AI Site Grade

potamkinautomotive.com — AI Site Grade

Potamkin Automotive's robots.txt blocks all major AI crawlers, but Squarespace serves them full content anyway, while the site has only 6 pages, no llms.txt, empty JSON-LD, and zero external web footprint.

Potamkin Automotive's AI visibility is severely limited by a contradictory robots.txt, minimal content, empty schema, and no external signals, despite Squarespace inadvertently allowing crawler access.

Findings
11
Evidence checks
25
Completed
30 May 2026

Analysis

I have enough to write the audit now. Let me compile the findings.

The most striking finding: The site explicitly blocks all major AI crawlers in robots.txt (GPTBot, ClaudeBot, Google-Extended, etc.), yet Squarespace serves them all full content anyway (200, 71KB) — the robots.txt is purely advisory and Squarespace doesn't enforce it. Meanwhile, the site has only 6 pages, no llms.txt, empty JSON-LD, and the cold LLM knowledge about Potamkin is factually wrong (wrong founding year, wrong location, wrong brands).

Potamkin Automotive — AI-Visibility Audit

The site's robots.txt explicitly blocks every major AI crawler (GPTBot, ClaudeBot, Google-Extended, anthropic-ai, Applebot-Extended, PerplexityBot, and 20 more), yet Squarespace serves all of them the same full 71KB HTML as a browser — the robots.txt is purely advisory and unenforced, meaning AI crawlers are technically welcome to index but told not to.

Crawler Access

All 11 tested bot UAs return HTTP 200 with identical 71KB content from Squarespace's edge, including GPTBot, ClaudeBot, Google-Extended, OAI-SearchBot, and PerplexityBot. The sole exception is Bytespider, which gets a 403 from Squarespace's infrastructure directly. The /llms.txt returns a 404 (Squarespace HTML error page). The sitemap.xml lists exactly 6 URLs: homepage, /home (a duplicate of the homepage), /history, /team, /brands, and two privacy pages. No dealership locations, no inventory, no service pages.

Cold-Knowledge Gap

A frontier LLM queried cold describes Potamkin as a Miami-based dealership group selling Mercedes-Benz, BMW, Lexus, Toyota, and Honda, founded in 1946. The actual site says the company was founded in 1954 by Victor Potamkin in New York, grew to become the largest U.S. auto dealer by 1987 with 41 franchises selling 76,000 cars, and has operated over 100 franchises across luxury and mainstream brands including Ferrari, Lamborghini, Pagani, McLaren, Porsche, and Cadillac. The model's prior is wrong on founding year (1946 vs 1954), wrong on geography (Miami vs New York origins), and omits the luxury/exotic portfolio entirely. The site itself provides no current dealership addresses, no inventory, no store locator — only a historical brand list.

Schema Posture

Every page carries the same two JSON-LD blocks: a WebSite and a LocalBusiness — both with empty description fields, empty address, and empty openingHours. The LocalBusiness type is used without specifying any sub-type (e.g. AutoDealer), without location data, without telephone, without sameAs links. The homepage has only 140 words of visible text. The /team page lists 18 staff with email addresses but no titles for several roles and no structured Person schema.

External Signals

The site links to five sibling domains — potamkindev.com, potamkinventures.com, potamkin.org, potamkinprize.org, and levelnyc.com — but none of these are dealership microsites. The history page hosts three PDFs (a 1987 AutoAge article, a Dealer Magazine article, and a 1954 Wall Street Journal piece) as evidence of press coverage. DuckDuckGo returns zero indexed results for the domain or brand queries — the site has essentially no external web footprint. The Wayback Machine shows the site has existed since at least January 2025, suggesting it is a recent Squarespace build that replaced an older web presence.

Findings

  1. Robots.txt blocks all major AI crawlers High

    The robots.txt explicitly disallows GPTBot, ClaudeBot, Google-Extended, and 20+ other AI crawlers, but Squarespace serves them all full 71KB HTML anyway, making the block purely advisory and unenforced.

    What to change: Remove AI crawler disallow rules from robots.txt to align with actual access, or implement server-side blocking if crawlers are truly unwanted.

  2. Bytespider receives 403 from Squarespace Medium

    Bytespider, the crawler for ByteDance's AI, gets a 403 Forbidden from Squarespace's infrastructure, while all other bots get 200. This inconsistency may limit visibility to certain AI platforms.

    What to change: Investigate why Bytespider is blocked and allow access if AI visibility is desired.

  3. No /llms.txt file Medium

    The site returns a 404 for /llms.txt, missing an opportunity to provide structured guidance to AI crawlers about which pages to index and how to interpret content.

    What to change: Create an /llms.txt file listing key pages and providing context for AI crawlers.

  4. Sitemap contains only 6 URLs High

    The sitemap.xml lists only 6 URLs: homepage, duplicate /home, history, team, brands, and two privacy pages. No dealership locations, inventory, or service pages are included, severely limiting discoverability.

    What to change: Expand the sitemap to include all relevant pages such as dealership locations, inventory listings, and service pages.

  5. JSON-LD has empty description fields High

    Every page carries WebSite and LocalBusiness JSON-LD blocks with empty description, address, and openingHours fields. The LocalBusiness type is not specialized to AutoDealer, and lacks location data, telephone, and sameAs links.

    What to change: Populate JSON-LD with accurate descriptions, address, telephone, opening hours, and use AutoDealer subtype. Add sameAs links to social profiles.

  6. Cold LLM knowledge about Potamkin is factually incorrect High

    A frontier LLM describes Potamkin as a Miami-based dealer founded in 1946 selling Mercedes-Benz, BMW, Lexus, Toyota, and Honda. The actual site states founded in 1954 in New York, with over 100 franchises including Ferrari, Lamborghini, and Pagani. The model's prior is wrong on founding year, geography, and brand portfolio.

    What to change: Publish accurate, detailed content about the company's history, locations, and brands to correct LLM misconceptions. Consider adding structured data and a Wikipedia page.

  7. Zero indexed results on DuckDuckGo High

    Multiple web searches for the domain and brand returned zero results on DuckDuckGo, indicating the site has no external web footprint. The Wayback Machine shows the site has existed since at least January 2025, suggesting it is a recent build.

    What to change: Build backlinks, submit to directories, and create social media profiles to establish an external web presence.

  8. Homepage has only 140 words of visible text Medium

    The homepage contains only 140 words, providing minimal context for AI crawlers to understand the site's purpose and offerings.

    What to change: Expand homepage content to include a clear value proposition, key services, and links to important pages.

  9. Team page lacks structured Person schema and some titles Medium

    The /team page lists 18 staff with email addresses but several roles have no titles, and no structured Person schema is used, reducing discoverability in AI knowledge graphs.

    What to change: Add Person schema for each team member with name, job title, and email. Ensure all roles have titles.

  10. No dealership location pages or store locator High

    The site provides no current dealership addresses, inventory, or store locator, only a historical brand list. This limits local SEO and AI visibility for physical locations.

    What to change: Create individual pages for each dealership location with address, phone, hours, and inventory. Add LocalBusiness schema with AutoDealer subtype.

  11. Privacy policy page returns 0 words Low

    The /privacy-policy page returns HTTP 200 but contains 0 words of visible text, likely an empty template. This may confuse crawlers and users.

    What to change: Populate the privacy policy page with actual content or remove it if not needed.

What's working

  • Squarespace serves full HTML to all AI crawlers — Despite robots.txt disallows, Squarespace returns full 71KB HTML to all tested AI crawlers (except Bytespider), ensuring content is technically accessible.
  • JSON-LD present on every page — Every page includes WebSite and LocalBusiness JSON-LD blocks, providing a foundation for structured data that can be improved.
  • History page hosts press PDFs as evidence — The history page links to three PDFs (1987 AutoAge article, Dealer Magazine article, 1954 WSJ piece) that provide credible historical press coverage.
  • Brands page lists over 100 franchises — The /brands page lists a comprehensive history of over 100 franchises including luxury and exotic brands, providing rich content for AI crawlers.
  • Team page lists 18 staff with email addresses — The /team page provides names and email addresses for 18 staff members, offering contact information that can be used for relationship building.
  • Site has been live since at least January 2025 — Wayback Machine snapshots show the site has existed since at least January 2025, indicating it is a recent but established presence.
  • Site links to five sibling domains — The site links to potamkindev.com, potamkinventures.com, potamkin.org, potamkinprize.org, and levelnyc.com, creating a small network of related sites.

Track potamkinautomotive.com across AI search

This is one snapshot. Open the interactive report to inspect evidence, or grade another site free.

Open this AI Site Grade Grade another site Track your brand