top of page

Volpini: ChatGPT using Entities - the hidden Entity Layer

  • Autorenbild: th3s3rp4nt
    th3s3rp4nt
  • 5. Nov. 2025
  • 3 Min. Lesezeit

Aktualisiert: 5. Nov. 2025

Key Takeaways:

  • By recording, parsing, and analyzing these real-time data flows [from ChatGPT (Server-Sent-Events)], [Volpini] uncovered a hidden layer of entity infrastructure that extends far beyond language understanding, now encompassing products, organizations, people, and even moderation logic. [Therefore, he built] a custom Playwright-based recorder in Python to intercept and log ChatGPT’s internal /backend-api/conversation SSE stream.

  • LLM Chats / ChatGPT evolve by building structured Knowledge Graphs to understanding and answers similarly to Google

  • As Profound found a great shift in ChatGPT rankings indicating an algo-update they calles "The Entitiy Update"

  • Gabe accordingly sees a shift towards LLM/AI Chats evolving towards Google: ... the need for AI Search platforms to develop core visibility systems, anti-spam systems, understand site-level quality, authority, and more. Again, big changes coming across all of those platforms

  • Ray comments: the GEO industry learns about E-E-A-T, entities and the Knowledge Graph

  • SEO/GEO implications:

    • structured data feeds are key for products / e-commerce / events etc.

    • good SEO best practices to form EEAT apply more and more to LLM Chats while spam tactics will eventually work less effective

    • optimize and maintain your clear stance in the Knowledge Graph

  • Promptwatch already included Entities in its LLM tracking



Product meta data from ChatGPT stream:

{
  "id": "2997526925583449256",
  "title": "Bialetti Moka Express (classic size)",
  "price": "€23.90",
  "rating": 4.7,
  "num_reviews": 5900,
  "merchants": "Unieuro + others",
  "featured_tag": "classic everyday model",
  "image_urls": ["https://...jpg"],
  "metadata_sources": ["p2"]
}
  • Product IDs are 18–20 digit numeric codes — matching Google Shopping catalog IDs, not GTINs.

  • All product URLs are empty strings — ChatGPT renders them internally, without external navigation.

  • The provider field ("p2") is consistent — suggesting a single, centralized product data source.


Mapping schema.org Mark-up to SSE entities:

SSE entity class

Primary Schema.org type

Key JSON-LD properties to include

Where it helps

person_entity

Person

name, sameAs, url, image, jobTitle, worksFor

Author panels, citations, E-E-A-T surfaces

organization_entity

Organization

name, url, logo, sameAs, brand, contactPoint

Brand cards, site links, knowledge answers

place_entity

Place or LocalBusiness

name, address, geo, openingHours, telephone, sameAs

Local AI results, maps, travel panels

event_entity

Event

name, startDate, endDate, location, organizer, offers

What’s on, ticketing answers

brand_entity

Brand

name, url, logo, sameAs, aggregateRating

Product clustering, brand overviews

product_entity

Product + Offer

name, description, image, brand, color, material, weight, category, inProductGroupWithID, sku, gtin*, isVariantOf; offers.price, offers.priceCurrency, offers.availability, offers.itemCondition, offers.inventoryLevel

AI shopping answers, price and availability, product comparisons


Product attribute map:

Feed field or detected attribute

Example value

Schema.org property

Notes

id

shopify_US_9020023177513_47811512729897

sku

Use variant id as sku

item_group_id

shopify_US_9020023177513

inProductGroupWithID

Stable group id across variants

title

Slim Ski Bag Martian Red

name

Keep variant-level name

description

Travel light and stylishly…

description

Plain text preferred

link

url

Canonical PDP URL

image_link

image

Array supports multiple images

additional_image_link[]

…01.jpg, …03.jpg

image

Append to image array

price

199.99 USD

offers.price and offers.priceCurrency

priceCurrency from suffix

sale_price

163.99 USD

offers.priceSpecification or a second Offer

Use UnitPriceSpecification or separate Offer for sale

availability

in_stock

offers.availability

Map to InStock, OutOfStock, PreOrder

inventory_quantity

2

offers.inventoryLevel

QuantitativeValue with value 2

condition

new

offers.itemCondition

NewCondition, UsedCondition, etc.

brand

Black Briar USA

Wrap in Brand object when possible

color

Martian Red

color

Free text ok

material

N/A

material

Omit if unknown

weight

10 lb

weight

QuantitativeValue if you can parse unit

product_category

luggage & bags > backpacks

category

Use your site taxonomy or GPC

age_group

adult

audience.audienceType

Or PeopleAudience with audienceType

gender

unisex

audience.suggestedGender

Or audienceType “unisex”



Sources:

© 2026 David Epding.            Erstellt mit Wix.com.

david epding logo

David Epding ist GEO & SEO, Data Analytics und Automation Manager mit über 10 Jahren Erfahrung in Technischem SEO mit breiter Expertise für LLMs und langjähriger Erfahrung in der Daten-Analyse.

bottom of page