Guides
How visibility in AI answers works: clear definitions, real mechanisms, and the levers that actually move them.
Before an AI can recommend a company, it has to know which one it is. What the Wikidata item, the sameAs property and your profiles connect, why the link has to run both ways, and the example of Simafri, recorded before and after 14 September 2026.
Continue reading
Twelve buying questions about AI phone assistants, put to ChatGPT, Claude and Perplexity on 26 August 2026: 36 answers, 366 citations, 168 domains. All of it is traceable, right up to the moment the prospect picks up the phone. What is visible before, what the call carries over, and the one field that survives it.
Continue reading
Profound publishes one answer engine at $99 a month and three at $399, with nine on the tier priced on request. What the twenty-four tools the engines name publish at the same money, read on 2026-08-18.
Continue reading
Which sources do the engines lean on when they answer your market? The tools that analyse AI citations, compared on published facts, read on 2026-08-18, with a worked example per engine.
Continue reading
A tracking series is worth what it keeps. The tools that track LLM mentions, compared on what each vendor publishes about retention, coverage and price, read on 2026-08-18.
Continue reading
Generative engine optimization is a measurement you keep, so the pricing model decides before the feature list. Subscriptions and prepaid usage compared on published facts, read on 2026-08-18.
Continue reading
The AEO tools the engines themselves name, compared on published facts: entry price, answer engines covered, raw answer traceability, API and MCP access. Read on 2026-08-18, selection measured on our panels.
Continue reading
Ads have run in ChatGPT since February 2026, across Europe since late August and in Southeast Asia and Taiwan since late September, on the Free and Go plans. Where the slot sits, how one ad is chosen over another, how the auction prices it, what it is not allowed to touch, and what advertisers have published about running it.
Continue reading
Nobody at an AI lab hand-picks websites. A shared crawl, documented filters, deduplication, licensing deals and sampling weights decide. What each stage keeps, what it drops, and what that implies for the way you publish.
Continue reading
The Model Context Protocol dropped its handshake on 28 July 2026. What the revision removes, what replaces it, the requirements that fail quietly, and a checklist to run against your own server.
Continue reading
Connect Claude to your AI visibility data and it becomes your GEO analyst: it reads your series engine by engine, quotes the answers behind a number, and turns a week of measurement into actions. The setup is one connection; the questions are ready to ask.
Continue reading
The half of AI visibility you do not write yourself: what a corroboration is, the families of sources you can obtain, how to weigh listening against obtainability, and how to date what you win. Measured on 923 archived AI answers.
Continue reading
Twenty-four AI visibility platforms side by side: pricing model, entry price, engines named, API and MCP access, raw answer traceability. Every line is sourced from the vendor's own pages, read on 2026-08-18, and the selection is measured on our panels.
Continue reading
GEO is the practice of making a business visible in AI-generated answers (ChatGPT, Claude, Gemini, Perplexity), the way SEO targets search results. Definition, synonyms, and how it is measured.
Continue reading
ChatGPT does not prefer your competitors: it assembles answers from sources where they are more visible than you. Here is what actually happens, and how to check it.
Continue reading