AI Search KPI’s: Measuring Success in the New SEO Era - Eduard Blacquière

Traditional SEO metrics no longer tell the whole story in an era of AI-powered search features. In this presentation, I’ll share what you actually can, and should, be measuring today. I’ll connect AI search mechanics to new metrics and KPIs beyond classic clicks and rankings. It's a practical, hands-on talk packed with examples, calculations, and actionable frameworks you can apply to your business today.

This presentation is a lightning talk given by SEO & AI Search consultant at the Search Central Live Deep Dive 2026 in Barcelona.

Transcript

1. Search Central Live Deep Dive 2026 Lightning Talk AI Search KPI’s: Measuring Success in the New SEO Era Eduard Blacquière SEO & AI Search Consultant | Edwords Consultancy
2. I’ll give you my KPI’s
3. Traditional search User Search results
4. Traditional: you do the searching the Web
5. AI Search adds a new layer User AI (LLM) Search results
6. AI is like a sales assistant the AI (LLM)
7. AI does searching for you
8. Traditional measurement: “Do we rank?” User (query) “Do we rank?” Search results
9. How most people measure AI Search… 󰣹 User “Do we rank?” ❌ AI (LLM) Search results
10. AI visibility = “Do we rank?” 󰣹
11. How you SHOULD measure AI Search User "Do AI systems understand, trust, and choose you?" AI (LLM) Search results
12. Does AI understand, trust & choose you? the AI (LLM)
13. How to measure? 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
14. Most AI measurement is only selection 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
15. But why are you shown? And why NOT? 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
16. AI Search success starts much earlier 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
17. 2 major signals 1. Primary bias 01 Know 02 Associate 2. Grounding presence 03 Search 04 Retrieve 05 Selection
18. Measure every stage 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
19. Know = does AI know you exist? 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
20. Know = crawling or fidelity problem Crawling / fidelity problem 01 Know ❌ 02 Associate 03 Search 04 Retrieve 05 Selection
21. Know metrics AI bot allow rate Target: 100% AI parsability rate Target: 100% Fidelity score 01 Know Target: >90% 02 Associate 03 Search 04 Retrieve 05 Selection
22. AI Bot Allow Rate GPTBot ClaudeBot Google-Extended CCBot ✅ ✅ ✅ ❌ AI Bot Allow Rate = 75% Target: 100%
23. Don’t block AI training bots Examples: GPTBot ClaudeBot Google-Extended CCBot
24. Associate = does AI connect you to topics? 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
25. Association: does AI already know what you are good at? the AI (LLM)
26. Association: what is its bias?
27. Association = brand problem Brand problem 01 Know 02 Associate ❌ 03 Search 04 Retrieve 05 Selection
28. Associate metrics Brand x association rate Target: 100% Consensus share 01 Know 02 Associate Target: 100% 03 Search 04 Retrieve 05 Selection
29. Measure Brand x Association Rate (BAR) Prompt: Name 10 topics associated with [BRAND] Credits: DEJAN
30. Measure with web browsing OFF
31. The other way around Prompt: Name 10 brands associated with [TOPIC] Credits: DEJAN
32. Measure Brand x Association over time 1 2 3 4 5 6 BAR Topic 1 X% Topic 2 Y% Topic 3 Z%
33. Association = 100% … awesome! Topic 1 1 2 3 4 5 6 BAR 🟢 🟢 🟢 🟢 🟢 🟢 100% Topic 2 Y% Topic 3 Z%
34. Association = 50% … OK’ish 1 2 3 4 5 6 Topic 1 Topic 2 Topic 3 BAR 100% 🟢 🔴 🟢 🔴 🟢 🔴 50% Z%
35. Association = 17% … Ouch 🙈 1 2 3 4 5 6 BAR Topic 1 100% Topic 2 50% Topic 3 🔴 🔴 🔴 🔴 🟢 🔴 0%
36. Search = is AI searching for you? 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
37. Search = fan-out queries Fan-out queries
38. Search = relevancy problem Relevancy problem 01 Know 02 Associate 03 Search ❌ 04 Retrieve 05 Selection
39. Search metrics (Fan-out) Query x Brand Co-occurrence Rate (QCR) (Fan-out) Query Match Rate (QMR) 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
40. Query x Brand Co-occurrence Rate (QCR) User query Fan-out queries: 1. query variant 2. query variant 3. query variant
41. Query x Brand Co-occurrence Rate (QCR) User query QCR = 33% Fan-out queries: 1. related term competitor1 competitor2 2. related term site:competitor1 your_brand 3. related term question 2026
42. Tools to measure fan-out queries Search engines Browser extensions Tool providers
43. Retrieve = AI live requesting your pages 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
44. Retrieve = crawl, index or claim problem Crawl/Index or Claim problem 01 Know 02 Associate 03 Search 04 Retrieve ❌ 05 Selection
45. Retrieve metrics Grounding Trust Score (GTS) AI parsability rate TTFB Indexation rate 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
46. Selection = AI choosing you 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
47. Selection = trust problem Trust / relevancy problem 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
48. Selection metrics Selection rate Constraint Fit Index (CFI) 01 Know 02 Associate 03 Search 04 Retrieve 05 Selection
49. We can measure AI Search (better!) AI bot allow rate AI parsability rate Fidelity score 01 Know Brand x association rate Consensus share 02 Associate Query match rate Query x Brand Co-occurrence Rate (QCR) 03 Search Grounding Trust Score (GTS) Selection rate AI parsability rate Constraint Fit Index (CFI) Indexation rate Mentions share 04 05 Retrieve Selection
50. There is even more to measure Know Associate Claim Evidence Search Index Retrieve Mention Citation
51. Main point: measure earlier in AI “funnel” the AI (LLM)
52. Main point: measure + improve if AI understands, trusts & chooses you! the AI (LLM)
53. Steal my KPI’s + Start measuring (better)!
54. Lightning Talk Happy to get in touch! Thank you 🙏 eduard@edwords.nl https://www.linkedin.com/in/eduardblacquiere/