Presentation · 13 slides

The click is no longer the conversion.

Discovery has fragmented across search, social, AI answers, communities, and email. The new job is to build influence where buyers already look.

Download the PDF Get the deck + Founder's Playbook

AISandwich

AI STRATEGY & EXECUTIVE COACHING

The click is no longer the conversion.

Discovery has fragmented across search, social, AI answers, communities, and email. The new job is to build influence where buyers already look.

Charles Gourlaouen

The old promise had a receipt

For years, digital marketing's superpower was supposed to be perfect measurability.

The squishy stuff—like impressions and share of voice—was left for traditional channels.

It was a tidy, predictable system that was perfectly legible to your CFO.

Channel

Search Engines

Social Networks

News Media

Forums & Blogs

Ad Networks

Direct Traffic

Tactic

Optimize signals to get rankings and clicks

Post links to draw traffic

Buy traffic

Brand?

Measurement

Quantity of traffic + attribution to signups, sales, and revenue.

The "squishy" metrics like impressions and share of voice.

AI-SANDWICH.COM02

The jaws keep opening

In Pew Research Center's observational study, 18% of Google searches produced an AI summary. When a summary appeared, a traditional link was clicked in 8% of visits — but a link inside the summary was clicked in only 1% of visits.

Source: Pew Research Center, "Do people click on links in Google AI summaries?" (July 2025; March 2025 browsing data). Note: Observational association, not proof of sole causation.

U.S. Google searches: ending with vs. without a click2019: SparkToro/Jumpshot · 2024: SparkToro–Datos · 2026: Similarweb panel

The core problem isn't that nobody sees your marketing. It's that more people see it, but fewer of them click.

Search didn't die. It spread.

Search still matters. Its answer layer is expanding.

Fresh 2026 evidence shows the behavior shifting inside the results page: more searches end in Google, while AI Mode and AI Overviews extend Google's answer surface.

And the answer layer barely refers: AI Mode sends a visitor to a website on roughly 2% of queries, versus 17–19% for classic search (Similarweb, Jan–Apr 2026).

The strategic question is not "Google or AI?" It is: where does your buyer get the answer—and are you present there?

68.01%

of U.S. Google searches ended without a click in the first four months of 2026.

>20%

of Google searches showed an AI Overview in the same 2026 study; CTR fell nearly 60% when one appeared.

1B+

monthly users for Google AI Mode, according to Google's I/O 2026 announcement; queries have more than doubled each quarter.

Sources: SparkToro, "In 2026, Less than One Third of Google Searches Still Send a Click" (8 Jun 2026; Similarweb U.S. desktop + mobile web panel, Jan–Apr 2026); Search Engine Land (9 Jun 2026) for AI Overview context; Google, "A new era for AI Search" (19 May 2026); Similarweb AI Mode referral analysis (2026). Caveats: SparkToro excludes Google's mobile search app and uses estimated mobile/paid splits; Google figures are company-reported.

AI-SANDWICH.COM04

Buyers now start in the answer box

The first stop of the buying journey is becoming a conversation.

G2's March 2026 survey of 1,076 B2B software buyers found AI chat is no longer a research sidekick—it is where shortlists are made and unmade.

Consumers agree: at the discovery stage, 35% now rate AI tools as their most useful channel, versus 13.6% for search engines (Similarweb, Jan 2026).

83% of AI-assisted buyers report more confidence in their final choice. The conversation is doing the qualification your website used to do.

51%

of B2B software buyers now begin their purchase research in an AI chatbot—not a search engine.

69%

chose a different vendor than initially planned, based on AI chatbot guidance.

1 in 3

bought from a vendor they had never heard of before the AI conversation.

If you're not in the citations and the training data, you're not on the shortlist.

Sources: G2, "The Answer Economy: How AI Search Is Rewiring B2B Software Buying" (March 2026, n=1,076); Similarweb, 2026 AI Brand Visibility Report (January 2026 consumer panel). Caveat: survey self-reports from one category (software); directional for other B2B categories.

Citation volatility is a terrible strategy

Today's citation winner can become tomorrow's 80% decline.

In August 2026, Reddit's share of ChatGPT citations fell by roughly 80% in a single week. Semrush found that running the same prompts through two different ChatGPT reasoning modes produced only a 25.6% overlap in cited domains.

Organizing a strategy around whichever external platform AI systems favor today is exhausting. Building a durable, useful presence across the internet is not.

57%

of citations across eight major AI models come directly from brand-owned websites—the largest single source.

32.8%

of ChatGPT citations in July 2026 went to product pages, nearly doubling since March.

69%

of Google Gemini citations come from brand-owned websites.

Sources: Promptwatch / Petra Labs, August 2026 (Reddit citation decline and product-page share); Profound, July 2026 (11.8B cross-model citation analysis); Semrush, September 2026 (reasoning mode overlap). As cited by Amanda Natividad (August 2026).

AI-SANDWICH.COM06

Your homepage moved

A buyer's first impression happens across dozens of surfaces before they ever visit your .com.

Google AI Summaries

The search snippet that answers their question directly without requiring a click.

LLM Interfaces

ChatGPT, Claude, and Perplexity giving a synthesized answer that either cites you—or a competitor.

Social Feeds

Native, linkless posts optimized to capture attention without sending traffic out.

Slack & Discord

Peers dropping your brand name or sharing a direct link inside private communities.

Review Sites

One unaddressed negative review becoming an amplified trend in AI models.

Newsletters & Podcasts

The expert mention that creates demand long before Google ever captures it.

AI-SANDWICH.COM07

Measurement is now full of blind spots

Dark Social Misattribution

In a controlled test, 100% of visits from Slack, WhatsApp, Discord, and TikTok shares were reported as "direct" traffic. Analytics cannot see where they came from.

Browser Privacy Constraints

ITP and third-party cookie restrictions partition storage and downgrade referrers, making cross-site identity tracking technically fragile.

Cross-Device Fragmentation

A buyer researching on a phone and buying on a laptop cannot be reliably stitched together without a persistent, logged-in first-party identifier.

The AI Answer Gap

Influence that happens inside a ChatGPT or Gemini conversation leaves no referrer at all. The visit arrives later—branded search, direct—looking like it came from nowhere.

Traffic marked "Direct" in Google Analytics(1,100+ visits across 11 networks)

TikTok100% hidden
Slack100% hidden
Discord100% hidden
WhatsApp100% hidden
FB Messenger75% hidden
Instagram DMs30% hidden
LinkedInhidden

Sources: SparkToro/Really Good Data Dark Social experiment (2023); WebKit Tracking Prevention documentation; Google Analytics cross-platform documentation.

Takeaway: your analytics undercount influence precisely on the surfaces where modern discovery actually happens.

The board question changed

Attributed ROAS assigns observed credit. Incrementality isolates causal lift.

Acquisition channel	Click-attributed ROAS	Incremental causal ROAS	The reality
Mobile Advertising	1.53x	0.70x	Negative actual return. Substantially overstated by attribution.
Search Engine Marketing	~2.00x	0.92x	Negative actual return. Large substitution into organic SEO.

The dashboard lies. Ads often intercept people who were already going to buy.

Source: "From Attribution to Causality in Digital Advertising", IEEE Access (March 2026). Dropbox U.S. mobile and SEM market blackout test, April 2025. Corroborated by field experiments at Facebook (Marketing Science, 2019) and eBay (Econometrica, 2015).

AI-SANDWICH.COM09

Five moves for the zero-click web

1. Build on rented land
Publish native value where the audience already pays attention. Earn the goodwill and influence before you ask for the click.
2. Keep one owned channel
Feed your rented-land influence into a durable, permissioned audience relationship—typically email.
3. Drop traffic as a KPI; measure what matters
Incremental sales lift is the primary KPI. Everything else is a leading indicator.
4. Shape the public record
AI answers summarize what others say about you. Seed the evidence in reviews, communities, and PR.
5. Content as a Service (CaaS)
If your content can't tell you what job it does and who it serves (Sales, CS, Comms), it's not ready to ship.

Traffic vs. Business Reality

Traffic can be a fine secondary metric, but it's a bad thing to optimize for unless you sell CPM ads.

HubSpot's Traffic

80%

Organic search cratered after algorithm updates and AI Overviews rolled out.

HubSpot's Revenue

ATH

Stock jumped on surprising growth. Revenue and traffic are not the same thing.

Fewer clicks, better clicks: visitors who arrive after reading an AI answer show higher conversion rates and longer sessions across 2026 vendor panels—they land pre-sold. Optimize for the demand, not the visit.

AI-SANDWICH.COM

The Founder's Weekly Loop

What to do next Monday, not just what to believe by Friday.

1
Pick 2–3
channels
Where your audience actually pays attention
2
Publish 1–2 zero-click
assets per week
Standalone value, no click required
3
Repurpose
each idea
Turn one insight into multiple native formats
4
Capture
demand
Email signup, branded search, demo request, direct traffic
5
Review
monthly
Correlation dashboard, not fake precision

Repeat weekly. Review the correlation dashboard monthly. Adjust channels quarterly.

AI-SANDWICH.COM11

AISandwich

THE AI SANDWICH METHODOLOGY

Don't just build AI. Build AI that is actually used.

Most organizations pour billions into the AI layer while neglecting the business problem above it and the human adoption below it. We bridge the gap.

Charles Gourlaouen

AI Strategist & Executive Coach

25+ years building & scaling AI-driven products across Europe & Asia (SCB Bank, Expedia, aCommerce, Agoda, Accor, SNCF, Eurostar, MediaPart).

ai-sandwich.com

contact@ai-sandwich.com

Book a Strategy Session

Sources & Evidence

Zero-Click AI Search Association
Pew Research Center, "Do people click on links in Google AI summaries?" (July 2025). Observational analysis of 900 consenting U.S. adults' browsing behavior in March 2025 showing fewer traditional clicks and more session conclusions on pages with an AI summary.
Zero-Click Panel Trends
SparkToro/Jumpshot (2019, 50.33% zero-click); SparkToro–Datos, "2024 Zero-Click Search Study" (58.5%); SparkToro/Similarweb U.S. panel (Jan–Apr 2026, 68.01%). "Zero click" includes reformulated or abandoned searches, not only answered ones.
AI-First Buying Behavior
G2, "The Answer Economy: How AI Search Is Rewiring B2B Software Buying" (March 2026; n=1,076 B2B buyers). Similarweb, 2026 AI Brand Visibility Report (January 2026): AI tools rated most useful at discovery by 35% vs 13.6% for search engines.
AI Referral Economics
Similarweb AI Mode analysis (2026): referral on ~1.6–2.5% of AI Mode queries vs 17–19% for classic search. Multiple 2026 vendor panels report higher conversion rates and session depth for post-AI-answer visitors (directional, vendor-reported).

Citation Volatility
Promptwatch / Petra Labs, August 2026 (Reddit citation decline; product-page share); Profound, July 2026 (11.8B cross-model citation analysis: 57% brand-owned, 69% for Gemini); Semrush, September 2026 (25.6% reasoning-mode overlap).
Social Feed Mechanics
Official ranking documentation from Meta, Instagram, TikTok, and X confirming feeds prioritize predicted relevance, retention, and native interaction signals. Includes a field experiment from the Journal of Business Research (2021) on link placement.
Dark Social Blind Spots
SparkToro & Really Good Data (2023) experimental URL tracking; Google Analytics direct/(none) technical definitions; WebKit Tracking Prevention cross-device documentation.
Incrementality vs. Attribution
IEEE Access (2026), "From Attribution to Causality in Digital Advertising", detailing Dropbox's U.S. market blackout study. Supported by field experiments at Facebook (Marketing Science, 2019) and eBay (Econometrica, 2015).
First-Party Data Measurement
IAB/BWG Strategy "State of Data 2024"; DMA "Email Benchmarking Report 2025"; Nielsen and Litmus guidance on shifting KPIs from platform proxies to outcome influence.

Methodological Note: Much of the marketing industry's published data relies on observational panels, vendor-supplied metrics, or selected client cohorts rather than randomized causal trials. Vendor-reported figures should be treated as strong directional evidence for the shift in buyer behavior, not as universal baseline guarantees for all domains, audiences, or industries.

1 / 13

The click is no longer the conversion

The numbers behind the shift

Directional evidence from 2025–2026 panels and surveys. Every figure is sourced on the research page.

of U.S. Google searches ended without a click (Jan–Apr 2026): 68%; Source: SparkToro / Similarweb (June 2026)
of B2B software buyers begin purchase research in an AI chatbot, not a search engine: 51%; Source: G2 (March 2026)
of visits clicked a link inside a Google AI summary when one appeared: 1%; Source: Pew Research Center (July 2025)
of AI Mode queries send a visitor to a website, vs 17–19% for classic search: ~2%; Source: Similarweb AI Mode referral analysis (2026)
of B2B buyers chose a different vendor than planned, based on AI chatbot guidance: 69%; Source: G2 (March 2026)
bought from a vendor they had never heard of before the AI conversation: 1 in 3; Source: G2 (March 2026)
of citations across eight major AI models come from brand-owned websites: 57%; Source: Promptwatch / Petra Labs (Aug 2026); Profound (Jul 2026); Semrush (Sep 2026)
of visits from Slack, WhatsApp, Discord and TikTok shares were reported as “direct”: 100%; Source: SparkToro & Really Good Data (2023)

Five moves for the zero-click web

1. Build on rented land
Publish native value where the audience already pays attention. Earn the goodwill and influence before you ask for the click.
2. Keep one owned channel
Feed your rented-land influence into a durable, permissioned audience relationship — typically email.
3. Drop traffic as a KPI; measure what matters
Incremental sales lift is the primary KPI. Everything else is a leading indicator.
4. Shape the public record
AI answers summarize what others say about you. Seed the evidence in reviews, communities, and PR.
5. Content as a Service (CaaS)
If your content can't tell you what job it does and who it serves (Sales, CS, Comms), it's not ready to ship.

The Founder's Weekly Loop

What to do next Monday, not just what to believe by Friday.

1
Pick 2–3 channels
Where your audience actually pays attention
2
Publish 1–2 zero-click assets per week
Standalone value, no click required
3
Repurpose each idea
Turn one insight into multiple native formats
4
Capture demand
Email signup, branded search, demo request, direct traffic
5
Review monthly
Correlation dashboard, not fake precision

Repeat weekly. Review the correlation dashboard monthly. Adjust channels quarterly.