Skip to content
OUHUDSearch Intelligence

How OUHUD assesses AI visibility

OUHUD combines website checks, content work supported by evidence and samples of AI answers. This page explains which evidence matters and what it can tell you.

The method applies to customer websites and our own website. Data available for a particular website depends on connected sources, the agreed scope and successful measurements.

Three pillars, each with supporting evidence

1. Make pages discoverable and accessible

Check
Check robots rules, the HTTP response, readable page text, internal links and language references together. Permission in robots.txt is insufficient if a firewall blocks access.
Suitable evidence
The specific URL, crawler tested, time and content actually received. For multilingual sites, check each language URL separately.
Limit of the finding
A successful fetch proves neither indexing nor a future citation.

2. Publish original knowledge with sources

Check
Explain services, limitations and working methods using approved company sources. Identify responsible people and sources; improve existing useful pages with specific changes.
Suitable evidence
A specific statement with a source, date and subject matter review. Real reference projects require verifiable records and permission to publish.
Limit of the finding
New text, structured data or more articles alone do not demonstrate increased AI visibility.

3. Measure outcomes separately

Check
Repeat a defined set of questions for each language and AI system. Evaluate linked pages, brand mentions and website visits separately.
Suitable evidence
Question, language, system or model, time, response status and source URL. For comparisons, keep the scope, period and failed requests visible.
Limit of the finding
An API sample does not represent every answer in the corresponding consumer app. Answers can change even when your website has not changed.

What each signal tells you

A crawler fetch, an AI source and a customer inquiry are different events. We treat them as separate evidence.

SignalWhat it demonstratesWhat it does not demonstrate
Cited pageThe answer names a specific URL as a source.It does not show whether anyone opened that link.
Brand mentionThe company name appears in the answer examined.It does not establish a linked source or a recommendation.
VisitConnected web analytics records a session under its own measurement rules.Organic search traffic is not automatically AI traffic. Missing referrers limit attribution.
InquiryAn event configured for this purpose, such as a submitted contact form, was recorded.Not every key event is an inquiry. An inquiry does not prove causation by one particular change.

How to make a comparison traceable

Compare the same questions, languages and systems over named periods. Record website changes and measurement failures. Without a successful measurement, the result is ‘no data’; zero means that a measurement was completed and nothing was found. More visits after a change are an observation, not yet evidence that the change caused them.

Check OUHUD itself

Our product descriptions are statements by the provider. These public documents help you assess them; they do not replace an independent customer reference.

Primary sources and their scope

These documents explain technical requirements and measurement terms. They do not establish an OUHUD success rate or an effect for a particular customer.

Sources checked:

Start with a specific page

Bring an important service page, its supporting company sources and your customers’ questions. Together, these provide a starting point you can verify.