satelliteoffice.com — AI Site Grade

Analysis

Satellite Office: An Offshoring BPO Wholly Misidentified by AI Models

The most consequential finding is that frontier LLMs cold-knowledge Satellite Office as a virtual-office/co-working provider (competing with Regus and WeWork), when the site actually sells full-time offshore team staffing in the Philippines — a complete category misidentification that means every AI-generated brand summary is fundamentally wrong.

Crawler Access

The site runs on nginx (SiteGround hosting) with a strict Content Security Policy and HSTS. The robots.txt is minimal — no AI-bot rules at all, just a generic Disallow: /wp-admin/. However, the server infrastructure silently drops connections from GPTBot, ClaudeBot, Google-Extended, anthropic-ai, Applebot-Extended, and Bytespider (all return status 0 / server disconnect). ChatGPT-User, OAI-SearchBot, PerplexityBot, and Perplexity-User all get 200 with full content (825KB). This creates a lopsided AI-crawl surface: OpenAI's search and ChatGPT-User can index the site, but GPTBot (training crawler) cannot. Google-Extended being blocked means Google's AI training pipeline has no access. The pattern is identical on the .com.au mirror domain. No llms.txt exists (404).

Cold-Knowledge Gap

When asked cold, a frontier LLM describes Satellite Office as "a provider of virtual office services, including mailing addresses, phone answering, meeting rooms, and coworking spaces" — a description that matches exactly zero of the site's actual offerings. The site is a Philippines-based offshore staffing and BPO provider that recruits, hires, and manages dedicated full-time teams (accounting, IT, customer support, legal, etc.) for US and Australian clients. The LLM also mentions "mixed customer reviews on Trustpilot and BBB" — no such review presence was found in external searches. The gap is total: the model has confused the brand name with an entirely different industry.

Schema Posture

Every page carries Yoast-generated WebPage, BreadcrumbList, and WebSite schema with SearchAction. Blog articles get Article schema with author, datePublished, and image. However, no Organization schema exists on the .com homepage (the .com.au version does include Organization with logo). No LocalBusiness, FAQPage, Product, or Service schema is present anywhere — despite the site having FAQ sections on multiple pages and listing dozens of service offerings. The homepage has no @type schema at all beyond the graph wrapper.

Content & JS-Rendering Risk

The homepage returns only 13 words of visible text from a plain GET — the bulk of the content is loaded via JavaScript (lazy-loaded sections, WP Rocket deferred scripts). The text_excerpt shows only "Learning and Development September 2, 2025 SO-LEAD: Your Career Journey Made More Rewarding." This means AI crawlers that execute JS (ChatGPT-User, PerplexityBot) see the full page, but crawlers that do not (GPTBot, even if it weren't blocked) would see essentially nothing. The site has strong content underneath: 2,000+ words on the outsourcing page, 1,381 on services, 906 on case studies, and a blog with 50+ articles. But the JS shell on the homepage is a significant rendering risk for non-executing crawlers.

External Signals

External search returns zero results for reviews, Reddit threads, or press coverage about Satellite Office as an offshoring company. The site itself lists notable clients (HelloFresh, King Living, Petbarn, Montu) and claims 160+ global clients, 2,000+ staff, ISO 27001 certification, and HIPAA compliance. The blog ranks itself as the #1 offshoring company in the Philippines. The leadership team page names 18 executives with LinkedIn profiles. But none of this validation appears in independent third-party sources discoverable via search — a visibility vacuum that leaves AI models reliant on the brand name alone, which they currently misinterpret.

Findings

1. AI models misidentify Satellite Office as a virtual-office provider High

   Frontier LLMs describe Satellite Office as a provider of virtual office services (mailing addresses, meeting rooms) when the site actually sells full-time offshore team staffing in the Philippines. This complete category error means every AI-generated brand summary is fundamentally wrong.

   What to change: Publish an Organization schema with the correct industry classification and a detailed llms.txt file that explicitly states the company's core offering. Build authoritative external content (press releases, case studies on third-party sites) to correct the AI training data.

2. Server silently drops connections from major AI training crawlers High

   GPTBot, ClaudeBot, Google-Extended, anthropic-ai, Applebot-Extended, and Bytespider all receive server disconnects (status 0) when attempting to crawl the homepage. ChatGPT-User, OAI-SearchBot, PerplexityBot, and Perplexity-User succeed. This blocks AI training pipelines while allowing search-oriented bots.

   What to change: Investigate the server configuration (nginx, SiteGround) to ensure AI crawlers are not blocked by rate limiting, IP restrictions, or TLS issues. Allow GPTBot, ClaudeBot, and Google-Extended to access the site.

3. Homepage returns only 13 words of visible text without JavaScript High

   A plain GET of the homepage yields only 13 words of visible text; the bulk of the content is loaded via JavaScript (lazy-loaded sections, WP Rocket deferred scripts). Non-executing crawlers see essentially nothing, while JS-capable bots (ChatGPT-User, PerplexityBot) see the full page.

   What to change: Implement server-side rendering or dynamic rendering for the homepage so that static HTML contains the core content. Alternatively, ensure critical text is not hidden behind JavaScript execution.

4. No Organization schema on the .com homepage High

   The .com homepage lacks any Organization schema, while the .com.au version includes one with logo. No LocalBusiness, FAQPage, Product, or Service schema is present anywhere, despite FAQ sections and dozens of service offerings.

   What to change: Add Organization schema to the .com homepage with correct name, logo, description, and industry. Add Service schema for each core offering and FAQPage schema for FAQ sections.

5. No llms.txt file published Medium

   The site returns a 404 for /llms.txt, missing an opportunity to provide AI models with a curated, authoritative summary of the company's offerings and key pages.

   What to change: Create an llms.txt file that describes the company as an offshore staffing provider and lists key pages (services, case studies, blog).

6. No independent third-party validation found in search results Medium

   Searches for reviews, Reddit discussions, and press coverage about Satellite Office as an offshoring company returned zero results. The site claims 160+ clients and ISO 27001 certification, but none of this appears in external sources, leaving AI models with no corroborating signals.

   What to change: Encourage clients to leave reviews on platforms like G2, Clutch, or Trustpilot. Publish press releases and case studies on third-party sites. Build a presence on relevant forums and directories.

7. Pricing page returns 404 Medium

   The /pricing/ URL returns a 404 error, which may frustrate both human visitors and AI crawlers looking for pricing information.

   What to change: Either restore the pricing page or redirect /pricing/ to a relevant page (e.g., services or contact).

8. Robots.txt lacks AI-bot directives Low

   The robots.txt only disallows /wp-admin/ and does not explicitly allow or disallow any AI crawlers. Combined with the server-side blocking, this creates an unclear crawl policy.

   What to change: Add explicit Allow directives for GPTBot, ClaudeBot, and Google-Extended to ensure they are not blocked by server configuration.

What's working

• PerplexityBot and ChatGPT-User receive full page content — PerplexityBot and ChatGPT-User successfully retrieve the full homepage (825KB) with a 200 status, ensuring AI search products can index the site's content.
• Australian domain includes Organization schema — The .com.au homepage includes an Organization schema with logo, providing structured data for AI models on that domain.
• Blog articles carry Article schema with author and date — Blog posts include Article schema with author, datePublished, and image, which helps AI models correctly attribute and date content.
• Sitemap contains 80 indexed URLs — The sitemap is accessible and lists 80 URLs, providing a clear crawl path for search engines and AI crawlers.
• Leadership team page lists 18 executives with LinkedIn profiles — The /our-leadership-team/ page provides detailed bios for 18 executives, offering authoritative information about company leadership.
• Case studies page provides 906 words of client success stories — The /case-studies/ page contains detailed client success stories, which can serve as authoritative content for AI models.
• Services page contains 1,381 words of detailed offerings — The /services/ page provides comprehensive descriptions of offshore staffing solutions, giving AI models substantial content to index.
• About page contains 1,253 words of company information — The /about/ page provides detailed company background, values, and differentiators, offering rich content for AI understanding.