Stradiji, an SEO and digital marketing firm, has released an AI visibility study and audit, and its headline finding is blunt: most websites still close their doors to large language models. For an industry that has spent two years arguing over whether to welcome or wall off AI crawlers, that is a concrete verdict, and it cuts against the common assumption that publishers have already picked a side.

The study, announced via press release and carried by the Gaston Gazette, examined how websites currently handle access by LLM-driven systems. Its core conclusion: a majority of sites are, deliberately or by neglect, not open to them. That single number — "most" — is the story. It also leaves the most useful questions unanswered.

What does the audit actually measure?

Stradiji frames the work as both a "visibility study" and an "audit," which suggests two layers: measuring how often brands appear in AI-generated answers, and checking the technical settings that determine whether AI systems can read a site at all.

The released announcement does not spell out the sample size, the crawl dates, or which LLM platforms the audit tested. Until Stradiji publishes the underlying data, "most websites" remains a directional claim rather than a precise one. Buyers of visibility audits should ask for exactly those three numbers before acting on any recommendation.

Why does this matter now?

Traffic referrals from AI answer engines are growing, and a site that blocks LLM crawlers will not appear in those answers. The trade-off is real: opening the door means handing content to systems that may reuse it without sending a visitor back. Blocking protects the content but forfeits the placement.

Stradiji's finding suggests most site operators have not resolved that tension — or have not looked at their robot settings at all. An audit that surfaces this gap is doing basic, useful work. Think of it as a locksmith pointing out that half the locks on the street were never keyed to the new doors.

What is not said?

The announcement does not report:

• How many sites the audit covered
• Which companies or sectors performed best on AI visibility
• Whether blocking was deliberate policy or default misconfiguration
• Any pricing for Stradiji's audit or remediation services

Each of those gaps matters. An audit from a firm that sells visibility services naturally benefits from a finding that most sites are failing. That does not make the finding wrong; it makes independent verification the next step.

What should operators do?

The practical takeaway is a checklist item, not a purchase decision. Site owners can review their robots.txt files, identify which AI crawler user-agents they admit or deny, and decide case by case. That review costs an afternoon. Rebuilding lost AI-search placement later costs considerably more.

Stradiji has not stated when it will publish full methodology or results. Until it does, treat this as a well-timed signal about market readiness, not a settled measurement of it.

For now, the door question is open, and the answer, apparently, is that most doors are not.