A 13-signal audit of how legible an ITAD company is to ChatGPT, Google AI, Gemini, Claude and Perplexity, run here against a representative fictional provider. The same engine runs on any real domain at aeo.mhndlabs.com.
Enterprise and government buyers increasingly ask an AI assistant before they build a vendor shortlist. These queries have no ranking page to win. The assistant composes an answer and names two to four companies. Either you are in that answer or you are invisible.
Scores below are for the fictional Cornerstone Asset Recovery, drawn to look like a typical 30-year ITAD firm: excellent operations, a website written for humans who already found it. Each signal shows what an AI system actually reads and the specific fix. No signal is scored by opinion; the selftest at the bottom recomputes every number in your browser.
Weights per dimension sum to 100. The profile is representative, not a real audit: your real numbers come from running your own domain through the live engine.
ITAD is unusual: the industry already has independently verifiable trust infrastructure. R2v3 (SERI), e-Stewards (Basel Action Network) and NAID AAA (i-SIGMA) each publish public directories of certified facilities. An AI engine can check those rosters. But on most ITAD sites the certifications exist only as footer badge images, and an engine cannot read a JPEG. The checklist and markup below turn them into signals.
Name each certification, the certifying body, and which facility holds it, in HTML text. R2v3 moved to per-facility scope; engines resolve facilities, not slogans.
One canonical Organization node, referenced site-wide by @id, carrying the credentials with their certifying bodies (template at right).
Confirm the SERI, e-Stewards and i-SIGMA directory listings match your legal name and address, then point sameAs at your profiles engines already trust: LinkedIn, Wikidata if eligible, the directories themselves.
NIST SP 800-88 Rev. 1 was withdrawn; Rev. 2 (September 2025) defers media destruction specifics to IEEE 2883. A surprising number of ITAD sites still cite the withdrawn revision. Citing current standards is a freshness and competence signal an engine can verify.
Secure destruction, decommissioning, and value recovery each get a Service node tied to the Organization, so service-specific queries resolve to you.
{
"@context": "https://schema.org",
"@type": "Organization",
"@id": "https://cornerstone-itad.example/#org",
"name": "Cornerstone Asset Recovery",
"url": "https://cornerstone-itad.example/",
"logo": "https://cornerstone-itad.example/logo.png",
"foundingDate": "1994",
"description": "IT asset disposition and value recovery for enterprise and government clients: secure data destruction, data center decommissioning, certified remarketing.",
"knowsAbout": [
"IT asset disposition",
"data center decommissioning",
"secure data destruction",
"IT asset value recovery"
],
"sameAs": [
"https://www.linkedin.com/company/example",
"https://sustainableelectronics.org/directory-profile-example"
],
"hasCredential": [
{ "@type": "EducationalOccupationalCredential",
"name": "R2v3 Certification",
"recognizedBy": { "@type": "Organization",
"name": "SERI, Sustainable Electronics Recycling International" } },
{ "@type": "EducationalOccupationalCredential",
"name": "e-Stewards Certification",
"recognizedBy": { "@type": "Organization",
"name": "Basel Action Network" } },
{ "@type": "EducationalOccupationalCredential",
"name": "NAID AAA Certification",
"recognizedBy": { "@type": "Organization",
"name": "i-SIGMA" } }
]
}
Engines do not invent recommendations; they assemble them from sources they retrieve and cite. For ITAD buying queries the recurring source classes look like this. An honest program works this map from the top down, because the top rows are verifiable and durable.
Strip away the mystique and there are four levers. Everything in this audit rolls up to one of them.
Each platform runs separate crawlers for training, search and citation. If your robots.txt blocks the citation crawlers (often via a blanket AI block from 2023), you cannot be recommended. This is checkable in one minute and each bot is controllable on its own.
Engines quote what they can parse: a first paragraph that answers the question, headings shaped like the questions buyers ask, clean structured data, certifications as text. Badge images and brochure PDFs are invisible.
Engines read dates. Pages without dateModified, sitemaps a year stale, and citations to withdrawn standards all age you. Current standards and visible update dates keep you quotable.
Who else says you are real: certifier directories, the entity graph, press, communities. For ITAD this is the strongest lever, because the certifiers publish verifiable rosters most competitors never wire up.
Of the major platforms, only Microsoft publishes a checklist for AI visibility; Google says its AI answers are rooted in its core search ranking; OpenAI, Anthropic and Perplexity document crawler access and citation behavior, not a recipe. So anyone promising a guaranteed placement is guessing. What can be promised: every lever above moved to verifiably done, and a measured citation baseline that shows whether recommendations follow. That is the deal I offer.
A panel of real ITAD buying questions is put to ChatGPT, Google AI, Gemini, Claude and Perplexity. For each: who gets named, who gets cited, with screenshots. Plus the 13-signal audit of your site and your top competitors.
Access and legibility fixes ship first (fast, fully in your control), then entity and directory wiring, then the citation-source program, which is slower and compounding.
The identical question panel re-runs monthly. Movement is reported as named-mention and citation counts per engine against the baseline, not as vibes. Refinements follow the data.
This page uses a fictional company because I will not guess at your website. The underlying engine is live and runs the audit on any URL in about a minute. Run your own domain, or send it over and you will have a real baseline within a day.
aeo.mhndlabs.com →Trust in an audit tool starts with the tool being checkable. This button recomputes every score above from the raw signal inputs embedded in this page, validates the schema template, and re-counts the source map, right in your browser.