Future InTech

RESEARCH NOTE · Real saved observations · Scope and limitations below

FIELD NOTES / SAVED EVIDENCE

AI visibility audit: what website and traffic evidence can establish

A real evidence inventory from Future InTech shows why website checks, crawler requests, referral visits and answer citations need separate reporting.

Observed window
2026-11-12T17:00:13+00:00 → 2026-11-19T17:00:13+00:00 · UTC
Report built
2026-11-19T17:02:16.071668+00:00

The observation

This evidence can support a website review and an access investigation. It cannot establish whether ChatGPT, Perplexity, Gemini or Claude mentioned or cited a brand. A useful AI visibility audit makes that boundary visible before it recommends a change.

Website evidence reports

350reportsSaved public copies; not platform answer samples.

Observed log requests

1134requestsOwn-site collection only; may include self-tests.

Browser pageviews

0eventsA separate population from log requests.

Original AI answers

Not measuredNo answer dataset or sampling plan in this research source.
01

Read the evidence

01

Website evidence

What did the saved page return?

Available in the selected archive
02

Crawler access

What did the server observe?

Mixed operational traffic
03

Referral traffic

Which recorded source sent a visit?

Source attribution incomplete
04

AI answers

Did an answer mention or cite the brand?

Not measured

ChatGPT

Not measured

No original platform answer dataset in these sources.

Perplexity

Not measured

No original platform answer dataset in these sources.

Gemini

Not measured

No original platform answer dataset in these sources.

Claude

Not measured

No original platform answer dataset in these sources.

Linked evidence · 2 references
02

What to do next

Investigation priorities reflect this evidence and its limits. They are not promises of traffic or AI citations.

P2

Define the answer sample before calculating visibility

Evidence

The sources here contain no original answers or complete sampling manifest.

A handpicked successful answer cannot establish a complete-group appearance rate.

  1. Record the questions, branded/nonbranded group, platform/product/mode, disclosed model, sampling time, language and region basis.
  2. Keep expected run identifiers and failures. Import only legally obtained records with processing permission.
  3. Evaluate brand mentions and verified original citations separately; retain incomplete citations as unknown.

Owner Research owner

Acceptance check Every ratio has an explicit valid denominator and coverage. Imported API records are not labeled native-product samples.

P2

Keep remediation tied to the observed layer

Evidence

Use the linked GEO audit findings for website changes and the monitoring report for access observations.

Fixing a page or responding to a bot does not demonstrate an answer citation.

  1. Use the technical finding URL and original evidence location to hand off a fix.
  2. Recheck the same technical condition after implementation.
  3. Use a separately defined answer sample to observe later visibility, without claiming the fix caused any change.

Owner SEO owner / developer

Acceptance check The handoff distinguishes technical acceptance from answer-sample observations and from business outcomes.

03

Scope, source & handoff

Source verification and dates

A hash detects a change to the saved artifact. It does not authenticate a claim, origin or author.

monitoring evidence · 2026-11-19T17:00:13+00:00
d62a44978c17cd7b1f97e57891bbe7f01a59cd7ecb92b24a5b7593bc6deb6bcb

audits evidence · 2026-11-15T16:50:05.151546+00:00
84a93572b505442f492e47e6a6e72e6c716bc34baa8294165d9e70ff4fa2dc7b

Combined source/configuration hash
2ca21d59b13db113dd2cae59ca0b990fc77e6eec5e11f7cd4f743260e9d3537a