ChatGPT Says Di Fara Is the Best Pizza in NYC. Gemini Says Joe’s. Both Are 100% Sure.

What happens when someone asks ChatGPT for the best pizza in New York City?

We decided to find out. Not by asking once and writing up the answer. We tracked dozens of real queries across ChatGPT and Google Gemini. Different phrasings, different angles. “Best pizza near me.” “Authentic New York slice.” The kind of questions real people actually ask every day.

The results were wild.

Two Platforms, Two Completely Different Winners

ChatGPT thinks Di Fara Pizza is the undisputed champion. Out of every query we ran, Di Fara came up every time. 100% mention rate, average position #1. Not #2, not #1.5. Dead first, every time.

Gemini disagrees. Its pick is Joe’s Pizza. Also 100% mention rate, average position #1.5. Gemini is just as confident about Joe’s as ChatGPT is about Di Fara.

Both are legendary spots. Both deserve praise. But the fact that two platforms looked at the same city, the same category, the same question, and arrived at completely different #1 picks? That changes what it means to be “findable” online.

The Invisible Famous

Here’s where it gets strange.

Lucali is one of the most celebrated pizzerias in New York. Long waits, celebrity fans, glowing press coverage for years. On ChatGPT, Lucali shows up in 88% of queries at position #2. A strong second place.

On Gemini? 12%. Position #14.

Prince Street Pizza went viral for its pepperoni square. Thousands of Instagram posts, constant foot traffic. ChatGPT mentions it 100% of the time. Gemini? 25%.

And then there’s the opposite story. Mama’s Too on Gemini shows up in 100% of queries. Every single time. On ChatGPT, Mama’s Too doesn’t exist. Zero mentions. Not once.

These aren’t obscure neighborhood spots. These are some of the most talked-about pizzerias in one of the most competitive food cities on Earth. And depending on which platform a potential customer opens, they either dominate or completely vanish.

There Is No Second Page

This matters because ChatGPT and Gemini work nothing like Google Search.

On Google, a business might land on page two or three. Not ideal, but still findable. Someone persistent enough will scroll. There are ads, maps, local packs, all sorts of ways to show up.

ChatGPT doesn’t work that way. Someone asks for pizza in NYC, they get a list of maybe five places. That’s it. No page two. No “show more results.” No sponsored slot to buy. The businesses in that list get all the attention. Everyone else might as well not exist for that particular customer.

Our data showed this concentration clearly. ChatGPT’s top three picks captured 43% of all mentions. There are over 1,800 pizzerias in New York City. Three of them got nearly half of ChatGPT’s attention.

Gemini spread recommendations a bit wider. But even there, the math is brutal. Nearly 40% of businesses mentioned by Gemini appeared exactly once. That’s not visibility. That’s noise.

Same Restaurant, Four Different Spellings

We noticed something odd while going through the Gemini results.

L’Industrie Pizzeria showed up under four different spellings. “L’Industrie,” “L’industrie Pizzeria,” “L’Industrie Pizza,” “L’Industrie Pizzeria.” Same restaurant, counted four separate ways. Rubirosa appeared as both “Rubirosa” and “Rubirosa Ristorante.” Lombardi’s showed up with and without “Coal Oven Pizza” appended to its name.

This means the real concentration is even tighter than raw numbers suggest. Fewer unique businesses are actually getting recommended. And for anyone trying to figure out how their business appears when people ask ChatGPT, this inconsistency makes simple name-searching unreliable. An owner could check for their exact business name and think they’re invisible, when they’re actually being mentioned under a slightly different variation.

Even 100% Isn’t Equal

Joe’s Pizza and Mama’s Too both hit 100% mention rate on Gemini. Every query, both names appeared. But Joe’s average position was #1.5. Mama’s Too averaged #8.

That gap matters. Most people reading a ChatGPT or Gemini response scan the first few recommendations and stop. Being mentioned isn’t the same as being mentioned first. There’s a quiet hierarchy even among the businesses that get named, and position in the list determines who actually gets the visit.

Nobody Told These Business Owners

Here’s the thing that keeps coming up in this kind of research.

The businesses winning in these recommendations mostly have no idea they’re winning. And the businesses losing have no idea they’re losing.

A pizzeria owner checking their Google Business profile sees reviews, star ratings, maybe some analytics about how many people viewed the listing. None of that tells them whether ChatGPT is sending customers their way or sending customers somewhere else. It’s a separate channel that runs on different rules, pulls from different signals, and produces completely different results depending on the platform.

The shift is already happening. Millions of people now ask ChatGPT or Gemini for recommendations instead of (or before) typing into Google. A question like “best pizza near me” that used to produce a Google Maps result with twenty pins now produces a ChatGPT answer with three names in a paragraph. The funnel got narrower. The stakes for being in that answer got higher.

What Makes ChatGPT Pick One Business Over Another

We’re still mapping this out. It’s early and the picture is complex. But a few patterns keep showing up.

ChatGPT and Gemini pull from a mix of sources. Review sites, food blogs, news articles, social media mentions, curated lists. The weight each platform gives these sources is different, which partly explains why they disagree so dramatically. ChatGPT seems to lean heavily on certain editorial sources and aggregated sentiment. Gemini appears to weight a broader range of mentions but with its own biases.

What’s clear is that the old playbook of “get good Google reviews and optimize your Google Business profile” is no longer enough on its own. Those things still matter. But a business with great Google reviews might still be invisible on ChatGPT if it lacks mentions in the editorial and conversational sources that ChatGPT draws from.

Different platforms read different parts of the internet with different priorities. What makes a business show up on one might do nothing on another.

So What Now

Showing up in ChatGPT and Gemini recommendations is becoming a real factor in how businesses get new customers. Not someday. Right now. And for most businesses, it’s a complete blind spot.

The pizza example is one category in one city, but the pattern repeats everywhere we look. A handful of businesses dominate recommendations. The rest don’t show up. The winners change depending on the platform. And almost nobody is tracking any of it.

Reachd.ai measures how restaurants and local businesses rank in ChatGPT, Gemini, and other recommendation engines, tracking mention rates, competitive positioning, and platform-specific blind spots so owners can see exactly where they stand when the next customer asks where to go.

Because over 1,800 pizzerias are competing for attention in New York. And right now, ChatGPT only talks about five of them.