Wildwood AI

Customer-safe methodology

How Wildwood turns provider evidence into a practical audit.

Wildwood AI records what each supported check attempted, where and when it ran, whether trustworthy evidence returned, and how that evidence supports a finding. Observed, not observed, unavailable, and not applicable are distinct states. Recommendations are withheld when the required evidence is missing, and raw provider payloads remain restricted to server-side storage.

Reviewed by James Goldstein ·

Start my free SEO audit
Evidence pathCustomer-safe method
  1. 1. Run contextQuery, page, place, device, language, depth, and time
  2. 2. Observation stateObserved, not observed, unavailable, or not applicable
  3. 3. Supported findingScope and evidence shown before the recommendation
The audit keeps failed evidence from becoming a false zero, absence, or recommendation.

Observation states

A missing answer is not automatically a zero.

Observed

A successful check returned trustworthy evidence for the stated query, page, or entity.

Not observed

A successful check completed, but the target result was absent within the disclosed sample and depth.

Unavailable

The check failed, returned malformed data, or lacked enough trustworthy evidence. It is excluded from scores and ordinary recommendations.

Not applicable

The check did not fit the business, website, or purchased section and was intentionally skipped without lowering quality.

Run context

Every conclusion needs a scope.

Customer-safe methodology records can include the query source, capture time, location, country, device, language, search depth, comparison period, and provider coverage. A locally sampled result stays separate from nationwide discovered-keyword data.

Search positions are preserved only when the provider supplies a real rank. Array order is never substituted for a Google position, and failed keyword requests do not become zero demand.

Website evidence

Page claims require a sufficiently complete page.

Wildwood separates DNS, HTTP, redirect, TLS, PageSpeed, crawl, and document evidence. Metadata, headings, canonical tags, robots directives, structured data, and copy findings are called missing only when a sufficiently complete document was received.

When page acquisition fails, Wildwood withholds unsupported copy criticism. Independent performance or search evidence can still be shown with its own scope.

AI evidence

Sampled answers and aggregate mentions do not share one denominator.

Wildwood records exact-citation, positive-text, negated-text, and no-mention evidence separately. Sampled answer citations stay separate from provider aggregate mention metrics so a broad metric cannot contradict the answer excerpts the customer sees.

AI checks name attempted and successful coverage. A platform failure is unavailable, not a zero, and answers are sanitized before customer display.

Freshness and reproducibility

The report keeps the versions needed to understand its result.

  • New reports snapshot the resolved report layout, labels, administrator content, revision, and hash at audit start.
  • Server-side evidence manifests record request scope, timestamps, parser versions, prompt/model versions when applicable, and provider coverage.
  • Customer pages show safe methodology and evidence descriptors; raw provider responses and diagnostics are not directly browser-readable.
  • A later audit is a new dated observation. Historical conclusions are not silently rewritten when providers, models, or templates change.

Privacy and limits

Evidence is collected for the requested audit—not to make claims beyond it.

Wildwood limits customer-facing data to the information needed to understand findings and recommendations. It validates outward links, removes raw provider errors and internal identifiers, and does not expose server-only evidence objects to unauthenticated browsers.

An audit samples supported data sources and cannot represent every search, user, location, crawl state, or AI response. Rankings, traffic, sales, and inbox delivery are never guaranteed.

Privacy policyRead how Wildwood handles information.Public self-auditSee these standards applied to Wildwood’s own baseline.

Primary guidance

The public sources used for search and AI eligibility principles.

Google: JavaScript SEO basicsGuidance for crawlable content, links, metadata, rendering, and meaningful HTTP status codes.Google: AI featuresGuidance for making indexed web content eligible for Google’s AI search experiences.OpenAI: publisher controlsGuidance for OAI-SearchBot discovery and ChatGPT search referrals.