It's ALL AI Search: building a unified view for growth

by Ray Grieselhuber

Slide 1

Slide 1 text

Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 2

Slide 2 text

2,666% increase in share of traﬃc from LLMs Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 3

Slide 3 text

LLM traﬃc still less than .1% but it is growing Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 4

Slide 4 text

Diﬀerent incentives create diﬀerent narratives Speakerdeck.com/raygrieselhuber @raygrieselhuber “SEO is dead!” LLM tool providers & AEO consultants “It’s all SEO!” Rank tracking tool providers & SEO consultants

Slide 5

Slide 5 text

This debate misses the point Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 6

Slide 6 text

There is no traditional vs. AI search Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 7

Slide 7 text

It’s all AI search - the question is the user experience Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 8

Slide 8 text

3 Search Experiences Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 9

Slide 9 text

Let’s talk about search in LLMs Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 10

Slide 10 text

The context window is not inﬁnite Speakerdeck.com/raygrieselhuber @raygrieselhuber PROMPT INPUT COMPLETION PREVIOUS TOKENS NEXT TOKENS

Slide 11

Slide 11 text

“Inﬁnite” meaning complete internet up to date in real time Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 12

Slide 12 text

So, we need Live Retrieval Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 13

Slide 13 text

The app wraps the LLM to determine tools Speakerdeck.com/raygrieselhuber @raygrieselhuber App (ChatGPT, etc) LLM (GPT-5, etc) USER INPUT MODEL QUERY LIVE RETRIEVAL

Slide 14

Slide 14 text

Live Retrieval “solves” the ﬁnite context window problem Speakerdeck.com/raygrieselhuber @raygrieselhuber LLMs Search Engines Retrieval Layer High quality & more relevant answers

Slide 15

Slide 15 text

Let’s talk about the Retrieval Layer Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 16

Slide 16 text

The Retrieval Layer abstracts search engine queries Speakerdeck.com/raygrieselhuber @raygrieselhuber APP LLM QUERY RETRIEVAL LAYER External channels - web tool: - search(query) - open(url) - file connectors: - file_search etc - recording connectors Workspace channels Visuals & Files Policy & Guardrails Search operators

Slide 17

Slide 17 text

The search tool queries one or more search engines Speakerdeck.com/raygrieselhuber @raygrieselhuber External channels - web tool: - search(query) - open(url) - file connectors: - file_search etc - recording connectors Workspace channels Visuals & Files Policy & Guardrails Search operators

Slide 18

Slide 18 text

Google & Bing form the backbone of AI search Google AI Overviews Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 19

Slide 19 text

Yes, ChatGPT is using Google Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 20

Slide 20 text

Mainstream recognition occurred in August 2025

Slide 21

Slide 21 text

No, they are not “using SerpApi” instead of Google Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 22

Slide 22 text

We were seeing it in April 2025 (at BrightonSEO) ● 10,000 prompts ● 17 industries ● Google ● Bing ● ChatGPT ● Perplexity

Slide 23

Slide 23 text

Measure index overlap in citations Speakerdeck.com/raygrieselhuber @raygrieselhuber Measure: ● number of citations ● citation overlap with SERP results ● overlap percentage

Slide 24

Slide 24 text

Data in April 2025 was showing heavy Google usage Speakerdeck.com/raygrieselhuber @raygrieselhuber 51.72% Google Index ChatGPT 4o 14.26% Bing Index ChatGPT 4o 34.02% Custom / Remix ChatGPT 4o 37.12% Google Index Perplexity 45.63% Custom / Remix Perplexity 17.25% Bing Index Perplexity

Slide 25

Slide 25 text

ChatGPT is a Google wrapper (Gasp) Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 26

Slide 26 text

What has changed? Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 27

Slide 27 text

Explosion in query fan-outs Speakerdeck.com/raygrieselhuber @raygrieselhuber Commercial best winter tires Query fan-out Query fan-out Query fan-out Query fan-out Which brand is the best for winter? Why are Blizzak tires so good? Which is better for snow awd or snow tires? Do snow tires really make a difference?

Slide 28

Slide 28 text

Before: source query = retrieval layer query Speakerdeck.com/raygrieselhuber @raygrieselhuber Commercial best winter tires Query for live retrieval “best winter tires” RETRIEVAL LAYER

Slide 29

Slide 29 text

Now: source query = ~10-12 retrieval layer queries Speakerdeck.com/raygrieselhuber @raygrieselhuber Commercial best winter tires Queries for live retrieval “best winter tires” “which tire brand is best for winter” “why are blizzak tires so good” “which is better for snow awd or snow tires” “do snow tires really make a difference” etc. RETRIEVAL LAYER x 10

Slide 30

Slide 30 text

We know this because we capture the fan-outs Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 31

Slide 31 text

ChatGPT (and others) are building their citations based on SERP data Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 32

Slide 32 text

Let’s talk about Google SERPs Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 33

Slide 33 text

It’s ALL AI search Speakerdeck.com/raygrieselhuber @raygrieselhuber AI Mode 2025 https://ai.google/our-ai-journey/ AI Overviews 2024 Gemini 1.0 2023 PaLM 2022 MUM 2021 Passage Ranking 2020 T5 2019 BERT 2018 Transformer Architecture 2017 TensorFlow 2015 Google Assistant 2016

Slide 34

Slide 34 text

The existence of Zero Click SERP Features does not mean that there are zero clicks 2024 Zero-Click Search Study Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 35

Slide 35 text

Metric tree analysis Speakerdeck.com/raygrieselhuber @raygrieselhuber 70.5% Organic results 41.5% - Click 1,000,000 Searches 21.4% - New search 28.5% Google property 41.5% Click 58.5% “Zero Click” 37.1% Session ends

Slide 36

Slide 36 text

Metric tree analysis Speakerdeck.com/raygrieselhuber @raygrieselhuber 70.5% Organic results - Net CTR of 29% 41.5% - Click 1,000,000 Searches 21.4% - New search 28.5% Google property 41.5% Click 58.5% “Zero Click” 37.1% Session ends

Slide 37

Slide 37 text

Users are only on the ﬁrst page Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 38

Slide 38 text

But the index is the prize Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 39

Slide 39 text

This is why Google removed num=100 Speakerdeck.com/raygrieselhuber @raygrieselhuber Commercial best winter tires Queries for live retrieval “best winter tires” “which tire brand is best for winter” “why are blizzak tires so good” “which is better for snow awd or snow tires” “do snow tires really make a difference” etc. RETRIEVAL LAYER x 10

Slide 40

Slide 40 text

Building a uniﬁed view Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 41

Slide 41 text

The mindset Speakerdeck.com/raygrieselhuber @raygrieselhuber Understand model responses vs. live retrieval Connect relationships between available data Focus on human attention

Slide 42

Slide 42 text

The data stack Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 43

Slide 43 text

Visibility monitoring Speakerdeck.com/raygrieselhuber @raygrieselhuber Data points: ● Mentions ● Citations ● Share of Voice ● Index Overlap Combine with: ● SERP data ● Log data ● GA4 data

Slide 44

Slide 44 text

Referral data from GA4 Speakerdeck.com/raygrieselhuber @raygrieselhuber Analyze: ● Sessions ● Pages ● Conversions ● Revenue Use URL-level data to deliver additional insights

Slide 45

Slide 45 text

Log data is crucial Speakerdeck.com/raygrieselhuber @raygrieselhuber Analyze: ● Bot behavior ● Crawl issues ● Content opportunities Combine with: ● Visibility monitoring ● Prompt research

Slide 46

Slide 46 text

SERP index overlap helps you predict visibility at scale Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 47

Slide 47 text

Prompt research is a process, not a tool Speakerdeck.com/raygrieselhuber @raygrieselhuber Valuable data sources: - People Also Ask (PAAs) - Reddit - Keyword Planner - Synthetic query generation - Query fan-outs - Topic & intent mapping (content-focused)

Slide 48

Slide 48 text

Prompt volume should be seen as relational and directional Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 49

Slide 49 text

Prompt volume challenge #1: there is no “Search Console for ChatGPT” Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 50

Slide 50 text

Prompt volume challenge #2: Clickstream data is very messy Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 51

Slide 51 text

Simpliﬁed metrics Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 52

Slide 52 text

Simple data models to support relationships

Slide 53

Slide 53 text

Deceptively simple uniﬁed view of factors impacting behavior Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 54

Slide 54 text

Takeaway #1: The lines are blurry between AI search experiences Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 55

Slide 55 text

Takeaway #2: Question the origin of data sources Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 56

Slide 56 text

Takeaway #3: Think in terms of uniﬁed views Speakerdeck.com/raygrieselhuber @raygrieselhuber

Slide 57

Slide 57 text

Come visit our booth! Speakerdeck.com/raygrieselhuber @raygrieselhuber x.com/demandsphere linkedin.com/in/raygrieselhuber Slides: