arrivia.com — AI Site Grade

Analysis

AI-Visibility Audit: arrivia.com

Arrivia has an llms.txt file (auto-generated by All in One SEO Pro) that is 141KB and contains the full text of every blog post, case study, and whitepaper — a rare and aggressive AI-friendly posture that most sites at this scale do not adopt, yet the robots.txt contains zero AI-bot-specific rules, leaving all crawlers to the wildcard * directive.

Crawler Access

All 11 AI bot user-agents tested (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Bytespider, Applebot-Extended, anthropic-ai, Perplexity-User) receive HTTP 200 with full HTML content identical to browser baseline (~124KB). No UA-based blocking, no Cloudflare challenge, no JS shell. The site runs on Azure DNS (NS pointing to azure-dns.com) with a single A record at 63.241.172.37. The anthropic-domain-verification and cursor-domain-verification TXT records confirm active AI-vendor relationship management. The llms.txt at /llms.txt returns 200 and is a full-content dump of the blog — not a curated summary, but a raw export of every post's excerpt and link.

Cold-Knowledge Gap

The LLM prior knows arrivia as a "travel technology and loyalty company founded in 2002, headquartered in Scottsdale, serving B2B clients." The site itself says it was founded in 1997 and rebranded in 2020 after acquisitions — a five-year discrepancy the model has no awareness of. The model also does not know about the ownership transition to a consortium of institutional investors (announced March 2026), the Agentic AI and booking intelligence investments the CEO describes, or the specific marquee clients (American Express, USAA, Marriott Vacation Club) that the homepage prominently names. The gap between "B2B travel tech company" and "world's largest stand-alone travel loyalty provider powering Amex and USAA" is significant.

Schema Posture

The site uses Organization, WebSite, WebPage, BreadcrumbList, FAQPage, and Service schema types across key pages. The homepage Organization schema includes address, phone, and a LinkedIn sameAs link — but no sameAs for Instagram or Facebook, despite those social profiles existing and being linked in the page footer. The Service schema on the financial-institutions page is well-structured with BusinessAudience and Offer types. FAQ schema is present on the homepage, product page, about page, and financial-institutions page — a strong answer-format signal. However, the about page schema lists a founding date of 1997 while the cold LLM knowledge says 2002, and the page text itself says "rebranded in 2020 to reflect several acquisitions."

Content & Signals

The homepage is a WordPress site (All in One SEO Pro v4.9.7.2) with 875 words of visible text, an H1 of "Driving Loyalty Through Travel Loyalty Platforms," and six FAQ accordions. The blog is actively maintained — posts from May 2026 appear, and the sitemap shows 27 pages of paginated blog content. The site claims "over 25 years" of operation and names American Express, USAA, and Marriott Vacation Club as clients. The dateModified field on the homepage schema reads 2026-01-15, which is a future date relative to typical audit expectations and suggests the site is actively maintained with forward-dated content. The addl-sitemap.xml is a near-duplicate of three URLs already in page-sitemap.xml, creating minor sitemap redundancy.

Findings

1. LLM cold knowledge lists wrong founding year (2002 vs 1997) High

   The LLM prior knows arrivia as founded in 2002, but the site states 1997 and a 2020 rebrand. This discrepancy means AI models may present incorrect historical information.

   What to change: Ensure consistent founding date across all site content and structured data; consider adding a 'foundingDate' property in Organization schema.

2. Organization schema missing Instagram and Facebook sameAs Medium

   The homepage Organization schema includes a LinkedIn sameAs but omits Instagram and Facebook, even though those profiles exist and are linked in the footer. This reduces the entity's digital footprint for AI models.

   What to change: Add Instagram and Facebook URLs to the sameAs array in the Organization schema on the homepage.

3. Addl-sitemap.xml contains duplicate URLs Low

   The addl-sitemap.xml is a near-duplicate of three URLs already present in page-sitemap.xml, creating redundancy that may confuse crawlers.

   What to change: Remove the addl-sitemap.xml or ensure it contains only unique URLs not already in other sitemaps.

4. Robots.txt has no AI-bot-specific rules Low

   The robots.txt contains only a wildcard '*' directive and does not explicitly allow or disallow any AI bot user-agents. While this means all bots are allowed by default, the lack of explicit rules may lead to inconsistent crawling behavior.

   What to change: Consider adding explicit allow rules for AI bots to ensure consistent access, or keep as-is if no issues arise.

What's working

• Large llms.txt with full blog content — The llms.txt file is 141KB and contains the full text of every blog post, case study, and whitepaper, providing AI models with extensive content for training and retrieval.
• All 11 AI bots receive full HTML content — All tested AI user-agents receive HTTP 200 with full HTML content identical to the browser baseline, with no UA-based blocking or JS shells.
• FAQPage schema on homepage, product, about, and financial-institutions pages — FAQPage schema is present on multiple key pages, providing structured answer-format signals that can enhance AI-generated snippets.
• Actively maintained blog with recent posts and sitemap — The blog has posts from May 2026 and the sitemap includes 27 pages of paginated blog content, indicating fresh content for AI models.
• Anthropic and Cursor domain verification TXT records present — The DNS TXT records include anthropic-domain-verification and cursor-domain-verification, confirming active AI-vendor relationship management.