{"id":31608,"date":"2026-09-15T09:56:20","date_gmt":"2026-09-15T07:56:20","guid":{"rendered":"https:\/\/wordlift.io\/blog\/en\/?p=31608"},"modified":"2026-09-15T09:56:27","modified_gmt":"2026-09-15T07:56:27","slug":"chatgpt-local-entities","status":"publish","type":"post","link":"https:\/\/wordlift.io\/blog\/en\/chatgpt-local-entities\/","title":{"rendered":"How ChatGPT Resolves Local Business Entities: Google Place IDs, lbe1 and schema:hasMap"},"content":{"rendered":"\n

We traced ChatGPT\u2019s local-business identifiers to real-world Google Place identities, observed how lbe1<\/code> powers the entity experience, and found that selecting an entity can shape the next search and conversation.<\/strong><\/p>\n\n\n\n

In our previous research on ChatGPT\u2019s named entities<\/a> and OpenAI\u2019s emerging semantic layer<\/a>, we showed that ChatGPT<\/a> increasingly operates on structured entities, not just pages.<\/p>\n\n\n\n

Our latest experiments expose another part of that layer: local businesses<\/strong>.<\/p>\n\n\n\n

\"\"
From query to entity-scoped retrieval: ChatGPT resolves a place, opens an entity panel, and carries that entity into follow-up search.<\/strong><\/figcaption><\/figure>\n\n\n\n

We can now observe this chain:<\/p>\n\n\n\n

Publisher entity\n      \u2193\nGoogle CID\n      \u2195\nGoogle Place ID\n      \u2195\nChatGPT local-business entity\n      \u2193\nlbe1_...\n      \u2193\nEntity panel\n      \u2193\nEntity-scoped conversation<\/code><\/pre>\n\n\n\n
The relevant AI visibility question is becoming:<\/p>\n\n\n\n
\n
Does the AI correctly identify the real-world entity behind the page \u2014 and carry that identity forward into the conversation?<\/strong><\/p>\n<\/blockquote>\n\n\n\n
ChatGPT has a structured local-business layer<\/h2>\n\n\n\n
Using the ChatGPT Search Observatory<\/strong>, we captured the Server-Sent Events generated during local search.<\/p>\n\n\n\n
For:<\/p>\n\n\n\n
\n
\u201cBest ski location near Mariapfarr\u201d<\/p>\n<\/blockquote>\n\n\n\n
ChatGPT represented Fanningberg as:<\/p>\n\n\n\n
category: local_business\n\nentity_lookup_id:\nlbe1_N0fAahpRpdylU3ed4w5PuLQQ\n\nname:\nFanningberg\n\naddress:\nFanningberg 151, 5573 Wei\u00dfpriach, Austria\n\nrating:\n4.7\n\nreview_count:\n1596\n\nwebsite:\nfanningberg.info<\/code><\/pre>\n\n\n\n
The same lbe1<\/code> was then reused in the map widget.<\/p>\n\n\n\n
We observed the same pattern for Gro\u00dfeck\u2013Speiereck and Obertauern.<\/p>\n\n\n\n
\n
ChatGPT currently uses lbe1_...<\/code> identifiers as entity_lookup_id<\/code> values for objects classified as local_business<\/code>.<\/strong><\/p>\n<\/blockquote>\n\n\n\n
This goes beyond Named Entity Recognition<\/a>. ChatGPT materializes an operational place object with coordinates, address, category, website, phone, opening hours, reviews and images.<\/p>\n\n\n\n
Independent research points in the same direction.<\/p>\n\n\n\n
GPTSpy<\/a> has documented ChatGPT entity<\/code> references containing local-business IDs, coordinates and addresses.<\/p>\n\n\n\n
RESONEO<\/a> has shown that ChatGPT local results frequently contain Google Place IDs and Google-compatible ratings, categories, reviews and images.<\/p>\n\n\n\n
Our experiment adds the bridge between ChatGPT\u2019s lbe1<\/code> identifier and the external place identity.<\/p>\n\n\n\n
Connecting lbe1<\/code> to Google Place IDs<\/h2>\n\n\n\n
For Fanningberg, ChatGPT gives us:<\/p>\n\n\n\n
lbe1_N0fAahpRpdylU3ed4w5PuLQQ<\/code><\/pre>\n\n\n\n
The independently resolved Google place is:<\/p>\n\n\n\n
ChIJ9c_cfw7ccEcRRc3dZpNRT4w<\/code><\/pre>\n\n\n\n
Both records match on name, address, phone, website, rating and review count.<\/p>\n\n\n\n
We repeated the comparison across multiple places:<\/p>\n\n\n\n
Entity<\/th>ChatGPT entity_lookup_id<\/code><\/th>Google Place ID<\/th>Google CID<\/th><\/tr><\/thead>
Fanningberg<\/td>lbe1_N0fAahpRpdylU3ed4w5PuLQQ<\/code><\/td>ChIJ9c_cfw7ccEcRRc3dZpNRT4w<\/code><\/td>10110389381997907269<\/code><\/td><\/tr>
Gro\u00dfeck\u2013Speiereck<\/td>lbe1_KpAdgJIp2gFh01ouRBpJ3XqR<\/code><\/td>ChIJLYQiDpjccEcRR4XhFPCesh0<\/code><\/td>2139947526934070599<\/code><\/td><\/tr>
Obertauern<\/td>lbe1_QJOd0CsvlGXSwcFbxel9KlKF<\/code><\/td>ChIJw4rPjr3XcEcRrC57dz_Epjg<\/code><\/td>4082165889124478636<\/code><\/td><\/tr>
AlpiNest Feriendorf Lungau<\/td>lbe1_R-wKU8fgyRChtCD4m9iN8QpD<\/code><\/td>ChIJZXvYPP7dcEcRqCiPOdTEjAk<\/code><\/td>688141258858637480<\/code><\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
lbe1<\/code> is not<\/strong> an encoded Google Place ID or CID. We found no reversible transformation between them.<\/p>\n\n\n\n
A better model is:<\/p>\n\n\n\n
Google Place identity\nCID \/ Place ID\n        \u2193\nentity reconciliation\n        \u2193\nChatGPT local-business object\n        \u2193\nlbe1_...<\/code><\/pre>\n\n\n\n
Different identifier systems can refer to the same real-world entity<\/strong>.<\/p>\n\n\n\n
Google CID and Place ID<\/h2>\n\n\n\n
For AlpiNest:<\/p>\n\n\n\n
Place ID\nChIJZXvYPP7dcEcRqCiPOdTEjAk\n\nCID\n688141258858637480<\/code><\/pre>\n\n\n\n
These are two representations of the same Google place identity.<\/p>\n\n\n\n
RESONEO provides a useful technical explanation in its research on Google Maps, Geostore, Place IDs and CIDs<\/a>.<\/p>\n\n\n\n
This matters because publishers can reference the CID directly in structured data.<\/p>\n\n\n\n
schema:hasMap<\/code> as an identity assertion<\/h2>\n\n\n\n
On alpina.travel<\/strong>, AlpiNest Feriendorf Lungau is represented as a LodgingBusiness<\/code>:<\/p>\n\n\n\n
{\n  \"@type\": \"LodgingBusiness\",\n  \"@id\": \"https:\/\/data.wordlift.io\/wl01855\/alpinest-feriendorf-lungau\",\n  \"name\": \"AlpiNest Feriendorf Lungau\",\n  \"hasMap\": \"https:\/\/www.google.com\/maps?cid=688141258858637480\"\n}<\/code><\/pre>\n\n\n\n
That gives us:<\/p>\n\n\n\n
Alpina Knowledge Graph\n        \u2193\nschema:hasMap\n        \u2193\nGoogle CID\n688141258858637480\n        \u2195\nGoogle Place ID\nChIJZXvYPP7dcEcRqCiPOdTEjAk\n        \u2195\nChatGPT reconciliation\n        \u2193\nlbe1_R-wKU8fgyRChtCD4m9iN8QpD<\/code><\/pre>\n\n\n\n
The publisher does not need to know ChatGPT\u2019s private lbe1<\/code>.<\/p>\n\n\n\n
It needs to make the real-world identity unambiguous.<\/p>\n\n\n\n
\"How
How Alpina’s Ontology helps ChatGPT resolve real-world entities<\/strong><\/figcaption><\/figure>\n\n\n\n
What changed with Samspitze 4<\/h2>\n\n\n\n
We have been tracking Samspitze 4<\/a><\/strong>, an apartment inside AlpiNest Feriendorf Lungau<\/strong>.<\/p>\n\n\n\n
Earlier in the experiment, we observed:<\/p>\n\n\n\n
Samspitze 4\n\u2193\nlbe1_hwcHov-PU9_ML4mChU2YPuVB<\/code><\/pre>\n\n\n\n
We then strengthened the Alpina entity graph with:<\/p>\n\n\n\n