<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet href="/feed.rss.xml" type="text/xsl" media="screen"?>
<rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:media="http://search.yahoo.com/mrss/" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Sunil Pratap Singh</title>
    <description>Sunil Pratap Singh is an independent SEO and GEO consultant based in India, with 15+ years in search. He helps businesses rank on Google and get named by ChatGPT, Perplexity, Gemini and Google AI Overviews when buyers ask for a recommendation, using his Search Signal Framework: Presence, Trust, Relevance and Momentum.</description>
    <link>https://speakerdeck.com/sunilpratapsinghseo</link>
    <atom:link rel="self" type="application/rss+xml" href="https://speakerdeck.com/sunilpratapsinghseo.rss"/>
    <lastBuildDate>2026-10-27 16:45:56 -0400</lastBuildDate>
    <item>
      <title>The Search Signal Framework: Diagnosing and Enhancing AI and Google Visibility</title>
      <description>Most search work starts with a tactic. Almost none of it starts with a diagnosis.

The Search Signal Framework organises Google and AI visibility into four cascading signals: Presence, Trust, Relevance and Momentum. Each one gates the next, so your lowest signal is your ceiling, and work spent above it returns nothing until the constraint moves. That single rule is what turns "should we do SEO or AI search" into a question with one answer.

Inside: the diagnostic tells for each signal, a 228-signal scoring model, four evidence slides with sourced data, and original research showing which platforms the major AI crawlers are actually allowed to read.

The framework maps 200 Google ranking factors across the four stages and adds 28 GEO-specific signals that traditional SEO audits do not measure, for a total of 228 signals.

WHY DIAGNOSIS BEATS TACTICS RIGHT NOW

68.01% of US Google searches ended without a click in the first four months of 2026, up from 60.45% two years earlier (SparkToro on Similarweb clickstream data). When an AI summary is present, users click a traditional result on 8% of visits against 15% when it is absent, and click a link inside the summary on just 1% (Pew Research Center, 68,879 searches, March 2025). The margin for guessing which tactic to run has narrowed.

THE FOUR SIGNALS

Presence (Signal 1). Can Google and AI systems find, read and understand your brand? Two tracks: crawl health, indexation, internal links, sitemap coverage and technical foundation on Google; entity recognition, Person and Organization schema, and a consistent footprint across platforms on AI systems. Roughly 52 Google factors plus 8 GEO signals.

Evidence: 52% of AI crawler requests are now made for training, up from 22% in spring 2025. A further 36% of activity comes from mixed-use crawlers. Pure search crawling, the only kind that can return a citation with a link, is a small and declining share of the total. Cloudflare also report that the content categories crawled most heavily saw human referral traffic fall by up to 40% in under a year. Crawl volume is not a Presence score, and being read and being visited have come apart (Cloudflare, network data through June 2026, published 1 July 2026).

Trust (Signal 2). Do Google and AI systems treat your brand as a credible, authoritative source? Trust is built from what other sources say about you, not what you say about yourself. Two tracks: backlinks from topically relevant domains, authority trajectory, unlinked mentions and topical authority; then citation frequency in AI responses, presence in industry listicles and roundups, and author entity signals. Roughly 78 Google factors plus 8 GEO signals.

Evidence: across 75,000 brands, branded web mentions correlate with AI Overview visibility at 0.664, branded anchors at 0.527 and branded search volume at 0.392, while number of backlinks sits at 0.218. The top three signals are all off-site and none of them is a link. Ahrefs state explicitly that correlation is not causation and that all measured factors showed only moderate to weak relationships (Ahrefs, May 2025).

Relevance (Signal 3). Are you visible for what buyers actually search, on Google and on AI systems? Relevance is intent alignment, not search volume. Roughly 51 Google factors plus 6 GEO signals.

Evidence: the average ChatGPT prompt runs 23 words against 3.4 words for an average Google query (Semrush, December 2025). A four-word query and a twenty-three-word prompt are different retrieval problems even when the buyer is identical, which is why "we already rank for that keyword" is not evidence of AI visibility.

Momentum (Signal 4). Is visibility compounding, or plateauing? Without deliberate momentum architecture, sites reach a ceiling around month twelve to eighteen and stop growing. Roughly 22 Google factors plus 6 GEO signals.

Evidence: AI assistants cite pages averaging 1,064 days old against 1,432 days for organic results, a 25.7% freshness advantage across 16.975 million citations on 7 platforms (Ahrefs, July 2025). And each engine concentrates on a different kind of source: Wikipedia is 47.9% of ChatGPT's top ten cited sources, while Reddit is 46.7% of Perplexity's and 21.0% of Google AI Overviews' (Profound, 680 million citations, August 2024 to June 2025). Winning one engine is not evidence you will win the next.

THE CASCADE PRINCIPLE

Presence gates Trust. Trust gates Relevance. Relevance gates Momentum. A score of 22 out of 25 on Trust alongside 8 out of 25 on Presence does not average out. The Presence score is the ceiling. Fix your lowest signal first, because work spent above the constraint returns nothing until the constraint moves.

ORIGINAL RESEARCH IN THIS DECK: AI CRAWLER ACCESS BY PLATFORM

A profile only counts as an entity signal if the engine you are trying to reach is allowed to read it. Live robots.txt rules, parsed per crawler against a representative content path on 23 August 2026:

- LinkedIn: open to OAI-SearchBot only. GPTBot, PerplexityBot, ClaudeBot and Google-Extended are all disallowed.
- Reddit: a blanket disallow. Every listed AI crawler is blocked.
- Quora: open to Google-Extended only. The rest are disallowed.
- Medium: open to OAI-SearchBot, PerplexityBot and Google-Extended. GPTBot and ClaudeBot are disallowed.
- Crunchbase: open to OAI-SearchBot and PerplexityBot. GPTBot, ClaudeBot and Google-Extended are disallowed.
- SlideShare and Scribd: open to GPTBot, OAI-SearchBot, PerplexityBot and Google-Extended. ClaudeBot and CCBot are disallowed.
- GitHub, Substack and Wikipedia: open to all five.

Three of the nine platforms are readable by all five major AI crawlers. Two are closed to four of the five or to every one of them. Reproducible by anyone in under a minute, and these rules change without notice, so re-check before acting.

Read those rows as separate doors, not as one verdict. GPTBot collects data for model training. OAI-SearchBot builds the index ChatGPT answers from. ChatGPT-User is the live fetch when someone pastes a URL. LinkedIn closes the training door and holds the search door open, which is why ChatGPT still cites LinkedIn articles while Claude, Perplexity and Gemini cannot reach them at all. Blocking a training crawler and blocking a search crawler have completely different consequences, and most robots.txt commentary treats them as the same decision.

The conclusion: a large share of the standard "build your entity everywhere" advice points at platforms that are closed to the engines being targeted. Audit access before you audit content.

SCORING

Each signal is scored out of 25, summing to 100. The total is the least useful part of the output. The distribution is what tells you where to work.

- 0 to 25, Presence stage. Fix technical SEO and entity formation.
- 26 to 50, Trust stage. Fix backlinks and topical authority.
- 51 to 75, Relevance stage. Fix intent audit and AI prompt mapping.
- 76 to 100, Momentum stage. Fix content clusters and citation growth.

HOW TO USE IT

1. Score all four signals before changing anything.
2. Identify the lowest signal, not the most interesting one.
3. Fix only that signal until it stops being the constraint.
4. Re-score and expect the constraint to move. When Presence lifts, Trust usually becomes the new ceiling.
5. Run both tracks inside every signal. Every signal has a Google track and an AI track, which is why treating search and AI visibility as competing budgets produces the wrong decision in both directions.

SOURCES

Zero-click search 68.01%: SparkToro on Similarweb clickstream, Jan to Apr 2026, via searchengineland.com/google-zero-click-searches-2026-study-479717
AI summary click rate 8% against 15%: Pew Research Center, 22 July 2025, pewresearch.org
Crawler purpose split, 52% training: Cloudflare, spring 2025 to June 2026, published 1 July 2026, blog.cloudflare.com/agentic-internet-bot-report
Brand mentions against backlinks, 0.664 to 0.218: Ahrefs, 26 May 2025, ahrefs.com/blog/ai-overview-brand-correlation
Prompt length 23 against 3.4 words: Semrush, 11 December 2025, semrush.com/blog/catch-up-on-ai-search
Citation freshness 1,064 against 1,432 days: Ahrefs, 28 July 2025, ahrefs.com/blog/do-ai-assistants-prefer-to-cite-fresh-content
Top cited sources per platform: Profound, tryprofound.com/blog/ai-platform-citation-patterns
Google ranking factor set: Backlinko, 2026
AI crawler access table: original research for this deck

Correlation is not causation. The Ahrefs authors state this explicitly and it applies to every correlational figure above.

Full framework, scoring detail and the free diagnostic:
https://sunilpratapsingh.com/framework?utm_source=speakerdeck&amp;utm_medium=deck&amp;utm_campaign=search-signal-framework</description>
      <media:content url="https://files.speakerdeck.com/presentations/419907be53014ddba8edec6a9bcc6dc8/preview_slide_0.jpg?41114227" type="image/jpeg" medium="image"/>
      <content:encoded>Most search work starts with a tactic. Almost none of it starts with a diagnosis.

The Search Signal Framework organises Google and AI visibility into four cascading signals: Presence, Trust, Relevance and Momentum. Each one gates the next, so your lowest signal is your ceiling, and work spent above it returns nothing until the constraint moves. That single rule is what turns "should we do SEO or AI search" into a question with one answer.

Inside: the diagnostic tells for each signal, a 228-signal scoring model, four evidence slides with sourced data, and original research showing which platforms the major AI crawlers are actually allowed to read.

The framework maps 200 Google ranking factors across the four stages and adds 28 GEO-specific signals that traditional SEO audits do not measure, for a total of 228 signals.

WHY DIAGNOSIS BEATS TACTICS RIGHT NOW

68.01% of US Google searches ended without a click in the first four months of 2026, up from 60.45% two years earlier (SparkToro on Similarweb clickstream data). When an AI summary is present, users click a traditional result on 8% of visits against 15% when it is absent, and click a link inside the summary on just 1% (Pew Research Center, 68,879 searches, March 2025). The margin for guessing which tactic to run has narrowed.

THE FOUR SIGNALS

Presence (Signal 1). Can Google and AI systems find, read and understand your brand? Two tracks: crawl health, indexation, internal links, sitemap coverage and technical foundation on Google; entity recognition, Person and Organization schema, and a consistent footprint across platforms on AI systems. Roughly 52 Google factors plus 8 GEO signals.

Evidence: 52% of AI crawler requests are now made for training, up from 22% in spring 2025. A further 36% of activity comes from mixed-use crawlers. Pure search crawling, the only kind that can return a citation with a link, is a small and declining share of the total. Cloudflare also report that the content categories crawled most heavily saw human referral traffic fall by up to 40% in under a year. Crawl volume is not a Presence score, and being read and being visited have come apart (Cloudflare, network data through June 2026, published 1 July 2026).

Trust (Signal 2). Do Google and AI systems treat your brand as a credible, authoritative source? Trust is built from what other sources say about you, not what you say about yourself. Two tracks: backlinks from topically relevant domains, authority trajectory, unlinked mentions and topical authority; then citation frequency in AI responses, presence in industry listicles and roundups, and author entity signals. Roughly 78 Google factors plus 8 GEO signals.

Evidence: across 75,000 brands, branded web mentions correlate with AI Overview visibility at 0.664, branded anchors at 0.527 and branded search volume at 0.392, while number of backlinks sits at 0.218. The top three signals are all off-site and none of them is a link. Ahrefs state explicitly that correlation is not causation and that all measured factors showed only moderate to weak relationships (Ahrefs, May 2025).

Relevance (Signal 3). Are you visible for what buyers actually search, on Google and on AI systems? Relevance is intent alignment, not search volume. Roughly 51 Google factors plus 6 GEO signals.

Evidence: the average ChatGPT prompt runs 23 words against 3.4 words for an average Google query (Semrush, December 2025). A four-word query and a twenty-three-word prompt are different retrieval problems even when the buyer is identical, which is why "we already rank for that keyword" is not evidence of AI visibility.

Momentum (Signal 4). Is visibility compounding, or plateauing? Without deliberate momentum architecture, sites reach a ceiling around month twelve to eighteen and stop growing. Roughly 22 Google factors plus 6 GEO signals.

Evidence: AI assistants cite pages averaging 1,064 days old against 1,432 days for organic results, a 25.7% freshness advantage across 16.975 million citations on 7 platforms (Ahrefs, July 2025). And each engine concentrates on a different kind of source: Wikipedia is 47.9% of ChatGPT's top ten cited sources, while Reddit is 46.7% of Perplexity's and 21.0% of Google AI Overviews' (Profound, 680 million citations, August 2024 to June 2025). Winning one engine is not evidence you will win the next.

THE CASCADE PRINCIPLE

Presence gates Trust. Trust gates Relevance. Relevance gates Momentum. A score of 22 out of 25 on Trust alongside 8 out of 25 on Presence does not average out. The Presence score is the ceiling. Fix your lowest signal first, because work spent above the constraint returns nothing until the constraint moves.

ORIGINAL RESEARCH IN THIS DECK: AI CRAWLER ACCESS BY PLATFORM

A profile only counts as an entity signal if the engine you are trying to reach is allowed to read it. Live robots.txt rules, parsed per crawler against a representative content path on 23 August 2026:

- LinkedIn: open to OAI-SearchBot only. GPTBot, PerplexityBot, ClaudeBot and Google-Extended are all disallowed.
- Reddit: a blanket disallow. Every listed AI crawler is blocked.
- Quora: open to Google-Extended only. The rest are disallowed.
- Medium: open to OAI-SearchBot, PerplexityBot and Google-Extended. GPTBot and ClaudeBot are disallowed.
- Crunchbase: open to OAI-SearchBot and PerplexityBot. GPTBot, ClaudeBot and Google-Extended are disallowed.
- SlideShare and Scribd: open to GPTBot, OAI-SearchBot, PerplexityBot and Google-Extended. ClaudeBot and CCBot are disallowed.
- GitHub, Substack and Wikipedia: open to all five.

Three of the nine platforms are readable by all five major AI crawlers. Two are closed to four of the five or to every one of them. Reproducible by anyone in under a minute, and these rules change without notice, so re-check before acting.

Read those rows as separate doors, not as one verdict. GPTBot collects data for model training. OAI-SearchBot builds the index ChatGPT answers from. ChatGPT-User is the live fetch when someone pastes a URL. LinkedIn closes the training door and holds the search door open, which is why ChatGPT still cites LinkedIn articles while Claude, Perplexity and Gemini cannot reach them at all. Blocking a training crawler and blocking a search crawler have completely different consequences, and most robots.txt commentary treats them as the same decision.

The conclusion: a large share of the standard "build your entity everywhere" advice points at platforms that are closed to the engines being targeted. Audit access before you audit content.

SCORING

Each signal is scored out of 25, summing to 100. The total is the least useful part of the output. The distribution is what tells you where to work.

- 0 to 25, Presence stage. Fix technical SEO and entity formation.
- 26 to 50, Trust stage. Fix backlinks and topical authority.
- 51 to 75, Relevance stage. Fix intent audit and AI prompt mapping.
- 76 to 100, Momentum stage. Fix content clusters and citation growth.

HOW TO USE IT

1. Score all four signals before changing anything.
2. Identify the lowest signal, not the most interesting one.
3. Fix only that signal until it stops being the constraint.
4. Re-score and expect the constraint to move. When Presence lifts, Trust usually becomes the new ceiling.
5. Run both tracks inside every signal. Every signal has a Google track and an AI track, which is why treating search and AI visibility as competing budgets produces the wrong decision in both directions.

SOURCES

Zero-click search 68.01%: SparkToro on Similarweb clickstream, Jan to Apr 2026, via searchengineland.com/google-zero-click-searches-2026-study-479717
AI summary click rate 8% against 15%: Pew Research Center, 22 July 2025, pewresearch.org
Crawler purpose split, 52% training: Cloudflare, spring 2025 to June 2026, published 1 July 2026, blog.cloudflare.com/agentic-internet-bot-report
Brand mentions against backlinks, 0.664 to 0.218: Ahrefs, 26 May 2025, ahrefs.com/blog/ai-overview-brand-correlation
Prompt length 23 against 3.4 words: Semrush, 11 December 2025, semrush.com/blog/catch-up-on-ai-search
Citation freshness 1,064 against 1,432 days: Ahrefs, 28 July 2025, ahrefs.com/blog/do-ai-assistants-prefer-to-cite-fresh-content
Top cited sources per platform: Profound, tryprofound.com/blog/ai-platform-citation-patterns
Google ranking factor set: Backlinko, 2026
AI crawler access table: original research for this deck

Correlation is not causation. The Ahrefs authors state this explicitly and it applies to every correlational figure above.

Full framework, scoring detail and the free diagnostic:
https://sunilpratapsingh.com/framework?utm_source=speakerdeck&amp;utm_medium=deck&amp;utm_campaign=search-signal-framework</content:encoded>
      <pubDate>Tue, 27 Oct 2026 00:00:00 -0400</pubDate>
      <link>https://speakerdeck.com/sunilpratapsinghseo/the-search-signal-framework-diagnosing-and-enhancing-ai-and-google-visibility</link>
      <guid>https://speakerdeck.com/sunilpratapsinghseo/the-search-signal-framework-diagnosing-and-enhancing-ai-and-google-visibility</guid>
    </item>
  </channel>
</rss>
