> ## Documentation Index
> Fetch the complete documentation index at: https://docs.anymorph.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# ChatGPT

> OpenAI assistant with optional web retrieval and parametric answers

## Background: provider and product surface

ChatGPT is OpenAI’s general-purpose AI assistant. It can optionally use web retrieval when current or external information is needed. Public analysis shows strong Bing index dependence for citation eligibility, while some answers remain parametric and citation-free.

## Usage scale and market context

| Metric              | Reported value                             | Context                                                                       |
| ------------------- | ------------------------------------------ | ----------------------------------------------------------------------------- |
| Weekly active users | 800M+ WAU                                  | OpenAI/public reporting describes hundreds of millions of weekly active users |
| Daily prompt volume | 2B+ prompts/day                            | Public/industry reporting cites multi-billion daily prompt volume             |
| Product context     | General assistant + optional web retrieval | Not all usage uses web retrieval or citations                                 |

## Product surface and user experience

* Can show retrieved results, a sources sidebar, and clickable citations.
* May provide vertical answer UI for news, weather, stocks, sports, and maps.
* Follow-up retrieval queries can change with conversation context.
* API/automated results may differ from the consumer UI because of routing, personalization, and browsing state.

## GEO characteristics

| Axis                 | Engine-specific behavior                                                  | GEO interpretation                                                  |
| -------------------- | ------------------------------------------------------------------------- | ------------------------------------------------------------------- |
| Retrieval substrate  | Bing-visible web + OpenAI crawl/retrieval + parametric knowledge          | Bing indexation is a key citation eligibility gate                  |
| Citation UI          | Sources/sidebar when web retrieval is used                                | Retrieval-free answers can still recommend brands without citations |
| Query fan-out        | Follow-up retrieval queries: best/reviews/pricing/alternatives/comparison | Zero-volume traditional keywords can still matter                   |
| Source concentration | ChatGPT source pools often appear relatively concentrated                 | Head sources such as Wikipedia, marketplaces, and listicles matter  |
| Measurement note     | Separate web-retrieved answers from parametric answers                    | Distinguish citation failure from entity visibility failure         |

### Key cited / influential source types

| Source type                   | Observed pattern                                                          | GEO action                                                    |
| ----------------------------- | ------------------------------------------------------------------------- | ------------------------------------------------------------- |
| Wikipedia / knowledge sources | Wikipedia is frequently prominent in ChatGPT source patterns              | Manage notability, Wikidata, and knowledge-graph consistency  |
| Directories / listings        | Directory and third-party listings often matter heavily for ChatGPT       | Manage G2/Capterra/Yelp/TripAdvisor by category               |
| Retail/marketplace pages      | Retail and marketplace domains often show high share in ecommerce prompts | Commerce prompts need marketplace/product-listing visibility  |
| Reddit/LinkedIn/community     | Third-party consensus and real-user narrative                             | Influences comparison, complaint, and recommendation prompts  |
| Bing-visible owned pages      | Bing-based candidate pool                                                 | Use IndexNow, canonical hygiene, sitemaps, and internal links |

## Engine-specific GEO strategy

| Strategy                                | Action                                                                    | Metric                         |
| --------------------------------------- | ------------------------------------------------------------------------- | ------------------------------ |
| Operate Bing indexation separately      | Use Bing Webmaster Tools, IndexNow, canonical, robots, and sitemap checks | Bing indexed URL count         |
| Tag web-retrieved vs parametric answers | Classify retrieval usage before interpreting citations                    | Retrieval-used rate            |
| Build directory/listing presence        | Create and align category-specific review/listing profiles                | Listing citation/mention       |
| Front-load answer passages              | Put comparison, definition, pricing, and alternative details near the top | Citation position/source count |

## What to measure

* Usage scale and surface-specific exposure
* Source/citation presence and source type
* Brand mention, recommendation, sentiment, and claim accuracy
* Engine-specific failure mode: retrieval, citation, or recommendation

## Sources

* [OpenAI announcement: Introducing ChatGPT search](https://openai.com/index/introducing-chatgpt-search/)
* [Erlin AI: ChatGPT optimization analysis](https://www.erlin.ai/blog/chatgpt-search-optimization)
* [ZipTie: How ChatGPT chooses sources](https://ziptie.dev/blog/how-does-chatgpt-choose-its-sources/)