Here's the thing: Cubeish has made smart infrastructure moves (robots.txt welcoming AI crawlers, llms.txt, AGENTS.md) that put it ahead of most small agencies on technical access. But the site's actual content is too thin at ~449 words, too generic, and completely unverified. No AI engine can cite what isn't there. The llms.txt file actually contains more useful information about Cubeish than the website itself. And the impressive client roster (UNICEF, Microsoft, ADNOC)? It exists only as untagged logos with zero external validation anywhere on the internet.
| Category | Score | Weight | Weighted |
|---|---|---|---|
| AI Citability | 29 | 25% | 7.25 |
| Brand Authority | 11 | 20% | 2.20 |
| Content E-E-A-T | 18 | 20% | 3.60 |
| Technical GEO | 64 | 15% | 9.60 |
| Schema & Structured Data | 22 | 10% | 2.20 |
| Platform Optimization | 34 | 10% | 3.40 |
| Overall GEO Score | 28/100 | ||
Best platform: ChatGPT (45), full crawler access with all 3 OpenAI bots explicitly allowed, plus llms.txt. Worst: Gemini (12), near-zero Google ecosystem presence. No YouTube, no Google Business Profile, no Knowledge Graph signals.
The entire website has fewer words than a single blog post. Think about that. AI engines need substantial, specific content to extract and cite, and ~449 words of generic marketing copy gives them nothing to work with. When someone asks an AI about eLearning providers, Cubeish simply won't come up.
No Wikipedia article. No Reddit mentions. No YouTube channel. No Clutch or G2 reviews. No industry press. The client list (UNICEF, Microsoft, ADNOC, Samsung) exists only in llms.txt and as unlabeled logos. Search "Cubeish UNICEF" and you get zero results. AI models can't verify or cite claims they can't find.
A 23K-character page of fake content is being crawled and indexed right now. AI crawlers consuming this will associate Cubeish with template gibberish. Remove the page, finish it with real content, or add a noindex directive. Today.
Client names (UNICEF, Microsoft, ADNOC, PwC Academy, Samsung, AUC) appear only in llms.txt, never on the actual website. The "Clients" section shows logos with zero text. AI models match queries to named entities, and unlabeled logos are invisible to them. You can't get credit for clients nobody knows you have.
UNICEF Learning Passport Egypt, Microsoft CSR, ADNOC. Zero project descriptions, zero outcomes, zero evidence. Case studies are the #1 citability asset for agencies. Honestly, this is the single action that would lift all 5 platform scores the most.
Zero named people anywhere. No bios, no credentials, no photos, no Person schema. E-E-A-T Experience and Expertise both scored 3/25. If visitors and AI models can't figure out who actually delivers the work, they move on.
Zero cross-platform entity links. Without sameAs, AI models treat Cubeish as an unverified, isolated entity. This is the biggest single schema gap holding back AI entity recognition.
All 3 answers boil down to "depends on complexity" and "depends on various factors." AI engines skip vague responses like these. There are only 3 questions instead of the recommended 12-20, and the ones that exist aren't giving anyone useful information. Replace with specific ranges and concrete details.
Canonical and OG tags point to www.cubeish.co but the site serves from cubeish.co. This splits indexing signals and confuses AI crawlers about the authoritative URL.
Returns 404. AI crawlers rely on sitemaps to discover and prioritize pages. Without one, crawlers can only follow links from the homepage.
Zero presence in AI training data beyond the homepage. When users ask "How to create effective compliance training?" or "eLearning trends in MENA 2026," Cubeish will never be cited.
H1(1), H2(9), H5(13), H6(7). Section headings use H6 instead of H2. Damages semantic parsing for all AI engines.
Working Arabic toggle but zero hreflang signals. Arabic content is invisible to search engines and AI systems. Arabic eLearning queries are a low-competition opportunity being missed.
Only Organization + FAQPage exist. 8 service lines, team members, and AI-assistant-ready content have no structured data representation.
Near-zero Google ecosystem presence (Gemini scored 12). LinkedIn is critical for Bing Copilot integration and ChatGPT entity recognition.
No Clutch, G2, GoodFirms, or Trustpilot profiles. Perplexity heavily cites review platforms when answering agency queries.
No HSTS, CSP, X-Frame-Options, or other security headers. HTTPS is active but all other trust headers are absent.
Every content block scores below the 70-point "citation-ready" threshold. The strongest passage (About/Mission) scores 41. It mentions the MENA region and lists industry verticals, but contains zero data points. Service descriptions average 23, which is too thin to quote. FAQ answers are "anti-citations" that AI models actively skip. The single biggest gap: zero statistical density across the entire site. Not one number, percentage, or measurable outcome. Anywhere.
Near-total absence from the platforms AI models use for entity verification. No Wikipedia, which is the strongest negative signal. No Reddit, YouTube, or industry press. LinkedIn exists (401 followers, 2-10 employees) but engagement is minimal. The claimed enterprise clients can't be verified through any third-party source. Search for "Cubeish" alongside any client name and you get zero co-occurrences.
Experience: 3/25, zero case studies despite UNICEF/Microsoft credentials. Expertise: 3/25, no named team, no credentials, no methodology. Authoritativeness: 4/25, a 3-sentence about section and no thought leadership. Trustworthiness: 6/25, no testimonials, no named people, evasive FAQ answers. Look, the content reads like it could belong to any agency. Broad phrasing, no specifics, polished structure but nothing that would make you remember Cubeish over a competitor.
The strongest dimension by far. Static HTML delivery means AI crawlers see everything without JavaScript execution (95/100 on SSR). Robots.txt is exemplary, with individual Allow rules for 10+ AI crawlers. llms.txt and AGENTS.md are forward-thinking additions. What drags it down: canonical mismatch, missing sitemap, broken heading hierarchy, no security headers, 11 CSS + 14 JS files for a 4-page site, and that Lorem Ipsum grow.html page.
Organization and FAQPage schemas exist in valid JSON-LD, which is good. But Organization lacks sameAs (zero entity links), description, knowsAbout, and hasOfferCatalog. No Person, Service, WebSite, BreadcrumbList, or speakable schemas exist. There's also a Glasgow address conflict between schema (48 West George St) and corporate docs (272 Bath St). FAQPage has been restricted since Aug 2023, though it still provides AI semantic value.
ChatGPT scores highest (45) thanks to full crawler access. Gemini is the worst at 12, mostly because there's almost no Google ecosystem presence at all. Every platform hits the same wall: content is too thin and unverified for citation. The cross-platform picture is bleak. No LinkedIn, no YouTube, no Google Business Profile, no review site profiles.
Add <meta name="robots" content="noindex"> to grow.html until real content is ready. 5-minute fix that stops AI crawlers from ingesting gibberish. Affects all platforms.
Replace all H6 section headings with H2 and H5 service titles with H3. Zero-cost semantic improvement that helps every AI engine parse the page. 30-minute fix.
Change "Join this diverse range of partners" to named text: "UNICEF, Microsoft, ADNOC, PwC Academy, Samsung, AUC." Instantly adds entity references AI can match to queries. 15-minute fix.
Link to LinkedIn, X/Twitter (@CubeishLearning), and any other verified profiles. When there's no Wikipedia article, this is how AI models figure out that you're a real entity. 20-minute fix.
Update canonical and OG URL to match the actual serving domain (cubeish.co). Create a sitemap.xml with all 4 pages. Add Sitemap directive to robots.txt. 30-minute fix.
| URL | Title | Issues |
|---|---|---|
| cubeish.co/ | Digital Learning & Corporate Training Solutions | Thin content, broken headings, no external links, vague FAQ, missing schema properties |
| cubeish.co/grow.html | GROW LMS | Lorem Ipsum placeholder content, no schema, no meta tags, address mismatch |
| cubeish.co/privacy.html | Privacy Policy | "Being finalised" — should be noindexed |
| cubeish.co/terms.html | Terms of Use | "Being finalised" — should be noindexed |