The Visibility Report #13

Week of April 14 -- April 20, 2026

Ahrefs dropped what might be the most practically useful AI search study we've seen: 1.4 million ChatGPT prompts analyzed to understand why some pages get cited and others don't. Meanwhile, Adobe data confirms AI traffic converts better than organic for U.S. retailers, Google expanded AI Mode into agentic territory, and the tool blogs had another busy week. Let's get into it.

🔥 Why ChatGPT Cites Some Pages and Ignores Others

Ahrefs analyzed 1.4 million ChatGPT prompts and found something that should change how you think about AI search optimization: ChatGPT retrieves dozens of pages in the background -- but only cites roughly 50% of them. The study reveals the signals that predict citation vs. being silently retrieved and ignored. Authoritative backlink profiles, structured content, clear attribution, and entity signals all move the needle. Retrieval is necessary but not sufficient -- citation is the actual goal, and it requires something more than just ranking well.

📎 Ahrefs​

💰 AI Traffic Converts Better Than Non-AI

Adobe data reported by Search Engine Land: AI-referred traffic converts better than non-AI traffic for U.S. retailers. I think this is the number that will finally move budget conversations in most organizations. The quality argument for AI search visibility has always been implicit -- this makes it explicit. If AI-referred visitors buy more, the ROI math for GEO investment changes significantly. Worth bookmarking before your next strategy review.

📎 Search Engine Land​

🤖 Google AI Mode Goes Agentic

Two Google moves this week signal a structural shift, not a feature update. First: AI Mode in Chrome now opens sources in a side-by-side split view -- making the source-browsing behavior more visible and more common. Second: Google's agentic restaurant booking through AI Mode is now rolling out worldwide. AI agents that take action on behalf of users (making reservations, not just surfacing information) are moving from demo to deployment. So if you're in any service business -- healthcare, hospitality, professional services -- this is the shift to watch.

📎 Google (AI Mode Chrome) | Semrush (Agentic Booking)​

🌐 The Agentic Web Is Here

Marie Haynes published a detailed breakdown of 5 AI shifts she's actively discussing with clients. It maps well to what we're seeing -- the move from AI as answer machine to AI as task executor changes the optimization surface entirely. Separately, Search Engine Journal's complete guide to agentic commerce goes deep on what "selling to AI" actually requires at the content and infrastructure level. If your clients have e-commerce or transactional goals, both are worth a read this week.

📎 Marie Haynes | Search Engine Journal​

📊 Bottom-of-Funnel Content Wins in AI Search

Search Engine Land reports that bottom-of-funnel content -- comparisons, reviews, "best of" lists, specific use-case content -- is earning disproportionate AI citation rates compared to top-of-funnel awareness content. This fits the Ahrefs citation finding: content that matches high-intent queries with clear, structured answers gets cited. Combined with the Adobe conversion data above, the strategic signal is consistent: AI search rewards specificity and intent-alignment, not volume.

📎 Search Engine Land​

🌐 Machine-First Architecture and the English Problem

Two technical pieces worth flagging this week. SEJ's machine-first architecture piece argues that websites aren't built for AI agents -- and that the gap is getting expensive as agents become primary discovery surfaces. Separately, SEJ also published a useful reality check: AI visibility strategies don't work outside English -- non-English content faces a materially different citation environment, and most published frameworks assume English-only.

📎 SEJ (Machine-First) | SEJ (Language Limits)​

📈 Zero-Click Survival and the Stanford AI Index

SparkToro's breakdown of 5 strategic features that predict survival in the zero-click era is one of the more useful tactical frameworks we've seen for the attention-without-clicks problem. And the Stanford 2026 AI Index confirms the macro context: AI adoption outpaced both the PC and internet adoption curves. We're not in the early adopter phase anymore -- this is the mainstream phase, which means the optimization window is narrowing.

📎 SparkToro | SEJ (Stanford AI Index)​

📋 AI Search Audits and Reputation Risk

iPullRank published a practical guide on what AI search audits can actually tell you about your site -- including blind spots you can't see in traditional analytics. And seoClarity flags something most teams aren't tracking: the reputation risk of AI inaccuracy. When AI engines generate wrong information about your brand, the correction cycle is slow and the damage is real. Monitoring AI-generated mentions is becoming a brand safety issue, not just an SEO issue.

📎 iPullRank | seoClarity​

💭 OpenAI Starts Running Ads

Search Engine Land reports OpenAI has begun rolling out ads in select markets. LLMrefs data shows ChatGPT ads now appear in nearly 20% of U.S. responses. This changes the organic/paid calculus for AI search -- the question of whether AI search is a pure organic play just got more complicated. I think the paid layer in AI search will develop faster than most teams expect.

📎 Search Engine Land | LLMrefs​

📝 From the Tool Blogs

Ahrefs:Why ChatGPT Cites One Page Over Another -- 1.4M prompts analyzed. The top story this week, referenced above. If you read one thing, read this.

Profound:Semrush integration nodes for Profound Agents -- connect Semrush data directly into AI visibility workflows. Also: Which retailers does ChatGPT actually send shoppers to? -- Target and Walmart analysis with specific signal breakdown. And Custom Dashboards for tailored AI visibility reporting.

Peec AI:Top domains cited by AI search -- 30M sources analyzed -- Reddit, YouTube, LinkedIn, Wikipedia consistently dominate. Also: MCP integration launches, and how to measure AI search visibility and revenue -- the KPIs that actually matter.

BrightEdge:You Do Not Need a Different Strategy for Every AI Platform -- consolidation argument with supporting data. Also: Three Key Insights from AI Overview and ChatGPT Analysis, and What Share of Voice Really Means for Search in 2026.

AirOps:Content Publish Tracking -- connect every content action to its downstream AI visibility outcome. Also: AirOps Champions launches.

Scrunch:GLP-1 conversation analysis -- AI search funnel data from millions of health conversations, useful pattern for any regulated industry. Also: Sentiment Trends for tracking topics and personas driving brand perception, and Q1 2026 AI search digest lookback.

Semrush:What Is Answer Engine Optimization? -- foundational explainer. Also: GEO: A Practical Guide and Google Knowledge Graph: What it is and why it matters.

Otterly AI:Recommendations data to done -- closing the loop between AI visibility insights and content action.

LLMrefs:ChatGPT ads now appear in nearly 20% of U.S. responses -- data point worth tracking as the paid layer in AI search develops.

seoClarity:Reputation Risk of AI Inaccuracy -- when AI gets your brand wrong, the correction cycle is slow. Monitoring AI-generated mentions is now a brand safety issue.

Foundation Inc:Introducing the Marketing Engineer role -- the hybrid of content strategy and technical execution that AI-era marketing requires. Also: Reddit Statistics: 80+ data points -- useful context given Reddit's dominance in AI citations.

🏢 Agency Insights

Search Influence:New Meta Update May Shift Your Reporting -- April Client Insider -- Meta's attribution changes are affecting how paid social performance appears in dashboards. If your reporting looks off this month, this is worth checking before you assume the campaigns are underperforming.

Foundation Inc:Introducing the Marketing Engineer -- The role emerging at the intersection of content, data, and AI execution. I think this is the job description that wins in the next 24 months.

Seer Interactive:How to Optimize for Both Human and AI Visitors -- practical framework for building content that works across both traditional search and AI-mediated discovery. One of the more balanced tactical pieces we've seen on the dual-audience problem.

💡 So What Do You Do About It?

The Ahrefs study is the thread to pull this week. ChatGPT is retrieving your content more often than you think -- the question is whether it cites it or silently moves on. Retrieval without citation is noise. The study gives you a clear signal: structured content, authoritative backlinks, and entity clarity are what turn retrieval into citation.

The Adobe conversion data makes the business case straightforward. AI-referred visitors convert better. So the question isn't "should we invest in GEO" -- it's "how fast can we close the gap between retrieval and citation?"

🔮 What to Watch

1. The citation gap is the real metric. Ahrefs found ChatGPT retrieves dozens of pages per prompt but only cites ~50%. Your goal isn't to be retrieved -- it's to be cited. Structured content, clear attribution, and authoritative signals are what determine which side of that 50% you land on.

2. Google AI Mode getting agentic -- fast. Restaurant booking going worldwide is a proof-of-concept for AI taking transactional actions, not just informational ones. Side-by-side source browsing in Chrome changes the UX around how people engage with cited sources. Watch how agentic actions expand beyond local service categories in the next 60 days.

3. AI traffic quality is now documented. Adobe's conversion data for U.S. retailers gives you the ROI argument. As more industries get their own data points, expect AI search budget conversations to accelerate. The teams that built their GEO foundation early will have the case studies to show for it.

Test one thing this week: run your top 5 pages through an AI citation check (Peec AI, Otterly, or just a manual ChatGPT/Perplexity test). Are you being retrieved? Are you being cited? If you're retrieved but not cited, the Ahrefs study gives you the checklist to close that gap.

The Visibility Report | Will Scott​
This newsletter is produced collaboratively by Will Scott and Bob, an AI agent. Human oversight, AI efficiency.