multipass.co — AI Site Grade

Analysis

Cold-Knowledge Identity Collapse

The LLM prior for multipass.co describes a US-based commuter benefits platform founded in 2019 in New York City, serving HR teams with transit and lifestyle spending accounts. The actual site is a UK-regulated multi-currency business account provider (EMI, FCA 900840) founded in 2017 as an in-house solution for Dyninno Group, headquartered at 87-89 Baker Street, London. The two entities share a domain name and nothing else — the cold-knowledge gap is a complete identity mismatch, not a nuance.

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, ChatGPT-User, OAI-SearchBot, Applebot-Extended — receive 200 responses with full HTML content from Cloudflare. No bot is blocked, rate-limited, or served a JS shell. The robots.txt contains zero AI-specific directives; only a catch-all * rule blocking parameterized URLs, registration paths, and /uae/* subdirectories. The llms.txt returns 404. The sitemap.xml and all four declared sub-sitemaps (blog-articles.xml, resource-articles.xml, help-centre-articles.xml) return HTTP 500 with a Nuxt.js server error page — meaning AI crawlers cannot discover the site's URL inventory through standard means.

Schema Posture

Every page carries the same JSON-LD graph: BankOrCreditUnion, Corporation, and WebSite types with identical address, telephone, and sameAs arrays. The homepage and batch-payment page additionally include a FAQPage schema with 7 and 7 questions respectively. The BankOrCreditUnion type is technically inaccurate — the site explicitly states "MultiPass is not a bank" in its own FAQ — and no FinancialService, PaymentService, or EMI schema type is used. The WebSite SearchAction target URL template (https://multipass.co/{search_term_string}) is malformed, missing a query parameter.

Content & Technical Architecture

The site is a Nuxt.js application served via Cloudflare. The homepage delivers ~1,010 words of visible text from a plain GET, with an H1 of "Business account without borders" and an FX rate calculator widget. Key product pages (global account, batch payments, corporate cards, open banking API) all return 200 with substantive content. The /help-centre page, however, returns HTTP 500 with a Nuxt error shell — a critical support resource is completely inaccessible to crawlers. The /solutions page contains only 55 words and a list of industry links, functioning as a thin navigation hub rather than a content page.

External Signals

The site links to Trustpilot (https://www.trustpilot.com/review/multipass.co) from multiple pages, but web searches for "multipass.co reviews" and "multipass platforms limited" return zero indexed results across Google, Bing, and general web search. The Dyninno Group parent site (dyninno.com) lists MultiPass as a division described as "a bank challenger that provides modern financial solutions for businesses with cross-border activity." No press coverage, Reddit threads, or third-party reviews were discoverable through standard search — the brand has near-zero external citation footprint, which likely explains why the LLM prior defaults to a different company entirely.

Findings

1. LLM prior describes a completely different company High

   The LLM prior for multipass.co describes a US-based commuter benefits platform, while the actual site is a UK-regulated multi-currency business account provider. This identity mismatch means AI assistants will fabricate incorrect information about the company.

   What to change: Publish an llms.txt file with a factual company description, and build external citations (press releases, listings, reviews) to correct the LLM prior.

2. Sitemap and all sub-sitemaps return HTTP 500 High

   The sitemap.xml and all four declared sub-sitemaps (blog-articles.xml, resource-articles.xml, help-centre-articles.xml) return HTTP 500 with a Nuxt.js server error. AI crawlers cannot discover the site's URL inventory through standard means.

   What to change: Fix the server errors on sitemap.xml and all sub-sitemaps to return valid XML listings of all public URLs.

3. llms.txt returns 404 High

   The llms.txt file, which AI crawlers use to discover key resources, returns a 404 error. This prevents AI assistants from finding a curated summary of the site's content.

   What to change: Create an llms.txt file with a factual company description and links to key pages.

4. Near-zero external citation footprint High

   Web searches for the brand name, reviews, and parent company return zero indexed results across Google, Bing, and general web search. No press coverage, Reddit threads, or third-party reviews were discoverable. This lack of external signals contributes to the LLM prior being incorrect.

   What to change: Build external citations through press releases, business listings, review platforms, and industry directories to establish an online presence.

5. JSON-LD uses BankOrCreditUnion type despite not being a bank Medium

   Every page carries a BankOrCreditUnion schema, but the site's own FAQ states 'MultiPass is not a bank.' The correct schema type would be FinancialService or EMI. This inaccuracy can confuse AI crawlers and reduce trust.

   What to change: Replace BankOrCreditUnion schema with FinancialService or a more accurate type, and ensure all schema properties match the actual business.

6. WebSite SearchAction URL template is malformed Low

   The SearchAction target URL template is set to 'https://multipass.co/{search_term_string}' without a query parameter, making it non-functional for AI crawlers that support site search.

   What to change: Update the SearchAction target to include a query parameter, e.g., 'https://multipass.co/search?q={search_term_string}'.

7. Help centre page returns HTTP 500 High

   The /help-centre page returns a Nuxt.js server error, making critical support content inaccessible to AI crawlers and users alike.

   What to change: Fix the server error on the help centre page to serve its content properly.

8. Solutions page contains only 55 words Low

   The /solutions page has minimal content (55 words) and functions as a thin navigation hub rather than a substantive page. This provides little value to AI crawlers and may be seen as low-quality.

   What to change: Expand the solutions page with detailed descriptions of each solution, use cases, and benefits.

What's working

• All major AI crawlers receive full HTML content — All tested AI crawlers (GPTBot, ClaudeBot, PerplexityBot, etc.) receive 200 responses with full HTML content from Cloudflare. No bot is blocked or served a JS shell, ensuring content is indexable.
• Key product pages have substantive content — Product pages for global account, batch payments, corporate cards, and open banking API all return 200 with 400-1000 words of descriptive content, providing rich material for AI crawlers.
• FAQPage schema present on homepage and batch-payment page — The homepage and batch-payment page include FAQPage schema with 7 questions each, which can help AI assistants extract structured Q&A content.
• Consistent JSON-LD graph across all pages — Every page carries a consistent JSON-LD graph with Corporation and WebSite types, providing structured identity data to AI crawlers.
• Robots.txt does not block any AI crawlers — The robots.txt file contains no AI-specific directives, and the catch-all rule only blocks parameterized URLs and registration paths, not core content.