Can Google really just cut off almost all of the internet from AI? Yes, yes they most certainly can.
September 2025 tossed another grenade into SEO and AI search. Google quietly killed the &num=100 parameter: the hack rank trackers used to yank 100 results at once. For years, scrapers rode that shortcut to mine the long tail on the cheap. Not anymore.
The result: dashboards, reports, and AI retrieval pipelines are suddenly scrambling. What appeared as radical drops in impressions or rank volatility usually stem not from a core algorithm update, but from a measurement disruption.
5 Key Takeaways:
- Google killed num=100; deep SERP firehose gone; scraping and RAG costs spike, access shrinks.
- Expect impression waterfalls and “better” average positions, traffic steady. Measurement shock, not demand collapse.
- Page one is the new floor; long-tail vanity ranks fade from visibility and influence.
- Win AI surfaces: structure entities, FAQs, TL; DRs; become citable across engines, not just Google.
- Rewire ops: paginate sampling, add GSC and logs, track Bing/Perplexity, prioritize clicks, conversions, brand signals.
This shift may feel like chaos, but it’s also a filter. The sites and strategies that survive will be those built for real visibility, not vanity metrics. Learn how this SEO and AI search shift reshapes rank tracking, AI retrieval, small sites, and what you must change now in your SEO and AI search optimization strategy.
The Most Recent SEO and AI Search Shift: What Changed (and Why It Matters)
1. The disabling of &num=100
Previously, appending &num=100 to a Google search URL let you return up to 100 organic results in one page (instead of Google’s default 10). Many tools use it to fetch deep SERP data in a single request.
But around September 10–12, 2025, Google deprecated that behavior. Now, URLs with &num=100 are often treated the same as &num=10, or ignored entirely.
The impact: to get 100 results, tools must paginate (i.e., make 10 requests of 10 each). That’s 10× more calls, rate-limit risk, and higher infrastructure/engineering cost.
2. Strange metrics: a drop in impressions, a rise in average position
Because many SEO and AI search tools and bots previously contributed to Google Search Console (GSC) “impressions” (by requesting deep SERPs), their disappearance triggers sharp metric shifts:
- In one analysis of 319 sites, 87.7 % saw impressions drop after num=100 was removed.
- In GSC, 77.6 % of sites lost unique ranking queries (i.e., fewer keywords surfaced).
- Because deep impressions vanish, average position often improves (mathematically) even though user-visible ranking hasn’t changed.
In effect, many SEOs are seeing waterfall drops in impressions and keyword counts, but click counts (actual user traffic) are largely stable. That implies the change is in reporting, not in user demand.
3. Why Google likely did this
Google has not publicly given a detailed explanation. But the compelling rationales are:
- Bot/scraping control: &num=100 made it efficient for tools and AI systems to parse deep pages. Disabling it raises cost and rate-limit friction and reduces abuse.
- Strategic control over AI search access: As Google integrates AI summaries, it may wish to control which pages external systems can access cheaply vs. via licensable APIs.
- Server load and efficiency: Serving deeper pages on demand is expensive; defaulting to more limited pages reduces overhead.
- Align GSC with human behavior: Real users rarely scroll past page 2. By removing machine-driven deep impressions, reported metrics better reflect real visibility.
This change is less about punishing SEOs than rebalancing those who control deep web data.
The Winners & Losers of SEO and AI Search (After the Cut)
SEO strategies built on scraping deep pages or mass-long-tail visibility are now under pressure. The new playing field favors sites that rank in the top 10, succeed in AI surfaces, or generate direct signals (brand, links, citations).
| Stakeholder/Asset | Effect | Why |
| SEO tools/tracker vendors | Negative | They must make many more requests, handle higher error rates, and raise costs or reduce tracking depth. |
| AI retrieval/RAG systems | Negative/constrained | They lose cheap access to deep-ranked URLs, compressing their input space and reducing diversity. |
| Small/mid-tier sites (rank > 10) | Harm | These sites disappear from scraped or aggregated datasets; they lose “visibility” in third-party tools. |
| Benefit/control | Gains power over what external systems can easily see. Enhances alignment between its SERP/AI surfaces and what external systems can access. |
What SEO and AI Search Optimization Marketers Must Do: 7 Tactical Moves
Below are seven concrete changes to your SEO and AI search playbook, tactical, not theoretical.
1. Treat Top-10 as your new visibility floor
Anything sliding beyond position 10 is now invisible to most tools and AI retrieval systems. Focus resources on pushing content into page one rather than chasing page-two long tail.
2. Expand multi-search-engine tracking
Diversify beyond Google-only metrics. Track rankings and visibility in Bing, Brave Search, Perplexity, ChatGPT citations, etc. These can reveal opportunities beyond Google’s strict SERP paradigm.
3. Use direct crawl & log data
Don’t depend solely on Google scraping. Use your crawlers, site logs, internal search queries, UX analytics, and first-party data to triangulate performance, instead of relying on third-party tools for deep insights.
4. Amplify entity, schema, and structured data
AI and generative systems prioritize structured information. Use a robust schema (FAQ, How-to, Product, Organization) and explicitly define entities and relationships in content so you become a clean extraction target.
5. Write to be cited in AI search summaries/answers
Now, writing for keywords alone is insufficient. Format digital content for AI extraction: TL;DR summaries, lists, comparison tables, bolded answers, and clear facts. Think about how a model would read and cite you.
6. Rebalance metric focus (clicks, conversions over impressions)
With impression and average position data disrupted, lean on real metrics: clicks (GSC), organic sessions (GA4), conversion paths, assisted conversions, retention, and brand searches. Impressions are now more noise than signal.
7. Build first-party visibility & authority
Rely less on scraped impressions. Invest in brand, email lists, social media, syndication, thought leadership, data-led content, and link-driven authority. The more your site is referenced, quoted, and linked, the less tool visibility matters.
How the Shift Affects Traditional SEO, AI Search & Generative Retrieval
To understand this change’s broader implications, consider the architecture of “retrieval-augmented generation” (RAG) systems that power ChatGPT, Bing, Perplexity, etc. Those systems rely heavily on external corpora of web content to answer queries.
Previously, those systems could cheaply scrape up to 100 Google results as context. Now, their cheap depth is restricted. They’ll rely more heavily on:
- Page-one sources
- Structured databases (Wikipedia, knowledge graphs)
- Licensed feeds and APIs
- AI summaries and citations
- Sites with strong entity representation
In other words, your position in AI search is now more tightly coupled to whether your content is extractable, credible, and highly cited, not just whether it ranks in position 50 or 70.
A recent academic paper by Cornell University argues that content influence (how likely your content is to be used in synthesized answers) matters more than rank position in the generative search era. In other words, SEO is evolving further: into AI-driven SEO (sometimes called AISO or Generative SEO). Under that paradigm, ranking is a means to an end (inclusion in answers), not the end itself.
Implementation Roadmap for SEO and AI Search Optimization (Next 30-90 Days)
| Phase | Focus | Actions |
| Weeks 1–2 | Audit & triage | Identify content that’s near the top 10. Flag pages whose metrics dropped hard in GSC. Annotate dashboards to show “pre-/post- num=100 baseline.” |
| Weeks 3–6 | Top-10 & entity push | Re-optimize content to cross the page-one threshold. Add entity markup, schema, FAQs, and structured format. Test snippet-rich versions. |
| Weeks 7–12 | Expand reach & retrieval | Begin multi-engine tracking. Monitor AI citation presence (manually search in ChatGPT, Perplexity). Push content to aggregators, syndicators. Collect internal logs, site search, and referral data. |
| Quarterly | Review & reset | Rebase line KPIs (clicks, conversions). Drop vanity keywords from tracking. Reallocate budget from deep-LT scraping toward content and authority plays. |
Key Takeaways on the Newest SEO and AI Search Shift
- The removal of &num=100 is not just a technical footnote. It is a structural pivot in how SEO and AI systems access the web. Many of the dashboards, bots, and tools we came to trust are now outdated.
- You’ll waste resources if you cling to tracking mass vanity rankings beyond position 10. The future rewards: high-ranking pages (top 1–10), content built for AI extraction, authority, citations, structured entities, first-party signals, and real traffic.
- This isn’t “the end of SEO.” It’s the recalibration of SEO for a more constrained, more extraction-oriented web. The winners will be those who see this disruption not as a setback, but as a filter.
Count on Digital Marketing Experts for Successful SEO and AI Search Optimization
Only SEO and AI search marketers can help you navigate the long-tail cutoff frenzy. Partner with a digital marketing agency to significantly increase your chances of ranking in AI search queries and improve the quality of your traditional SEO strategy.
Get in touch today to elevate your brand awareness, online visibility, and profits.