ChatGPT Already Knows Who It’ll Recommend Before It Searches

I requested ChatGPT for the most effective AI note-taking app. Seven phrases, no model names.

Earlier than it fetched something, it wrote itself this search.

finest AI word taking apps 2026 official pricing options Granola Notion AI Otter Fireflies Fathom Mem Limitless

Learn the tail of that string. Granola. Notion AI. Otter. Fireflies. Fathom. Mem. Limitless. Seven merchandise in a single search. I’d named none of them, and nothing had come again from the online search but, so no consequence put them there. These names got here out of the mannequin.

Then it ran 9 extra searches.

website:granola.ai pricing options AI assembly notes 2026
website:otter.ai pricing AI assembly notes 2026
website:fathom.video pricing AI assembly assistant 2026
website:notion.com product AI Assembly Notes official 2026 pricing
website:fireflies.ai pricing official AI assembly notes 2026
website:mem.ai pricing AI notes official 2026
website:notebooklm.google official options pricing 2026

That’s the fan-out I wrote about within the first two elements, and now you’ll be able to see what it’s made from. One search to call the shortlist, then one search per title on that shortlist, every pointed straight at that firm’s personal web site.

So the fan-out was by no means a seek for candidates; it was ChatGPT taking place an inventory it already had, one title at a time.

Within the first two elements, I wrote that that you must survive a website:yourdomain.com probe, as a result of ChatGPT runs them. It solely runs one on you in case you have been named in that first search. Miss the shortlist and your web site by no means will get checked out, nevertheless nicely constructed it’s. The choice occurs earlier than something touches your server.

I’ve spent two months studying this visitors for half 1 and half 2. That is the very first thing I’ve discovered that adjustments what I’d inform a shopper to do.

Earlier than you learn any additional, all of this comes off one account, so each proportion beneath is a route moderately than a measurement. The mechanism is a distinct matter. You may reproduce that by yourself account in two minutes, and I’ll present you the way on the finish.

The 4 Concepts Behind This

Each units up the subsequent.

1. ChatGPT writes its personal search queries, and you’ll learn them.
2. These queries already comprise manufacturers no one talked about.
3. Being in that question is price about 33 occasions greater than being findable.
4. When you’re in, a second and far harsher filter decides who will get cited.

Thought 1: It Writes Its Personal Searches, And They’re Readable

While you ask a query, ChatGPT rewrites it into search queries of its personal, runs them, reads what comes again, then writes a solution.

These queries sit within the response your browser downloads, below a key at the moment known as search_queries. OpenAI renamed it from search_model_queries in early August 2026. Nothing about this can be a leak. Your personal browser wants that JSON to attract the web page, and you’ll learn it in DevTools by yourself account in about two minutes. Recipe’s on the finish.

Right here’s one from a query about stay chat software program.

finest AI stay chat buyer assist software program 2026
Intercom Fin Zendesk AI pricing official

I requested for “finest ai primarily based stay chat assist software program.” It added the 12 months, “official,” “pricing,” and three names.

All the pieces beneath comes from studying a couple of hundred of those.

Begin right here. Open ChatGPT, ask the “finest [your category]” query your consumers ask, and browse the question it writes again. That one string reveals you whether or not ChatGPT is aware of your model exists, which is what AI visibility audits are offered to reply. All the pieces else on this article is about what to do with what you discover in it.

Thought 2: The Shortlist Exists Earlier than The Search Runs

The apparent objection to my note-taking instance is timing. Perhaps ChatGPT searched as soon as, noticed these manufacturers, then wrote a better second question. That’d make the names a results of retrieval moderately than a trigger.

So I examined it correctly. For every dialog, I took the primary consumer message and the primary search question, ordered by timestamp. At that time nothing’s been fetched, so there’s no earlier consequence to have discovered from.

In 21 of 27 conversations, that first question contained manufacturers the consumer by no means typed.

Have a look at rows two and three. Identical query, barely completely different phrasing, and the listing grew from three names to seven. The listing stretches with the way you ask as an alternative of popping out of a set desk.

Then I ran  12 classes that don’t have anything to do with one another, to examine this wasn’t a software program factor.

11 of 13 did the identical.

The robotic vacuum row is the one which obtained me. Recalling that Roborock exists can be unremarkable. It recalled Saros, Dreame X50, and Eufy S1 Professional. Present mannequin numbers, within the first question, unprompted. No matter this data is, it goes right down to the product line.

The electrical SUV row does one thing completely different from all of the others. For accounting and remedy and internet hosting, it named distributors and went to their pricing pages. For automobiles, it named magazines. Automobile and Driver, Edmunds, Prime Gear. In a single class, its intuition is to go to the makers; in one other, the reviewers.

I obtained enthusiastic about that after which re-ran three classes to see if any of it holds nonetheless.

Language studying barely moved. 5 of six names got here again the identical. Accounting collapsed from six distributors to a single focused probe at QuickBooks. Internet hosting switched sides fully, dropped each vendor and went to a overview website as an alternative.

So distributors versus magazines is a bent that may flip between runs, not a set property of the class. Two issues survive that. The injection occurred each single time, and the class with the clearest market leaders saved its names. My guess is that settled classes have steady shortlists and contested ones wobble, however three repeats isn’t sufficient to say that.

Don’t decide your AI visibility from one reply. Run the query 5 occasions, as a result of the listing adjustments between runs.

It Occurs Each time ChatGPT Has To Provide The Merchandise

Each question up so far had the phrase “finest” in it, in order that was the very first thing I attempted to interrupt. I ran twenty-four extra queries that prevented the phrase, throughout seven completely different shapes.

“Finest” turned out to don’t have anything to do with it. What issues is whether or not ChatGPT has to give you the merchandise itself.

When it doesn’t search in any respect, none of this is applicable. Ask how noise-cancelling works, or what a vector database is, and it solutions from coaching with no net search. Identical for open-ended moaning. “We’re spending an excessive amount of on buyer assist tooling” obtained no search both. Seven of my twenty-four by no means touched the online, so there was no shortlist to get onto.

Identify the manufacturers your self, and it takes them. “Xero or QuickBooks for a small enterprise” went straight to website:xero.com/uk. “Ought to I take advantage of HubSpot for a small company” went to website:hubspot.com.

Go away the candidates to it, and it reaches into reminiscence. That occurred in ten of the eleven circumstances the place I requested for a advice with out naming something.

The third row is my favourite. I intentionally prevented saying “robotic vacuum” and named no one. It went to a selected Roborock mannequin’s personal web site on the primary search.

Row two ought to fear you in case you promote software program. That’s a criticism, not a buying query. Any person grumbling about their assembly notes obtained become Granola official pricing earlier than a single web page was fetched.

And there’s a model of this aimed straight at your opponents. All three of my displacement queries pulled in new names. “Alternate options to Zendesk” produced Assist Scout; “what can i take advantage of as an alternative of quickbooks” produced Zoho Books; and “one thing like duolingo however higher for grammar” produced Kwiziq and Babbel. When your buyer goes searching for a means out of your rival, ChatGPT nominates whoever it already is aware of. That might be you, and also you don’t get a say in it on the day.

If ChatGPT searches for a product and also you haven’t named one, it brings its personal.

Do that. Run your class query 5 occasions and write down the names that seem within the question every time. The names that present up each run are your actual aggressive set in ChatGPT’s head. Something that comes and goes is contested floor, which is the place a push can transfer one thing. In case your model by no means seems in 5 runs, you could have your reply, and it isn’t a technical one.

Thought 3: Being In The Question Is Price About 33 Occasions Extra

I sorted each model into two teams. Ones that appeared in a question ChatGPT wrote, and ones that have been solely fetched through the search with out ever being named. Then I checked how usually every group made it into the ultimate reply.

About 33 occasions the distinction.

I additionally discovered 86 circumstances the place a model obtained beneficial, and its web site was by no means fetched in that dialog in any respect. A point out doesn’t want a crawl.

That is the uncomfortable half for my very own trade. Most of what’s offered as GEO proper now’s retrieval work, which is the two.1% column. The 68.9% column is set earlier than any of it runs.

Do that. Cut up your funds to match the 2 columns. Should you’re absent from the question, the cash belongs within the work that will get you written about, reviewed, in contrast, and listed. Digital PR, class content material, review-site placement, analyst protection, displaying up within the roundups your consumers learn. Should you’re already within the question, that spend is essentially accomplished, and the technical work beneath is what’s left.

Thought 4: Being In The Question Is The Entry Ticket, Not The Win

If it stopped there, the recommendation can be “construct model fairness,” and we might all go dwelling. A second filter runs after the question, and it’s brutal.

I constructed a labelled dataset from 57 conversations. Each retrieved web page as a row, with whether or not it earned a quotation because the label. 3,554 pages.

110 obtained cited. That’s 3.1%.

ChatGPT reads round 600 pages to write down one reply and credit about 30. Nearly every part will get learn. The hole between being learn and being credited is the place the work is.

Three issues separated the cited from the ignored.

Place, exhausting. ChatGPT teams outcomes by area, and the place you sit in that group predicts practically every part.

Beneath the highest two, quotation is a rounding error. Being within the retrieved set isn’t a win in case you’re ninth.

Piling on pages hurts. When a number of pages from one area present up in the identical group, per-page conversion collapses.

Two tightly matched pages is the candy spot. Previous six, you’re largely competing with your self. I’ve been telling purchasers to consolidate for a 12 months on intuition. That is the primary time I’ve watched it occur within the information.

Relevance qualifies you; it doesn’t choose you. I scored the cited web page towards each different web page retrieved for a similar declare, on how nicely its textual content matched the sentence being supported. The cited web page sat within the prime 5% of the pool. Nevertheless it was the only finest match solely 20% of the time, and its common overlap was nicely beneath the most effective accessible.

So declare relevance builds a shortlist and one thing else picks the winner. Get into the highest 10% for a selected declare, and also you’re within the dialog. That half you management by writing. The ultimate choose includes issues I can’t see.

Do that. Take the questions your consumers ask and discover each web page of yours that solutions the identical one. That’s the underside row of the desk above, and it’s costing you. Choose the web page that matches the intent most tightly, make it the reply, and fold or redirect the remainder into it. Then ensure that the sentence that truly solutions the query sits close to the highest of that web page in plain HTML textual content, with the numbers in it.

Discovering these overlaps by hand throughout an actual website is the depressing half. Cannibalization is often judged on key phrase overlap, and that’s the mistaken unit right here, as a result of ChatGPT teams by declare moderately than by key phrase.

That is what Key phrase Insights was constructed for.

It clusters your key phrases by search intent moderately than by string match, so the pages combating over one intent floor as a bunch even once they share no key phrases.

That grouping maps carefully onto what I’m watching ChatGPT do when it collapses a site down to at least one cited web page.

Full disclosure: It’s my firm, which can also be why I do know it handles this particular job.

What I’d Really Do With This

The findings cut up AI visibility into two video games that maintain getting handled as one.

Sport One Is Being In The Class Vocabulary

If ChatGPT doesn’t already join your model to your class, it received’t title you within the question, and also you’re a 2% probability of a point out.

Schema received’t repair that, and neither will web page pace. An llms.txt file has even much less of an opportunity, as a result of your server by no means will get contacted earlier than the choice’s made.

What appears to construct it’s gradual and unglamorous. Being written about, reviewed, in contrast, and argued over throughout the open net for years, till the affiliation exists within the coaching information.

That’s digital PR and category-defining content material, which is awkward for everybody at the moment promoting technical audits as an AI technique.

Sport Two Is Profitable The Quotation As soon as You’re In

That’s the three.1%, and it’s actual. One tightly matched web page per intent, the claim-bearing sentence early, details and numbers in plain HTML textual content, and no cluster of near-identical pages combating one another.

Right here’s the entire thing as a sequence you’ll be able to run this week.

1. Learn the question. Ask your class query 5 occasions and word the manufacturers ChatGPT writes into its personal search. Ten minutes.
2. Discover which recreation you’re in. Within the question each time means you’re taking part in recreation two. By no means in it means recreation one, and no quantity of web page work will transfer you.
3. Should you’re absent, go and get written about. Evaluations, roundups, comparability pages, the publications your class reads. In a class like electrical SUVs, the place ChatGPT went to the magazines earlier than it went to any producer, that’s the one door.
4. Should you’re current, examine conversion, not retrieval. Rely how usually your pages get fetched towards how usually they get credited. Fetched lots and credited hardly ever is a web page downside, not a visibility downside.
5. Repair the web page, not the positioning. One web page per intent, the reply sentence close to the highest, actual numbers in HTML textual content moderately than in a picture or loaded by JavaScript.
6. Re-check month-to-month. The steadiness check confirmed the shortlist strikes, and the format strikes with it. In early August 2026, OpenAI renamed the important thing these queries sit below, and the fan-outs in my captures dropped from 12 searches per reply to 4. What was true in July received’t essentially maintain in September.

In my information, one model obtained fetched 66 occasions and by no means cited as soon as. That’s not an consciousness downside. The engine saved going again and saved deciding they’d nothing price quoting. Step 4 exists to catch that.

Steps one and 4 are what I constructed FanoutFox for, as a result of working them by hand for each reply will get previous quick. It’s a free Chrome extension; every part stays in your browser, and it reads your individual ChatGPT session moderately than guessing from the skin. It reveals you the fan-out queries verbatim, so the first step is a look as an alternative of a DevTools session.

For step 4, it now has a converts column, retrieved towards cited per area, the quantity this text argues you have to be monitoring as an alternative of retrieval counts. A website that will get learn repeatedly and by no means credited is the prognosis you’re searching for, and it’s sorted proper there.

The Caveats

I’d moderately you belief the elements that deserve it.

The mechanism is stable. ChatGPT writing model names into its first question is seen in a single seize, and you’ll reproduce it your self in two minutes. Each proportion comes from one account and some hundred conversations, weighted in direction of software program and AI instruments as a result of that’s what I ask about. A unique question combine strikes the numbers.

My first move was completely “finest X” queries, which meant I’d have revealed a declare about ChatGPT whereas solely testing one phrasing of 1 form of query. The 24-query boundary check exists as a result of that bothered me, and it modified the discovering moderately than confirming it.

Personalization is actual, and it confirmed up in my information. My meal equipment question got here again with UAE, Dubai, and a neighborhood firm. Journey insurance coverage went straight to a UAE insurer’s website. I by no means mentioned the place I stay. So these are my account’s shortlists, not ChatGPT’s.

FanoutFox has a personalization label, so you’ll be able to see in case your question is affected by personalization.

The 68.9% towards 2.1% comparability makes use of entire conversations, so a model discovered in flip one and queried in flip two counts as pre-known. That’s why the first-query check is the headline declare as an alternative. A 33 occasions hole is simply too huge for that contamination to clarify, however the clear quantity is the 21 of 27.

My model matching splits “Automobile and Driver” into two tokens and as soon as let “SaaS” via as a model. On the native question, it counted “Arabian Ranches” and “Dubai” as manufacturers, which they clearly aren’t. The queries themselves are the proof. Deal with my counts as approximate and browse the question strings as an alternative.

Test It Your self In 2 Minutes

You don’t need to take my phrase for any of this, which is the entire level of studying visitors as an alternative of guessing.

1. Open ChatGPT in Chrome, open DevTools, go to the Community tab.
2. Ask “finest [your category] 2026”.
3. Filter for dialog, open the response, seek for queries.
4. Learn the primary question the mannequin wrote and search for names you by no means typed.

In case your opponents are in that listing and also you’re not, now you recognize, and it took much less time than making a espresso.

That’s the handbook model, and it’s effective for one examine. FanoutFox runs the identical examine on each reply you give it, retains a historical past so you’ll be able to watch a class shift over weeks, and by no means sends your conversations anyplace. It’s free, and it’s the device I take advantage of for all of this myself.

What’s Subsequent

The hole in all that is that it’s one account. I’ve requested my full ChatGPT information export to run the identical evaluation throughout each dialog I’ve ever had. The follow-up that may settle it’s the identical class sweep on a number of accounts directly. If the identical manufacturers get injected for a similar questions on different folks’s accounts, it’s model-level information. In the event that they don’t, personalization is doing extra work than anybody thinks.

Both reply is price having, and I’ll publish whichever one turns up. Something that strikes goes within the ChatGPT analysis tracker.

Learn off a logged-in ChatGPT Plus account in Dubai between 24 and 25 July 2026. 57 conversations for the quotation numbers, 27 for the first-query check, twelve recent class queries, and three repeats. Structural findings are stable at this pattern. Each proportion is directional.

Extra Sources:

This put up was initially revealed on Suganthan.

Featured Picture: Prostock-studio/Shutterstock