Comparisons
AI Visibility and an Editorial Plan Using the Features Page
How to use Pico Neural's features page to track a site's AI visibility when Google and AI engines like ChatGPT, Perplexity, Gemini, Grok and Google AI each read content differently.
Say you run a blog and want to understand why a page reads well on Google but never turns up on ChatGPT, Perplexity, Gemini, Grok or Google AI. Pico Neural's features page shows these readings as separate observations, and from there you build your editorial plan. The ground rule is simple: don't treat Google's reading as if it were what the AI engines return. They are separate observations, and they need to be read separately.
What the features page shows
Pico Neural's public pages name ChatGPT, Perplexity, Gemini, Grok and Google AI. Claude and Grok are measured only after you connect them. Gemini and Google AI Overviews are not the same observation, so keep them apart even when you look at them at the same moment. These are the limits of what is observed, not a list of promises.
From observation to editorial plan: the Pico Neural method
What follows is a working proposal from Pico Neural, not a universal standard. Its job is to turn what you see on the features page into editorial decisions you can check.
- Fix a set of queries per engine. Pick a small group of real questions from your audience and record, for each one, what every engine named on the features page returns. Check: for each query you have one row per engine, and Gemini and Google AI Overviews are never merged into the same box.
- Mark where your site is missing. Compare whether your site appears in AI engine answers with whether it appears on Google. Check: you list separately the cases where Google reads the page but the AI engines do not return it, and the other way round.
- Connect the engines that require it. Claude and Grok are measured only after you connect them; without that step you have no observation for them. Check: in your sheet, the Claude and Grok columns stay empty until the connection is active.
- Turn the gaps into editorial briefs. For every query where your site is missing on an AI engine, write a brief that answers that question with your own content. Check: each brief is tied to one query and one specific engine, not to “all engines” as a block.
- Read again after publishing. Repeat the observation on the same queries and compare. Check: you have a first-reading date and a second-reading date for the same query and the same engine.
Why not merge the readings
If you treat Google and the AI engines as a single source, your editorial plan loses its starting point: you no longer know which reading you are fixing. Keeping the observations separate lets you write content aimed at what a specific engine does not return, instead of chasing one single ranking.
The practical contribution of this piece is a two-way worksheet we call the “query-engine matrix.” The columns hold the engines named on the features page (ChatGPT, Perplexity, Gemini, Grok, Google AI), with Claude and Grok marked “to connect” until the connection is active; the rows hold your audience's queries. In each cell you note three things: whether your site appears, which page appears, and the date of the observation. The matrix is not there to produce a score. It is there to show where the editorial plan has a real hole. A query that is covered on Google but not on an AI engine becomes a brief; a query covered everywhere needs no new content. Keep Gemini and Google AI Overviews in two separate columns, because they are not the same observation. This sheet is a working proposal from Pico Neural, not a universal standard: adapt it to the number of queries you can actually read again.
One limit to keep in mind
The features page shows observations on the engines it names. It does not tell you why an AI engine reads a page or does not, and an API response is not what a person sees on screen. The editorial plan that comes out of it therefore has to be read again over time, not treated as a final snapshot.
For details on the observation and on the boundaries between readings, see the in-depth page.
Editorial planning across Google and AI engines: the Pico Neural method
AI Visibility: Build Your Editorial Plan From the Features Page