tmbl.com.au — AI Site Grade

Analysis

Teachers Mutual Bank Limited (tmbl.com.au) — AI-Visibility Audit

The site has exemplary AI crawler access policies but zero structured data on any page, creating a paradox where bots are welcomed yet handed nothing machine-readable to understand the entity.

Crawler Access

All 11 tested AI user-agents (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, anthropic-ai, Applebot-Extended, Bytespider, Perplexity-User, and a browser baseline) receive 200 status with identical 23,251-byte HTML from Cloudflare. No UA-based blocking, no JS shell, no CAPTCHA. The robots.txt explicitly allows GPTBot, ClaudeBot, CCBot, Applebot-Extended, and GPT-user with Allow: / directives — a rare and commendable openness for a financial institution. The llms.txt returns 404, a missed opportunity to declare an AI-friendly content map for a site with deep, well-structured content across five retail brands.

Schema Posture

Every page inspected — homepage, about, banking-for-good, merger, news articles, product FAQs — contains zero JSON-LD or any structured data. No Organization, FinancialService, BankOrCreditUnion, FAQPage, NewsArticle, or BreadcrumbList schema is present. This is the single largest AI-visibility gap. A site that is a B Corp-certified mutual bank with five retail brands, a merger in progress, and rich FAQ content is invisible to knowledge-graph construction by search engines and AI crawlers at the semantic level.

Cold-Knowledge Gap

A frontier LLM queried cold knows TMBL as a "customer-owned bank serving teachers and education staff" with "over 100 years of history (founded 1967)." The actual site says the bank was founded in 1966 (as Hornsby Teachers Credit Union) and now serves education, higher education, emergency services, and healthcare — not just teachers. The model also claims B Corp certification was achieved in 2016; the site states it was first certified in 2022 and recertified in 2025. The model knows nothing about the Australian Mutual Bank merger (approved March 2026, effective 1 May 2026), the five-brand structure, the $14.2B in assets, or the B Corp recertification with a 121.6 score. These are material facts absent from the model's prior.

Sitemap Pollution

The sitemap at /sitemap.xml contains approximately 50% non-navigational URLs — AEM content fragments at paths like /content/tmb/global-fragments/product-faqs/momentum-saver and /experience-fragments/payment-services-availability. These pages render as standalone HTML (confirmed: GPTBot gets 200 with real FAQ content), but they duplicate content that exists on the proper brand subdomains (e.g., tmbank.com.au). The canonical tags point to the fragment path, not the canonical product page, creating a fragmented citation surface for AI engines.

External Signals

The DNS TXT records reveal verifications for OpenAI (openai-domain-verification), Anthropic (anthropic-domain-verification), and Apple (apple-domain-verification) — confirming proactive engagement with AI crawler ecosystems. The site runs on Cloudflare with HSTS, X-Frame-Options SAMEORIGIN, and strong cache headers. No external review sites or Reddit threads surfaced in search, suggesting limited off-domain brand chatter that AI engines could cite.

Findings

1. Zero JSON-LD structured data on any page High

   Every page inspected — homepage, about, banking-for-good, merger, news articles, product FAQs — contains no JSON-LD or any structured data. No Organization, FinancialService, BankOrCreditUnion, FAQPage, NewsArticle, or BreadcrumbList schema is present, making the site semantically invisible to AI crawlers and knowledge graphs.

   What to change: Add JSON-LD structured data to all pages: Organization schema on the homepage and about page, FinancialService or BankOrCreditUnion schema on product pages, FAQPage schema on FAQ pages, NewsArticle schema on news articles, and BreadcrumbList schema site-wide.

2. llms.txt returns 404 Medium

   The site does not serve an llms.txt file, missing an opportunity to provide an AI-friendly content map for a site with deep, well-structured content across five retail brands.

   What to change: Create an llms.txt file listing key pages and summaries for AI crawlers, including brand pages, product FAQs, and news articles.

3. LLM cold knowledge contains factual errors and omissions High

   A frontier LLM queried cold incorrectly states the bank was founded in 1967 (actual: 1966), B Corp certification achieved in 2016 (actual: 2022), and service scope limited to teachers (actual: education, higher education, emergency services, and healthcare). The model knows nothing about the Australian Mutual Bank merger, the five-brand structure, $14.2B in assets, or the B Corp recertification with a 121.6 score.

   What to change: Add Organization schema with founding date, B Corp certification details, service sectors, and merger information. Publish a dedicated page or press release with key facts and ensure it is well-linked and included in the sitemap.

4. Sitemap contains ~50% AEM content fragment URLs Medium

   The sitemap includes many non-navigational AEM content fragment paths (e.g., /content/tmb/global-fragments/product-faqs/momentum-saver) that duplicate content from proper brand subdomains. Canonical tags point to the fragment path, not the canonical product page, creating a fragmented citation surface for AI engines.

   What to change: Remove AEM content fragment URLs from the sitemap or replace them with canonical product page URLs. Ensure canonical tags point to the authoritative brand subdomain pages.

5. Product FAQ pages lack FAQPage schema Medium

   The Momentum Saver FAQ page contains rich Q&A content but has no FAQPage structured data, preventing AI crawlers from extracting the questions and answers as structured knowledge.

   What to change: Add FAQPage schema to all product FAQ pages, marking up each question and answer pair.

6. News articles lack NewsArticle schema Medium

   News articles such as the merger vote and B Corp recertification have no NewsArticle structured data, reducing their chance of appearing in Google News and AI-generated news summaries.

   What to change: Add NewsArticle schema to all news articles with headline, datePublished, author, and image fields.

7. Limited external brand mentions and reviews Low

   Web searches for reviews and merger information returned zero results, indicating low off-domain brand chatter that AI engines could cite as external signals.

   What to change: Encourage customer reviews on third-party platforms and publish press releases that are picked up by news outlets to increase external citations.

What's working

• All 11 tested AI crawlers receive full access with 200 status — The site returns identical 200 HTML to all tested AI user-agents, with no UA-based blocking, JS shells, or CAPTCHAs. Robots.txt explicitly allows GPTBot, ClaudeBot, CCBot, Applebot-Extended, and GPT-user with Allow: / directives.
• DNS TXT records verify domain with OpenAI, Anthropic, and Apple — The DNS TXT records include openai-domain-verification, anthropic-domain-verification, and apple-domain-verification, confirming proactive engagement with AI crawler ecosystems.
• Cloudflare with HSTS, X-Frame-Options, and strong cache headers — The site runs on Cloudflare with HSTS, X-Frame-Options SAMEORIGIN, and strong cache headers, ensuring secure and fast delivery to all crawlers.
• Product FAQ pages render full HTML content for GPTBot — The Momentum Saver FAQ page returns 200 with 10,834 bytes of HTML when fetched as GPTBot, confirming that AI crawlers can access the full Q&A content.
• Sitemap contains 80 URLs covering key pages — The sitemap at /sitemap.xml lists 80 URLs, providing a broad map of the site's content for crawlers.