Field notes on the AI answer economy
Original research and practical analysis on how AI engines cite brands, and how to earn a place in the answer. Grounded in what we measure across ChatGPT, Claude, Gemini, Perplexity and Google AI.
The State of AI Search & GEO in 2026
How AI engines like ChatGPT, Claude, Perplexity and Google AI cite brands in 2026: what Generative Engine Optimization (GEO) is, the AI Citation Gap, and how to show up in AI answers.
Latest
The newest research, hot off the daily desk.
Citations and sources
Where AI citations come from, who earns them, and why the engine names one brand over another.
What AI engines actually cite: 37,547 citations analyzed
We analyzed 37,547 citations across 5,200 AI answers from ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews. Only 5% pointed to the brand's own site. Here is what AI actually cites, and what it means for GEO.
From crawl to citation: how an AI answer gets built
An AI citation has a lead-up: an engine crawls your page, retrieves it, then names you. How to win each stage, and read the crawl-to-citation timing honestly.
The AI Citation Gap: where you win on Google but lose the AI answer
Your AI Citation Gap is the set of queries you rank for on Google but never get cited for in AI answers. Here is how to find it with Search Console, prioritize it by traffic value, and close it.
Earned media drives most AI citations. Here is how to make yours count
Study after study puts earned media behind the large majority of AI citations, with journalism a big slice. That makes GEO partly a PR problem. How to earn the right coverage and measure whether AI actually cites it.
Being cited by AI is not being recommended, and the gap decides the sale
AI can pull your page as a source and still recommend a competitor by name. Citation and recommendation are different outcomes. Why the gap happens, why self-ranking best lists fail, and what to measure instead.
AI search is winner-take-most. Here is how challenger brands break in
A small share of brands capture most AI recommendations while most are invisible, and the concentration is intensifying. Why it happens, why it is not hopeless for challengers, and what to measure.
Reddit runs the AI answer now, and its spam purge just changed the GEO playbook
Reddit is one of the most cited sources in AI answers, and it is now deleting the posts planted to game them. What the GEO spam crackdown means for brands, and how to earn citations that last.
Do you need an llms.txt file to get cited by AI? What 300,000 domains say
A vendor tells you to ship an llms.txt file and the AI engines will read your site. The 2026 data says it does not move your citations. Here is when it is worth doing anyway, and how to actually test it.
Measurement and visibility
How to measure AI visibility credibly: per engine, with confidence, and over time.
How to track your brand across AI search engines
A practical playbook for monitoring how ChatGPT, Claude, Gemini, Perplexity and Google AI cite your brand: which prompts to track, how often, and how to read the numbers honestly.
AI share of voice: how to measure it without fooling yourself
Share of voice in AI answers tells you how often engines pick you over rivals. Here is how to calculate it correctly, why branded prompts ruin it, and how to read it per engine and per topic.
Your AI visibility is not one number, it is five, and they barely agree
The same brand can be everywhere on Perplexity and nearly invisible on ChatGPT. Citation rates and sources diverge enormously by engine, so a single blended score misleads. How to track and win each one.
How llemmy measures AI visibility (and what we don't claim)
Exactly how llemmy derives every number: presence as a rate, with sample size and a 95% Wilson confidence interval, over your recent answers. Plus an honest account of what a confidence interval does not fix.
Why llemmy doesn't track branded prompts like other tools
Branded prompts return ~100% AI visibility by construction, which quietly inflates your reporting. Here is why llemmy keeps them out of the headline score and routes them into sentiment and accuracy instead.
How to measure content effectiveness in AI search
Publishing content and hoping AI notices is not measurement. Here is how to run a content campaign against AI visibility: baseline first, read the change against a confidence interval, and find which pages actually win citations.
What belongs in a monthly AI search report
A monthly report is not five dashboard screenshots stapled together. Here are the surfaces that belong in one, why each matters, and the honesty rules that make a client trust it.
Local visibility in the AI era: your map pack, rating, and reviews
As AI answers absorb local search, your Google local pack and review profile decide who nearby customers see. Here is how to track them honestly, and what public tools can and cannot measure.
How fast can AI cite you? Crawl-to-citation lag, and why a daily read matters
SEO taught you to wait a quarter. AI search is faster in both directions: retrieval engines can cite a page within days, and a daily read catches it. What crawl-to-citation lag looks like, honestly.
Getting cited: optimization
The work of earning the citation: readability, structure, freshness and authority.
What makes a page AI-readable: a practical GEO checklist
AI engines cite pages they can fetch, understand and trust. Here is a practical checklist for making your content AI-readable, and the common reasons engines skip a page that ranks fine on Google.
Content optimization for AI answers, beyond the checklist
Getting retrieved and getting cited are different battles. How structure, atomic claims, formats and refreshes win citations, and how to prove an edit worked.
Content freshness and AI citations: what the data actually shows
AI answer engines rerank retrieved pages and the citation data skews hard toward recently updated content. What the studies say, where the effect is real vs correlational, and a practical refresh playbook.
E-E-A-T for generative engines: what still counts when the ranker is an LLM
Google's E-E-A-T half-transfers to AI answers. The trust signals LLMs actually pick up: entity consistency, authorship, schema, and what does not carry over.
Query fan-out: the searches an AI runs before it answers you
Before an AI engine answers a question, it often runs its own background searches first. Those searches are a direct signal of how the model frames your category, and most tools that show them are not being straight about where the numbers come from.
Fan-outs tell you what to write. Corroboration is what gets you cited
A tactical GEO guide to AI fan-out queries: read them as the model's own keyword research, cover the subjects with answer-shaped content, then earn the cross-source corroboration that actually wins the citation.
Brand, reputation and risk
What AI says about you across the answer layer, and the cost of it getting you wrong.
Brand monitoring has a new surface: AI answers
Social listening and news alerts miss the place your buyers now ask about you: AI answers. What to watch, how often to sample, and how to read it honestly.
Reputation management when the reviewer is a machine
AI engines form durable brand impressions from training data and cited sources. How to audit them, counter negative narratives, and why old ORM tactics fail.
When AI states wrong facts about your brand, and how to catch it
AI answers confidently get brands wrong: the wrong price, a product you do not sell, a founder who never worked there. It moves buyers, and you cannot fix an error you never see. How to monitor and correct it.
The AI search market
Where AI search is heading, and what it means for how buyers find and choose you.
Google says GEO is still SEO. Here is the half that guidance leaves out
Google's 2026 guidance calls optimizing for AI search still SEO. It is right about the optimizing, and silent on two things that decide whether you win: measuring citation, and every engine that is not Google.
Zero-click search is here, and being cited is the difference between gaining and losing
Most Google searches no longer send a click. The twist: brands cited in the AI answer can gain traffic while uncited ones lose it. Why citation is the new scoreboard, and how to measure it.
Agentic commerce is here: will an AI agent recommend your brand, or a rival?
AI agents are starting to shop on people's behalf. They pick from a shortlist the model surfaces, so whether AI answers already name your brand decides whether you make the cart. How to measure it.
Your buyers are shortlisting vendors inside ChatGPT. Are you on the list?
B2B buyers now ask AI for the best tools in a category and act on the shortlist it returns. If AI does not name you on the buying-intent queries, you are cut before a rep is ever involved. How to measure and win it.
Put it to work
Instrument your site, measure the traffic AI drives, and switch tools without losing history.
The llemmy Tag: measure the traffic & revenue AI search actually drives
A practical guide to the llemmy Tag: a one-line, cookieless script that measures the humans who arrive from ChatGPT, Perplexity, Gemini and other AI answers, and ties them to revenue.
Track and improve your AI visibility from WordPress
Most of the web runs on WordPress, and it is where your content actually lives. The free llemmy for WordPress plugin measures how AI engines cite you and turns citation gaps into content briefs, without leaving wp-admin.
Switching GEO tools without losing your history
The real cost of switching AI-visibility tools is the blank graph. Here is how to move your prompts, competitors and months of answer history into llemmy from Peec, and what to expect when two tools measure the same brand.
You can't fix everything in GEO. Here is what to prioritize first
Tracking AI visibility tells you where you stand. It does not tell you what to do Monday morning. A practical framework for reading demand, rank headroom, AI citation status and content coverage together, so you fix the topics worth fixing first.
See how AI describes your brand
Run a free GEO audit, no signup needed to see your score, or start tracking your brand across every AI engine.