Are AI recommendations personalized?

By the Pickgap team · Updated Oct 15, 2026 · 1 min read

Partly. Location is the biggest factor: 'near me' questions depend on where the person is. Account features like ChatGPT's memory, past chats and custom instructions can also shape answers, and Google may personalize some results. The core of the answer, which businesses the web supports for that question and place, is shared. To test fairly, use a fresh session and name the city.

What personalizes an answer

FactorEffect
LocationStrong: decides which area "near me" means
WordingStrong: different conditions, different answers
Memory and past chatsCan shape answers if the assistant knows preferences
Custom instructionsCan shape tone and priorities
Signed-in Google historyMay personalize some Google results

What's shared

The pool of businesses an AI can recommend for a given question and place comes from the same web for everyone: the same directories, reviews, lists and websites. Personalization shifts the answer around that pool; it doesn't invent a different set of local businesses for each person.

Why it matters for testing

Your own results are among the most personalized: you've probably searched your business many times, your location is your office, and your assistant may remember you. That's why owners sometimes see themselves named when customers don't.

To see something closer to a typical customer's answer:

Pickgap asks every question as a new searcher located in the business's city, with no account history, so results are comparable over time and between businesses.

What you can influence

You can't control anyone's location, memory or wording. You can control the shared part: what the web says about you. Reviews, listings on the sites AI reads, and a website that states what you do and where are the inputs every personalized answer draws from. See how to get recommended by ChatGPT.