Which Data Sources Should You Care About For AI Search?

TierTypical useSupplyProof standingWhat the proof saysReference1Internet & search discoveryGoogle SearchConfirmed + presentGrounding with Google Search connects Gemini to real-time net content material, returning inline citations to supply URLs.Google — Gemini API docs, Grounding with Google Search1Bing SearchConfirmed + presentMicrosoft paperwork Bing outcomes getting used to reinforce Copilot responses. Not re-verified on this go.Microsoft Bing3Widespread CrawlConfirmed historicGPT-3 used filtered Widespread Crawl as roughly 60% of its sampling combination; LLaMA 1 reported 67%.Common Crawl3Historic net corpora (C4 and so forth.)Confirmed historicC4 is a cleaned by-product of Widespread Crawl; LLaMA reported C4 at 15% of its pretraining combination.TensorFlow Datasets — C44Internet grounding providersRobust proof / probablyClass inference masking third-party grounding/retrieval intermediaries. No single canonical supply.—1Merchandise & purchasingGoogle Service provider MiddleConfirmed + presentService provider feed information underpins Google’s purchasing surfaces. Retained on person instruction; Google Procuring eliminated as it’s a floor, not a supply.Google Merchant Center Help2Service provider / retail feeds (OpenAI)Confirmed + presentRetailers share a safe, often refreshed CSV/JSON feed of identifiers, descriptions, pricing, stock, media and fulfilment so ChatGPT can floor merchandise precisely. Refreshes accepted as typically as each quarter-hour.OpenAI Developers — Agentic Commerce, product feeds4Microsoft Service provider MiddleRobust proof / probablyEqual business feed infrastructure; inferred parallel to Google Service provider Middle relatively than individually evidenced.—4Market feedsRobust proof / probablyClass inference. Shopify catalog information is already built-in into ChatGPT, which helps the sample.OpenAI Help — Shopping with ChatGPT Search1Native & locationsGoogle MapsConfirmed + presentGrounding with Google Maps is a documented software alongside Search grounding, giving fashions geospatial context.Google Cloud — Grounding API1Google Enterprise ProfileConfirmed + presentEnterprise profile information feeds Google’s native surfaces. Carried from the supply desk; not individually re-verified.—1YelpConfirmed + presentYelp licenses critiques, pictures and enterprise info to OpenAI for real-time native suggestions. Past grounding it additionally drives actions: ChatGPT customers can ebook a desk or be a part of a waitlist, and Request a Quote lets customers contact suppliers in-chat. Yelp’s 10-Q confirms it’s stay.Axios; Yelp blog; Yelp 10-Q FY20264OpenStreetMapRobust proof / probablyBroadly used open geospatial corpus; inferred relatively than confirmed for any named mannequin.—4FoursquareRobust proof / probablyLeft in Tier 4 intentionally: the OpenAI deal is Yelp’s, and no equal proof exists for Foursquare.—4TripadvisorRobust proof / probablyClass inference for evaluation/journey information. No confirmed deal recognized on this go.—1Information & referenceWikipediaConfirmed + presentExplicitly current in GPT-3’s disclosed combination and LLaMA (June-Aug 2022 dumps, 20 languages); additionally broadly used as a stay reference/RAG corpus.Wikimedia dumps1WikimediaConfirmed + presentIdentical corpus household as Wikipedia. Licensing is unusually clear: principally CC BY-SA with attribution/share-alike obligations.Wikimedia dumps4WikidataRobust proof / probablyStructured entity layer; strongly implied by knowledge-graph use however not individually confirmed.—1Neighborhood / Q&A / socialRedditConfirmed + presentThe Google deal gave entry to the Reddit Information API for ‘real-time, structured, distinctive content material’, and permits Reddit content material to be displayed throughout Google merchandise — i.e. stay grounding, not solely coaching.Tom’s Guide (Google/Reddit deal)2RedditConfirmed + presentIdentical deal, coaching aspect: Google might use Reddit posts to coach its AI fashions and enhance providers similar to Search; reported at roughly $60m/yr. NOTE: Reddit is reportedly weighing whether or not to resume — deal with as unstable.Fortune; Neowin/WSJ on renewal doubt4Social platformsRobust proof / probablyClass inference masking platform-wide social corpora.—4Boards / communitiesRobust proof / probablyClass inference. Overlaps Reddit however generalised to non-Reddit boards.—1Information & writer content materialStay writer pagesConfirmed + presentReached at inference time through search grounding relatively than pretraining; retrieval choice and crawlability govern inclusion.Google — Grounding with Google Search2Licensed writer content materialConfirmed + presentOpenAI has a number of express licensing partnerships (FT, Axel Springer, AP, Information Corp). Phrases differ per accomplice on coaching vs grounding vs attribution.OpenAI — FT content partnership2Writer partnershipsConfirmed + presentAxel Springer’s deal contains in any other case paywalled materials in solutions; AP licensed a part of its textual content archive.OpenAI — Axel Springer partnership3Historic information corporaConfirmed historicArchive materials absorbed in pretraining; distinct from stay licensed entry.—2Developer / technicalGitHubConfirmed + presentLLaMA used GitHub’s public BigQuery dataset, restricted to Apache/BSD/MIT tasks; The Pile individually contains GitHub. Public visibility will not be an open licence.GitHub2Stack OverflowConfirmed + presentNamed in licensing-deal mapping alongside Reddit and Shutterstock as a knowledge platform powering a number of consumers.LLM Pulse — AI content licensing deals mapped2Technical docsConfirmed + presentVendor documentation corpora; broadly used however not tied to a single disclosed settlement.—4npm / PyPI registriesRobust proof / probablyWEAKEST ENTRY IN THE TABLE. Relabelled from ‘package deal registries’ to call examples. No disclosed settlement or documented retrieval use discovered — think about reducing.—1Journey & commerce actionsGoogle Resort Middle feedsConfirmed + presentWhen Gemini or AI Mode present resort choices with real-time costs, that information comes from the Google Resorts feed. In Aug 2026 Google added resort reserving inside AI Mode accomplished with Google Pay, so that is grounding plus actions.TechCrunch — AI Mode travel update4Reserving / accomplice feedsRobust proof / probablyReserving Holdings and IHG are reported as members in Google’s agentic reserving pilot, which helps the path however stops wanting a documented feed spec.InfosTourisme (IHG/Booking pilot)4OTA / commerce sourcesRobust proof / probablyClass inference. OTAs run their very own charge feeds into these surfaces.—4Reservation / stock APIsRobust proof / probablyClass inference masking reserving/stock endpoints uncovered to brokers.—