way.com — AI Site Grade

Analysis

Way.com is a single-page Angular app that selectively gates AI crawlers behind Cloudflare's JS challenge — GPTBot, ClaudeBot, and PerplexityBot get full rendered content while Google-Extended, Perplexity-User, and regular browsers are blocked entirely.

Crawler Access

The site runs on Cloudflare with a JS challenge that blocks browser UAs, Google-Extended, Perplexity-User, anthropic-ai, and Bytespider (all 403). Meanwhile GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, ChatGPT-User, and Applebot-Extended receive a full 200 response with ~506KB of pre-rendered HTML served via SEO4Ajax — a server-side rendering proxy for Angular. The /robots.txt and /llms.txt are both behind the same Cloudflare challenge and return 403 to browser UAs, making them invisible to anyone who doesn't spoof an allowed bot. No AI-bot-specific directives exist because the file cannot be read. The /sitemap.xml is also 403'd. This creates a paradoxical posture: some AI crawlers see the full site while Google's own crawler (Google-Extended) is locked out.

Cold-Knowledge Gap

The LLM prior describes Way.com as a "digital platform" for parking, car washes, EV charging, and auto insurance, founded 2016, based in San Francisco, with $100M+ funding. The actual site (via Wayback and GPTBot fetches) reveals a far more ambitious scope: Way is a full-stack car ownership platform covering parking (3,500+ facilities, 300+ cities), car washes (40,000+ locations), gas cashback (100,000+ stations), EV charging (70,000+ stations), an IRS-compliant mileage tracker, auto refinance, auto repair, roadside assistance (Way+), and a Way Wallet earning 8% APY. The founding date on the site's own JSON-LD schema is 2013 (not 2016), and the headquarters is Fremont, CA (not San Francisco). The site claims "Trusted by 10M+ drivers" and "9M+ customers" — the LLM prior knows nothing about the mileage tracker, gas cashback, Way Wallet, or the Way+ subscription.

Schema Posture

The homepage and key subpages carry Organization and WebSite JSON-LD schema with founder name (Binu Girija), founding date (2013), address, phone, email, and social links. The WebSite schema includes a SearchAction potential action targeting /mainsearch?citynamesearch={search_term_string}. However, the schema is static and identical across all pages — the parking page, insurance page, and homepage all carry the exact same Organization block with no page-specific schema like Product, FAQPage, LocalBusiness, or Service. The insurance page has a real FAQ section with 6 questions but no FAQPage schema marking it up. The blog/journal uses WebPage, BreadcrumbList, and Organization schema but no Article or BlogPosting markup on individual posts.

External Signals

The App Store listing shows 57,884 ratings with 4.8 stars — a strong positive signal. The Google Play store shows 8,611 ratings at 3.9 stars, with recent reviews flagging issues around parking lot quality, suppressed reviews, and gas cashback redemption friction. The app description on both stores is more comprehensive than the website itself, listing all seven service categories clearly. The DNS TXT records reveal integrations with HubSpot, SendGrid, Mailchimp, Klaviyo, and Zoho — plus an anthropic-domain-verification record confirming Way.com has explicitly verified its domain with Anthropic for Claude access. The blog lives at journal.way.com (WordPress, PHP 8.0.30) while the main app is an Angular SPA — a split architecture where the blog is directly accessible to GPTBot but the main site requires pre-rendering.

Findings

1. Google-Extended blocked by Cloudflare JS challenge High

   Google-Extended receives a 403 response from Cloudflare's JS challenge, preventing Google's AI crawler from indexing the site. This is paradoxical because other AI crawlers like GPTBot and ClaudeBot are allowed.

   What to change: Allow Google-Extended through Cloudflare by adding a firewall rule that permits its user-agent or by serving a static version without JS challenge.

2. Robots.txt returns 403 to browsers High

   The robots.txt file is behind Cloudflare's JS challenge and returns 403 to browser user-agents, making it unreadable for anyone who does not spoof an allowed bot. No AI-bot-specific directives are visible.

   What to change: Serve robots.txt without Cloudflare challenge so all crawlers can read it, and add explicit directives for AI bots.

3. Sitemap.xml returns 403 High

   The sitemap.xml is behind Cloudflare's JS challenge and returns 403, preventing crawlers from discovering the site's URL structure.

   What to change: Serve sitemap.xml without Cloudflare challenge so all crawlers can access it.

4. LLMs.txt returns 403 Medium

   The llms.txt file is behind Cloudflare's JS challenge and returns 403, preventing AI crawlers from discovering a structured overview of the site.

   What to change: Serve llms.txt without Cloudflare challenge and populate it with a summary of the platform's services.

5. Regular browser access blocked by Cloudflare JS challenge High

   The main site returns 403 to browser user-agents due to Cloudflare's JS challenge, making the site inaccessible to users who do not execute JavaScript. This also blocks crawlers that use browser-like UAs.

   What to change: Remove the JS challenge for browser user-agents or implement a less restrictive challenge that allows direct access.

6. Static identical JSON-LD schema across all pages Medium

   The homepage, parking page, insurance page, and other subpages all carry the exact same Organization and WebSite JSON-LD schema with no page-specific markup. This limits AI understanding of individual page content.

   What to change: Add page-specific schema such as Product, FAQPage, LocalBusiness, or Service to relevant pages.

7. Insurance FAQ section lacks FAQPage schema Medium

   The insurance page contains a real FAQ section with 6 questions, but no FAQPage schema is present to mark them up for AI crawlers.

   What to change: Add FAQPage schema to the insurance FAQ section.

8. Blog posts lack Article or BlogPosting schema Medium

   The blog/journal uses WebPage, BreadcrumbList, and Organization schema but no Article or BlogPosting markup on individual posts, reducing AI understanding of article content.

   What to change: Add Article or BlogPosting schema to individual blog posts.

9. LLM prior knowledge is outdated and incomplete Medium

   The LLM prior describes Way.com as a parking and insurance platform founded in 2016 in San Francisco, but the site reveals a full car ownership platform founded in 2013 in Fremont, CA, with services like mileage tracker, gas cashback, Way Wallet, and Way+ subscription not mentioned in the prior.

   What to change: Ensure the site's content and schema accurately reflect the full scope of services to correct AI knowledge gaps.

10. No web search results found for Way.com High

    Multiple web searches for Way.com and its services returned zero results, indicating extremely low external visibility and citation on the open web.

    What to change: Investigate why the site is not appearing in search results and improve SEO and backlink profile.

11. Blog subdomain blog.way.com fails DNS resolution Medium

    The blog subdomain blog.way.com failed DNS resolution when accessed as GPTBot, indicating it may be misconfigured or not publicly accessible.

    What to change: Fix DNS records for blog.way.com to ensure it resolves and is accessible to crawlers.

12. About Us page returns 404 Low

    The about-us page at way.com/about-us returns a 404 error, missing an important page for brand information and schema.

    What to change: Restore the about-us page with accurate company information and schema.

What's working

• Server-side rendering for allowed AI crawlers — GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, ChatGPT-User, and Applebot-Extended receive a full 200 response with ~506KB of pre-rendered HTML via SEO4Ajax, ensuring they see complete content.
• Organization and WebSite JSON-LD schema present — The homepage and key subpages carry Organization and WebSite JSON-LD schema with founder name, founding date, address, phone, email, and social links, providing basic structured data.
• Strong App Store ratings with 4.8 stars — The iOS App Store listing shows 57,884 ratings with 4.8 stars, a strong positive external signal for AI crawlers that consider app store data.
• Anthropic domain verification record present — DNS TXT records include an anthropic-domain-verification record, confirming Way.com has explicitly verified its domain with Anthropic for Claude access.
• Blog accessible to GPTBot with full content — The blog at journal.way.com is directly accessible to GPTBot and returns 200 with full HTML content, including article text and schema.
• Comprehensive app description on stores — The App Store and Google Play descriptions list all seven service categories clearly, providing a rich source of structured information for AI crawlers that index app store content.
• SearchAction potential action in WebSite schema — The WebSite schema includes a SearchAction potential action targeting /mainsearch?citynamesearch={search_term_string}, enabling AI crawlers to understand the site's search functionality.
• Wayback Machine snapshot available — A recent Wayback Machine snapshot from 2026-05-06 shows the site's content, providing a fallback for historical analysis.