mazuma.org — AI Site Grade

Analysis

Mazuma.org: Full AI-visibility access, but a cold-knowledge gap on a flagship product and zero external citation footprint

Every major AI crawler — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended — receives a 200 with full HTML content identical to the browser baseline (214 KB, Cloudflare edge). No UA-based blocking, no JS shell, no 403. The site is hosted on HubSpot CMS behind Cloudflare, with HSTS and a s-maxage=36000 cache policy. This is an unusually open posture for a financial institution.

Crawler Access

The robots.txt contains only a catch-all User-agent: * rule disallowing HubSpot preview and cache-busting paths. No AI bot is explicitly named or restricted. No llms.txt exists (404). The sitemap at /sitemap.xml lists 115 URLs and is well-formed. Every bot gets the same full page the browser does — no conditional gating, no login-wall, no rate-limiting detectable at the homepage level. The BankOrCreditUnion JSON-LD schema is present on every page examined, carrying the same aggregateRating (4.8/208 reviews) and sameAs social links.

Cold-Knowledge Gap

An LLM queried cold recalls Mazuma as a Kansas City credit union founded to serve DST Systems employees, with a "Mazuma Money" high-yield checking account and a "Skip-a-Pay" loan feature. The actual site has no mention of "Mazuma Money" or "Skip-a-Pay" anywhere. The checking page offers Classic, Signature, and Premier checking — none branded "Mazuma Money." The homepage promotes a 6.00% APR HELOC introductory rate, a FIFA World Cup ticket giveaway, and recreational vehicle loans. The LLM's prior about the DST Systems origin is also absent from the site's own founding story (which credits Wyla and Joe Pinkerton, not DST). The model knows a product line the site has apparently retired or renamed, and the site's current promotional emphasis (HELOC teaser, World Cup, RV loans) is entirely absent from the model's cold recall.

Schema & Content Posture

The BankOrCreditUnion schema is thorough: name, address, geo, phone, hours, currencies, payment methods, aggregate rating, and six sameAs URLs. However, the sameAs list omits any Google Business Profile or Yelp link — only TikTok, Facebook, Instagram, LinkedIn, and YouTube. The checking account page is the only page with FAQPage schema (6 Q&A pairs). Blog posts carry BlogPosting schema with author, datePublished, and dateModified. The rates page has a live HTML table of APY/APR figures but no Product or Offer schema for individual loan products. No WebSite schema with potentialAction (site search) is present. Heading structure is flat — most pages use H2s for major sections with no H1 on the blog listing page.

External Signals

Web searches for "Mazuma Credit Union" reviews, Reddit mentions, and news articles returned zero indexed results from DuckDuckGo. The brand has no detectable press coverage, review-site presence, or forum discussion in the search results sampled. The only external signals are the social profiles linked in the schema (sameAs). This near-zero off-domain footprint means AI engines have almost no third-party corroboration to draw on — the cold LLM knowledge is likely derived from older training data rather than live web signals. The site's own blog is active (posts from 2024-2025 on fraud awareness, cooperative principles, community events), but none of this content appears to be cited externally.

Findings

1. LLM cold knowledge references a retired product line not found on the site High

   An LLM queried cold recalls a 'Mazuma Money' high-yield checking account and 'Skip-a-Pay' loan feature, but the current site offers Classic, Signature, and Premier checking with no mention of those products. The model's prior about the DST Systems origin is also absent from the site's founding story.

   What to change: Update the site's content to align with current product offerings and ensure the founding story is accurately reflected. Consider adding a page or section that clarifies the history and any product name changes.

2. No external citations found in web search results High

   Web searches for reviews, Reddit mentions, and news articles returned zero indexed results. The brand has no detectable press coverage, review-site presence, or forum discussion, meaning AI engines have almost no third-party corroboration.

   What to change: Build an external citation footprint by encouraging reviews on Google, Yelp, and other platforms, and by engaging in community events that generate press coverage.

3. No llms.txt file published Medium

   The site returns a 404 for /llms.txt, missing an opportunity to provide AI crawlers with a structured summary of the site's content and key pages.

   What to change: Create and publish an llms.txt file that lists important pages and a brief description of the site's content.

4. Rates page lacks Product or Offer schema for loan products Medium

   The rates page has a live HTML table of APY/APR figures but no structured data markup for individual loan products, reducing the chance of AI engines extracting specific rate information.

   What to change: Add Product or Offer schema markup to each loan product row in the rates table.

5. No WebSite schema with site search potentialAction Low

   The site does not include WebSite schema with a potentialAction for site search, which would help AI engines understand the search functionality.

   What to change: Add WebSite schema with a SearchAction potentialAction to the homepage.

6. Flat heading structure on blog listing page Low

   The blog listing page uses H2s for major sections with no H1, which may hinder content hierarchy understanding for crawlers.

   What to change: Add a clear H1 heading to the blog listing page and ensure proper heading hierarchy.

7. sameAs links omit Google Business Profile and Yelp Medium

   The BankOrCreditUnion schema's sameAs list includes TikTok, Facebook, Instagram, LinkedIn, and YouTube but not Google Business Profile or Yelp, which are key for local search visibility.

   What to change: Add Google Business Profile and Yelp URLs to the sameAs list in the schema.

What's working

• All major AI crawlers receive full HTML content with no blocking — Every major AI crawler receives a 200 with full HTML content identical to the browser baseline. No UA-based blocking, JS shell, or 403 errors.
• BankOrCreditUnion JSON-LD schema present on every page — The schema includes name, address, geo, phone, hours, currencies, payment methods, aggregate rating (4.8/208 reviews), and six sameAs social links.
• FAQPage schema on checking account page with 6 Q&A pairs — The checking account page includes FAQPage schema with six question-answer pairs, helping AI engines extract common questions.
• Blog posts carry BlogPosting schema with author and dates — Blog posts include BlogPosting schema with author, datePublished, and dateModified, which helps AI engines understand content freshness and authorship.
• Sitemap is well-formed with 115 URLs — The sitemap at /sitemap.xml lists 115 URLs and is well-formed, aiding crawler discovery.
• Active blog with recent posts on relevant topics — The blog has posts from 2024-2025 on fraud awareness, cooperative principles, and community events, providing fresh content for AI engines.