Why publishing a page doesn’t mean AI has read it

We followed business information through a published AI browsing tool to see where details can get lost.

Research notes

By OptiviewFor business owners, marketers and agenciesAbout 6 minPublished October 29, 2026Sources checked October 29, 2026Updated October 31, 2026

Download guide (.md)

OPTIVIEW FIELD NOTESQuestions before conclusionsA / CONTROLB / COMPARISONDefine the question. Retain the observation.Study design illustration — no results shown.

Editorial illustration · Check AI answers

The takeaway

Check what an AI tool actually receives. A published page, a successful download and an AI answer are different pieces of evidence.

A few terms in plain English

Markdown A simple text format that keeps headings, lists and links.

HTML The code used to describe a webpage.

retrieval When a tool fetches information from a page or another source.

Narrative Graph Optiview’s name for the business information your team organizes, reviews and approves.

What the published browsing code does

As part of its gpt-oss project, OpenAI published a simple browsing tool. It gives a model actions for searching, opening pages and finding text within them. It also includes connections to Exa and You.com, services that can supply search results and page content.

This is a working open-source example, not a disclosure of the browsing system behind production ChatGPT. Our tests use a fixed version of that example so the processing can be reproduced. They do not prove how ChatGPT, Claude, Gemini or Perplexity browse.

We can inspect what the tool actually does instead of guessing. Its browser actions, content processing and Exa and You.com connections are available in its published code.

How information moves from a webpage to the model

Think of it as a research assistant working through text, not someone looking at the finished website design.

  1. Search for information. The tool sends a query to a search provider. In this version, the Exa search listing uses provider-generated summaries; the You.com listing uses returned snippets. A search result is not the same as reading the whole page.
  2. Retrieve a page. When asked to open a URL, the tool requests content from the provider. That returned content is the starting point for the next step. The code alone does not tell us how recently the provider obtained it or which website representation it used.
  3. Turn it into readable text. The processing code converts returned HTML into text with structure and link references. The model does not need the website’s visual design to read that representation.
  4. Read a portion and explore further. The default display budget in this implementation is 1,024 tokens—pieces of text, not a fixed number of words. The model can request another part, find a phrase or follow a link.
  5. Use the information in an answer. The browsing tool makes information available. It does not guarantee that the model will consider it relevant, use it or cite it.

That last distinction matters. A fact can exist somewhere in the retrieved text without appearing in the first portion shown to the model. And a page can appear in search results without ever being opened.

Where Optiview can help

Optiview works on the information a website makes available. People see the normal website. AI tools that support it can get a cleaner, easier-to-read version of the same website content, along with any extra information the business has approved.

The automatic conversion does not require a second website to maintain. The optional Narrative Graph lets a business add explanations: what it offers, what makes it different, who it serves and which limitations or responsibilities customers should understand.

For example, “we introduce customers to independent contractors” means something different from “we perform repairs.” Publishing that distinction clearly gives a system information it can potentially use to avoid a misleading description.

Optiview can also record what it served to an observed request. That verifies a delivery. It does not, by itself, verify what an intermediary passed on or which information ultimately shaped an answer.

We use OpenAI’s code as a research reference. It is not Optiview’s production delivery engine, and using it does not imply an OpenAI partnership, endorsement or certification.

First, we checked whether the information survived processing

We took captured content from two websites and supplied it to the published text-processing code. When the supplied content included our added business context, all 13 supporting passages we had selected in advance remained in the full processed text.

None of those passages appeared in the initial viewing window. They were still accessible later. This showed why “present in the full text” and “immediately visible to the model” are different observations. We did not test whether a model would choose to navigate to those passages.

We supplied the content ourselves for this local test. That tells us what survived processing, not whether a search provider would retrieve the same version. See the controlled study and its limitations.

Next, we checked what live retrieval services returned

We tested the pinned browser’s live Exa and You.com paths with 12 fixed requests across those two websites: four direct page requests and eight searches that did not name the target businesses. All 12 requests succeeded without retries.

All four direct requests returned page content. But none of those four responses included a complete added brand passage we were looking for. Separate checks confirmed that the current prepared website versions contained the published context.

Some related information was present in ordinary page content. We therefore do not interpret the result as “the providers understood nothing” or proof that all business meaning was lost. We also cannot tell whether caching, selection of a different version, provider extraction or another difference explains the absent complete passages.

Of the eight searches, one returned a target website among the first 20 results: Exa returned Learnings.org at position 13. That result was below the browser’s initial viewing window but accessible later. This was a search-result appearance, not an assistant recommendation or citation.

These two tests measure different steps. Their counts should not be treated as a before-and-after performance comparison. Inspect the live-provider findings and records.

The takeaway: publishing content is not the same as retrieving it

The prepared context survived this particular transformation when we supplied it. The four live provider responses did not supply complete tested passages to that transformation in the first place. We can locate that observable gap without claiming to know its cause.

This is why a visit from an AI-related bot is useful evidence of activity, but not proof of a citation. It is also why converting a website to Markdown cannot, on its own, guarantee inclusion in an answer. A cleaner format may make information easier to process; it cannot make another service retrieve or use it.

For Optiview, the next responsibility is clear: check what we publish, test what supported retrieval services return, and compare those observations separately with what assistants say. These findings do not establish better citation performance or justify replacing the parser or Narrative Graph.

What we still need to learn

We have observed real AI answers using published business language. Those examples are interesting, but they do not by themselves establish that Optiview caused a recommendation or improved an answer.

We also ran a separate 72-answer comparison of different source conditions. Problems with automated scoring prevent us from treating it as reliable evidence that one condition produced better answers. We retain that study and its limitations rather than turn its scores into an improvement claim.

The bigger question remains: when approved business context demonstrably reaches a model, does it improve the answers people receive? A future controlled study could examine that. This article reports the completed work, not a new experiment or a promise of a particular result.

Businesses need more than facts scattered across the web. They need those facts and their meaning to remain clear as other systems handle them. That is the problem we are exploring with Optiview.

Explore real AI conversations, or read the complete research methods and results.

More in this collection

Go to the source

Keep exploring.

References linked in this guide. Open the original source for its full scope and latest details.

github.combrowser actionsOpens in a new tab github.comcontent processingOpens in a new tab github.comExa and You.com connectionsOpens in a new tab

Keep going

A useful next read.

Research design

How to run a useful test of what AI reads from your website

https://optiview.ai/resources/ai-retrieval-experiment-design-template/

Evaluation guide

Is that AI answer about your business actually correct?

https://optiview.ai/resources/ai-answer-accuracy-evaluation-rubric/

Research design

Does simpler website text lead to better AI answers? A test plan.

https://optiview.ai/resources/markdown-vs-html-controlled-study-protocol/

Back to the resource library

Additional source context

Your website was built for people. Optiview makes it ready for AI, too.

Try Optiview

See my site with Optiview · Run your free AI Audit

Explore

Publish · Parse · Retrieve · Pricing

Learn

Guides · Developer docs · Research & thinking · Experiments & methods · AI Audit · AI agent directory · FAQ

Company

About Optiview · Agencies · Contact · Privacy · Billing terms

Additional page links


Published brand narrative

Site-wide brand information

Who we are

Optiview — Parse, Retrieve and Publish for AI

Optiview helps businesses and developers make website information easier for AI to use. Parse converts supplied HTML into clean Markdown and source metadata. Retrieve fetches an eligible public HTTPS webpage and prepares its content. Publish helps a verified business prepare its own website content and make approved Narrative Graph context available through a supported connection.

Parse and Retrieve use application keys and a prepaid developer wallet. Publish has separate website setup, ownership checks and a subscription. Optiview does not control whether an outside AI service discovers, uses or cites the resulting content.

Applies to: Optiview

Why we matter

A clearer source, from website content to business context

Webpages mix useful information with navigation, layout and application code. Businesses may also have important qualifications, answers and context that are not clearly published.

Optiview gives developers a way to prepare source content for their applications, and gives businesses a way to publish the context their teams approve. Your application chooses how to use Parse and Retrieve results. Your business controls what it publishes through Publish.

Clearer publishing makes information available; it does not determine an AI system’s answer. Each outside service chooses its own retrieval, recommendations and citations.

What we offer

Three products, with separate jobs and clear pricing

Developer credits do not purchase Publish subscriptions. Parse and Retrieve do not add a customer’s private Narrative Graph to arbitrary third-party pages. See https://optiview.ai/pricing/ and https://optiview.ai/developers/ for the current offer and limits.

Why trust us

Inspect the output and understand what the evidence establishes

Developers can inspect the prepared Markdown, job status, source metadata, processing warnings and credit ledger. Retrieve records the public source response it observed; Parse identifies supplied HTML rather than claiming to verify the current origin.

Publish administrators can review parser output, the combined published-output preview, source checks and publication history. A preview is separate from a recorded live delivery. Saved edits remain separate from the active publication until explicitly published.

Our experiment pages contain selected first-party AI conversations and source comparisons. Delivery records establish what was served, not downstream ingestion, recommendations, citations or revenue. Observed language matches do not establish causation or a consistent visibility improvement. Explore the evidence and limitations at https://optiview.ai/proof/.

Helpful knowledge

Common questions about the three products

Which product should I use? Use Parse when you already have HTML. Use Retrieve when you need Optiview to fetch one eligible public page first. Use Publish when your business wants to prepare its verified website and publish approved context.

Do Parse and Retrieve require website ownership or DNS changes? No website connection is needed for these developer products. You remain responsible for source rights and permitted downstream use. Public availability does not grant a reuse license.

How do developer accounts and keys work? Create a developer workspace and save the personal sign-in key and offline recovery code. Create scoped service keys for server-side applications. Email is an unverified contact identifier, not account-ownership proof or an email-only recovery method. Keep all credentials private.

Can my agent use Optiview? The authenticated developer API returns prepared content for your application or model. A downloadable JavaScript client and local MCP adapter are available. Your host must authorize sources and any credit expenditure. Compatibility does not establish adoption or a partnership with an AI provider.

Does the API include my private Narrative Graph? Parse and Retrieve are source-only. Verified website publishing and approved publisher context belong to the separate Publish workflow.

Does Publish replace our website or CMS? No. Your website stays in its existing system. Your web team establishes and tests a supported connection. Your team reviews and explicitly publishes any Narrative Graph additions. Ordinary visitors retain the normal website experience.

Does every AI tool get the prepared version? No. Delivery depends on the website connection, source eligibility and the request. Some tools explicitly request Markdown; supported AI requests may also receive prepared Markdown or HTML through eligible delivery paths. Traditional search crawlers retain the original website response. A claimed tool name does not prove the requester’s identity.

Does the demo publish or save to our workspace? No. The public demo is a separate preview. Its AI-generated examples require review and do not transfer automatically into an approved publication.

Will Optiview guarantee citations or recommendations? No. Optiview prepares and publishes sources. Outside AI services choose what to retrieve, use and cite.

What to do next

Choose your starting point

Already using Publish? Review your Narrative Graph, publish approved changes explicitly, inspect the combined source in Markdown and check recorded delivery in Activity. Keep your existing visibility tools to study downstream AI answers.