Answers
Pico Neural FAQ: AI engines are separate, not one reader
The Pico Neural FAQ lists five AI engines as separate entries. Read each one on its own, because a citation in ChatGPT and silence in Perplexity are two observations, not a contradiction.
The FAQ lists five separate engines, and each one has to be read on its own because each reads and answers differently. If a page gets cited in ChatGPT but not in Perplexity, you are not looking at contradictory data. You are looking at two observations from two different readers. Anyone who opens the Pico Neural FAQ finds a map of observations, not a single way of reading a site: ChatGPT, Perplexity, Gemini, Grok and Google AI appear as separate entries, and the question about ChatGPT and Google cannot be settled by looking at one engine alone.
What the FAQ actually covers
The public pages name ChatGPT, Perplexity, Gemini, Grok and Google AI. That is the perimeter the FAQ makes visible. Two details change how you read it:
- Claude and Grok are measured only after you connect them. Until the account is connected, they are not observed.
- Gemini and Google AI Overviews are not the same observation. Read them as two entries, not as a duplicate.
The practical point follows from this: the FAQ does not tell you how AI reads you. It tells you which readers you are looking at and which ones you are not.
The common mistake: treating the FAQ as one verdict
The typical error is to add up the entries and conclude that the site is cited or that it is not. The FAQ does not work that way. If Perplexity answers by citing a page and ChatGPT does not cite it, the FAQ is showing you two reading behaviours, not an average result. Treating the engines as one reader erases exactly the information you need: which engine, on which page, with what kind of answer.
How to use the FAQ, step by step
- Start from the list of engines. Check: the FAQ shows ChatGPT, Perplexity, Gemini, Grok and Google AI as separate entries. If you are reasoning about an engine that is not on the list, the FAQ does not cover it.
- Check the status of each engine. Check: Grok and Claude count as measured only after you connect them. If you have not connected them, the entry is a pending item, not an observation.
- Keep Gemini and Google AI Overviews apart. Check: they are two distinct observations. If you note them on a single line, you lose the comparison.
- Record one line per engine, not a total. Check: each line gives the engine, the status and the type of answer. No line combines several engines.
A worked example: one row of the table
This is a Pico Neural working proposal, not a universal standard. Imagine a services page that ChatGPT cites but Perplexity does not. The completed row would read: engine: ChatGPT — status: observed — answer type: page cited. The next row: engine: Perplexity — status: observed — answer type: same page not cited. Read together, the two rows do not say the site is halfway there. They say two readers handled the same page differently. The FAQ is there to tell you that you are watching two readers, not to decide which one is right.
The contribution of this piece is a way of reading the FAQ without aggregating. When you open a page that lists several engines, the temptation is to build an average: three out of five cite it, so it is fine. The Pico Neural FAQ offers no common denominator, because the engines are not variants of the same reader. Our method is the single row: one engine, one status, one answer type. Applied to the ChatGPT and Perplexity example, the method shows at once that the difference is not an error to correct but information to keep. It also guards against a second mistake: confusing the absence of an observation with a negative result. If Grok is not connected, its row does not say not cited. It says not yet observed. Keeping these three pieces of information separate — engine, status, answer type — is what makes the FAQ usable as a map rather than a report card.
The limit to keep in mind
The FAQ tells you which engines are observed and how to read them separately. It does not tell you why one engine cites a page and another does not, and it does not turn a row into a forecast. An API answer is not a person's screen: what you read is an observation of the engine, not the experience of a real user.
For the full picture of the entries and their statuses, see the detailed page.