Downloaded 24 times
Uploaded bySearch Engine Journal
How Do You Track What Doesn’t Rank? Measuring Visibility in AI Search.
AI search has/is/will continue to rewrite the rules of visibility. As SEOs, you're used to a world where rankings rule. But, ChatGPT, Perplexity, and Google's AI Mode don't serve ranked results, they serve answers. And if you can't measure your presence there, you're flying blind. That leaves SEO and marketing leaders with one critical question: How Do We Prove Visibility When Traditional SERP Rankings No Longer Apply? The hard truth: You cannot fundamentally replace traditional rankings with AI visibility metrics. That's a losing game. Instead, you need to reframe your mindset: think stochastic over deterministic, intent-driven over keyword-driven, and AI-as-customer over AI-as-channel. In this webinar, AJ Ghergich, Global VP of AI & Consulting Services at Botify, and Frank Vitovitch, VP Solutions Consulting at Botif, will unpack how to compete in AI search, not by chasing fading metrics, but by shedding old habits and building new measurement frameworks that are actually insightful. You’ll learn: -Why your formerly reliable SEO data doesn’t work for AI search (and what to track instead). -A new visibility framework that helps your team measure meaningful exposure. -What you should be measuring (and how to start). -Ways to use intent mapping to uncover high-value visibility opportunities. This session will help your team make confident, data-driven decisions even when the data itself is changing. Can't make it live? Register anyway and we'll send you the on-demand recording.
More Related Content
Similar to How Do You Track What Doesn’t Rank? Measuring Visibility in AI Search.
How Do You Track What Doesn’t Rank? Measuring Visibility in AI Search.
- 2. 2 Yes No 01 02 Are you now trackingAI prompts alongside organic rankings? POLL
- 3. 3 If you're stillmeasuring success the way you did two years ago, you’ve got it all wrong.
- 4. 4 AI search diverges from SEO,but they have the same foundation.
- 5. 5 Search & rankings Reframe howyou approach search FROM Conversations & visibility TO
- 6. 6 The AI searchdisruption isn’t coming — it’s already in your data.
- 7. 7 Between Sept 8–12,Google disabled num=100. Here’s what happened next.
- 8. 8 Impressions growth dropped67% Source: Botify E-Commerce & Retail 2025 Data — bit.ly/googlenum100 Impressions YoY growth, retail/ecommerce 75% 50% 25% 0% 25% Jan Mar May Jul Sep Sept 10th
- 9. 9 Clicks have stayedflat Clicks YoY growth, retail/ecommerce 20% 15% 0% 10% 20% Sept 10th Jan Mar May Jul Sep Source: Botify E-Commerce & Retail 2025 Data — bit.ly/googlenum100
- 10. 10 CTR grew 150% CTRYoY growth, retail/ecommerce 20% 0% 20% 40% Sept 10th Jan Mar May Jul Sep Source: Botify E-Commerce & Retail 2025 Data — bit.ly/googlenum100
- 11. 11 Average position improved366% Average position YoY growth, retail/ecommerce 50% 25% 0% 25% 50% Sept 10th Jan Mar May Jul Sep Source: Botify E-Commerce & Retail 2025 Data — bit.ly/googlenum100
- 12. 12 AI bot activityhas increased 75% since July Daily logs volume, AI bots 80M 60M 40M 20M 0 May Jun Jul Aug Oct Sep Source: Botify E-Commerce & Retail 2025 Data — bit.ly/googlenum100
- 13. 13 How much ofthe data you relied on was synthetic noise?
- 14. 14 How do I measure success? Whatactions should I take? 1 2 Answer two questions:
- 15. 15 AI vs TraditionalSearch
- 16. 16 AI search compressesthe customer journey: Fewer touchpoints, fewer interfaces FROM TO
- 17. 17 What’s the difference? Traditionalsearch AI search Deterministic Content is king Probabilistic Context is king User signals Session awareness Intent + query fan out VS
- 18. 18 You don’t “rank” inAI search. AI platforms are not search engines. They don’t rank content — they answer questions. Ranking has no meaning in this context.
- 19. 19 Conversations don’t happen ina vacuum. Share of voice only tracks one prompt. But consumers have entire conversations. These conversations are personalized. Your brand may only appear at step #15 — there’s no tracking for that.
- 20.
- 21. 21 Traditional search enginesrely on crawling and indexing... bit.ly/how-ai-search-works
- 22. 22 And so doAI search platforms! bit.ly/how-ai-search-works
- 23. 23 Log data isyour strongest source of truth for how you show up in AI search. focus on collecting data for LLM training. Training bots create and maintain a search index. Indexing bots RAG bots supplement pre-trained models with real-time data. bit.ly/ai-crawler-bots
- 24. 24 Benefits ● 100 %of bot hits recorded server‑side ● Shows depth & frequency of crawls Key metrics to monitor ● Bot hits by user-agent ● Bot hits by page type ● User hits referred by AI platforms Log files: Your best source of truth Source: Botify Important! The best proxy for a Chat GPT impression is a ChatGPT-User or OAI-Searchbot request
- 25. 25 GPTBot is usedto train future models (like ChatGPT-6) OAI-Searchbot is your “traditional” indexing bot. → Triggered by users asking questions that need context → Retrieves live information to use as context for answers → Most important bot for performance measurement. GPTBot OAI-Searchbot ChatGPT-User ChatGPT-User visiting your sites means ChatGPT is using your content. → Crawls websites and indexes content for OpenAI's platforms → Similar to crawlers used by traditional search engines, like Googlebot or Bingbot → Training bot, powering large language models (LLMs) like OpenAI's GPT → Crawls sites to use their content to train the next model → Only used in the future, not now. bit.ly/ai-crawler-bots
- 26. 26 Combine log data& directional data to measure AI visibility
- 27. 27 Used to maketechnical recommendations & understand bot behavior on your website. Helps you analyze: → Bot behavior → RAG impressions → Training / indexing / RAG Cross-reference log data with analytics data to understand: → How much traffic you’re getting from AI search → What people are interested in based on URL & topics → How much revenue you’re getting from AI platforms. Directional data is further away; you’ll choose one of several options as new info surfaces. Ask: → What questions do you need answered? → What data answers them? Log file analysis Analytics platforms Visibility dashboard
- 28. 28 What to measure Percentage ofanswers linking to website / top citation sources Citation share Percentage positive / neutral / negative; flag answer quality issues Sentiment mix Model coverage The models surfacing the brand across ChatGPT, Perplexity, Gemini, etc. Percentage of answers that mention brand or URLs Presence rate Visibility across awareness, research, and decision intents Journey coverage AI visibility
- 29.
- 30. 30 Normalize your data Source:Botify E-Commerce & Retail 2025 Data Impressions YoY growth, retail/ecommerce 75% 50% 25% 0% 25% Jan Mar May Jul Sep Sept 10th Set September 2025 as your new baseline 1 Estimate synthetic data percentage 2 Apply that discount to past data 3 Report both raw and adjusted numbers going forward 4 Compare YoY with adjusted data 5
- 31. 31 Build AI-ready techfoundations Determine your AI bot governance plan (robots.txt) 1 Analyze AI bot behavior and make action plan (inlinks, depth, etc.) 2 Look into JavaScript reliance, what can bots see? 3 Look into structured data, what can you improve? 4 If possible, use server-side rendering for AI bots 5
- 32. 32 Build an AIperformance framework Build an AI performance framework 1 Look at data by page type, language, etc. 2 Create AI-specific analytics dashboard (referrer + UTM) 3 Determine prompts to track (keywords, intent, PAA) 4 Create AI visibility / SoV dashboard 5
- 33.
- 34.
- 35.