- Restructure topic pipeline toward depth instead of breadth
- Define own thesis substance per article — no aggregation content
- Implement schema.org and semantic HTML structure consistently
- Establish AI search visibility as an independent measurement dimension next to classical search
- Community presence strategy where strategically relevant
- For news publisher mandates: activate Subscription Linking
These tasks aren't worked through in 4 weeks — they are a 6-to-12-month strategic shift. But they are realizable and economically sustainable. The alternative — hoping for further declining clicks in the old logic — is not.
Sources and further reading
- Google Blog: "How AI is making Search more useful — and helping you explore the web" (May 2026)
- Pew Research: "Google users are less likely to click on links when an AI summary appears in the results" (July 2025)
- Google Developers: Subscription Linking Documentation