u.com.my — AI Site Grade

Analysis

I have enough data. Let me compile the audit.

U Mobile's AI crawlers get full 200 access but find a JavaScript-heavy shell with barely any readable content

Crawler Access

All major AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended, ChatGPT-User, anthropic-ai — receive a 200 status with identical byte payload (664KB) to a browser. No UA-based blocking exists. However, the homepage and all key pages return under 300 words of visible text from a plain GET, dominated by JSON configuration objects, carousel markup, and template interpolation strings like {{prefixRM(addCommaForCurrency(0))}}. The site is a heavy JavaScript single-page application served behind AWS CloudFront; AI crawlers see a shell, not the actual plan pricing, feature comparisons, or call-to-action content.

Infrastructure and Discovery Gaps

The domain runs on AWS (CloudFront + Route53) with no robots.txt (404), no llms.txt (404), and no sitemap.xml (404). The only URL discovery mechanism is internal navigation links. The site has no Organization schema on any page — every page uses only a bare WebPage schema with datePublished dates in 2026 (future-dated), suggesting a CMS timezone or build-pipeline issue. The canonical URLs strip language/segment paths inconsistently (e.g., /en/postpaid instead of /en/personal/mobile-plans/postpaid), creating a fragmented URL structure.

Cold-Knowledge Gap

The LLM prior describes U Mobile as a younger-audience, data-focused MVNO-style operator with known products like "U28" and "Hero P70" plans, and notes network congestion complaints and slower 5G rollout. The actual site positions itself as "Malaysia's newest 5G network provider" operating a national 5G Wholesale Network with enterprise slicing, private networks, and an Enterprise Innovation Platform (EIP) for 5G-AI co-creation. This is a fundamentally different brand story — the cold model knows nothing about the wholesale network, the EIP initiative, the ULTRA5G brand, or the pivot from consumer MVNO to infrastructure operator. The site's actual competitive differentiators (5G-Advanced, network slicing, wholesale) are entirely absent from the model's prior.

Content and Schema Posture

The site has no FAQPage schema despite FAQ sections on multiple pages. No Product, MobileApplication, Service, or PriceSpecification schema exists anywhere — plan pricing is rendered client-side and invisible to crawlers. The heading structure is repetitive (identical nav H3s duplicated on every page) and the h1 tags are marketing slogans ("Go Big. Game Big.", "ULTRA-LEVEL SPEED") rather than descriptive page titles. The broadband wireless page is the only page with substantive text (~5,300 words), but it is buried in JSON-encoded HTML fragments.

Findings

1. Key pages render as JavaScript shells with under 300 words of visible text High

   All major pages return 200 status but contain minimal readable content (under 300 words) due to heavy JavaScript rendering. AI crawlers see JSON configuration, carousel markup, and template interpolation strings instead of actual plan pricing, features, or CTAs.

   What to change: Implement server-side rendering or dynamic rendering for AI crawlers to expose plan details, pricing, and feature comparisons as static HTML.

2. No robots.txt file (404) Medium

   The domain returns a 404 for robots.txt, meaning crawlers have no explicit instructions on which paths to crawl or avoid. This can lead to inefficient crawling or accidental blocking of important resources.

   What to change: Create a robots.txt file that allows all well-behaved crawlers and disallows irrelevant paths (e.g., /assets/).

3. No sitemap.xml file (404) Medium

   The domain returns a 404 for sitemap.xml, preventing crawlers from efficiently discovering all pages. Only 25 URLs were discovered via internal links.

   What to change: Generate and submit a sitemap.xml covering all important pages, including product, plan, and enterprise pages.

4. No llms.txt file (404) Low

   The domain returns a 404 for llms.txt, missing an opportunity to guide AI crawlers to key content and provide context about the site's structure.

   What to change: Create an llms.txt file that lists important pages and provides a brief summary of the site's offerings.

5. No Organization schema on any page High

   Every page uses only a bare WebPage schema. The absence of Organization schema means AI crawlers cannot reliably extract the company name, logo, social profiles, or contact info.

   What to change: Add Organization schema with name, logo, URL, and social media links to all pages.

6. Schema datePublished values are set to 2026 (future dates) Medium

   Multiple pages have datePublished values in 2026, indicating a CMS timezone or build-pipeline issue. This can confuse crawlers and reduce trust in the content's timeliness.

   What to change: Fix the date generation logic to output correct current dates.

7. Canonical URLs strip language/segment paths inconsistently Medium

   Canonical URLs use shortened paths (e.g., /en/postpaid instead of /en/personal/mobile-plans/postpaid), creating a fragmented URL structure that may dilute ranking signals.

   What to change: Ensure canonical URLs match the full, correct path of the page.

8. No FAQPage schema despite FAQ sections Medium

   Multiple pages contain FAQ sections but lack FAQPage structured data, reducing the chance of appearing in rich results or being understood by AI crawlers.

   What to change: Add FAQPage schema to pages with FAQ content.

9. No Product, MobileApplication, or PriceSpecification schema for plans High

   Plan pricing and features are rendered client-side and invisible to crawlers. No structured data exists for mobile plans, broadband, or enterprise services, making it impossible for AI to extract pricing or feature comparisons.

   What to change: Add Product or Service schema with price, currency, and description for each plan, rendered server-side.

10. LLM prior lacks knowledge of U Mobile's 5G wholesale network and enterprise pivot High

    The cold LLM prior describes U Mobile as a consumer MVNO with known plans, but the site positions itself as a national 5G network provider with wholesale, network slicing, and an Enterprise Innovation Platform. This brand story is entirely absent from AI knowledge.

    What to change: Improve crawlable content and structured data on enterprise and 5G pages, and consider publishing an llms.txt or knowledge graph to seed AI models.

11. Repetitive heading structure with marketing slogans as H1s Low

    The heading structure is repetitive (identical nav H3s on every page) and H1 tags are marketing slogans like 'Go Big. Game Big.' rather than descriptive page titles, harming semantic understanding.

    What to change: Use unique, descriptive H1 tags that summarize the page content, and reduce duplication in navigation headings.

12. Minimal external web presence and citations Medium

    Web searches for U Mobile reviews, 5G coverage, and Reddit discussions returned zero results, indicating low external signals that could help AI models validate the brand's claims.

    What to change: Encourage customer reviews, press coverage, and community discussions to build external signals.

What's working

• All major AI crawlers receive 200 status with no UA blocking — GPTBot, ClaudeBot, PerplexityBot, and others are not blocked by user-agent or IP, allowing them to access the site freely.
• Broadband wireless page contains ~5,300 words of substantive text — The U Home 5G page has extensive descriptive content about the service, which is valuable for AI crawlers despite being embedded in JSON fragments.
• Enterprise Innovation Platform page provides detailed 5G-AI co-creation content — The EIP page contains ~591 words describing the platform, which helps communicate U Mobile's enterprise capabilities to AI crawlers.
• 5G Wholesale Network page has ~1,542 words of detailed technical content — This page explains network slicing, private networks, and wholesale offerings, providing rich material for AI crawlers.
• Site hosted on AWS CloudFront with good performance — The site uses AWS CloudFront CDN, which provides fast global delivery and scalability.
• Historical snapshot available on Wayback Machine — A snapshot from June 2023 exists, indicating some archival presence.