workhuman.com — AI Site Grade

Analysis

Workhuman's AI-visibility posture is unusually mature — but a critical schema gap undermines how AI engines interpret its content

The site has a full llms.txt (23,667 bytes, last updated February 2026), an llm-info/ page explicitly allowed for all major AI crawlers, and a robots.txt that grants GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and ChatGPT-User unrestricted access. Every AI bot tested returns a 200 with identical byte payload (974,654 bytes) — no UA-based blocking, no JS shell, no Vercel WAF discrimination. This is best-in-class technical AI-readiness.

Crawler Access

Hosted on Vercel with HSTS and a strict CSP, the site serves all AI crawlers the same full HTML as a browser. The robots.txt explicitly allows GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and ChatGPT-User to crawl / and /llm-info/. No AI bot is disallowed from any content path. The llms.txt at /llms.txt is a comprehensive 23KB document covering company overview, product suite, governance posture, and verifiable claims — one of the most thorough llms.txt files observed in the wild. A dedicated /llm-info/ page mirrors this content in HTML format, explicitly positioned as "the authoritative source for product definitions" for AI assistants.

Cold-Knowledge Gap

The LLM prior on Workhuman is dated and thin. It knows the company was founded as Globoforce in 1999, rebranded in 2019, and hosts a "Workhuman Live" conference. It mentions "Workhuman Cloud" as a product name — a term that no longer appears anywhere on the live site. The site now brands its platform around "Human Intelligence," "AI Assistant," "Recognition Advisor," and "Culture Hub." The cold model knows nothing about the ROI Guarantee, the $1B+ rewards marketplace, the Inclusion Advisor DEI tool, or the "Human Intelligence 2025 Release." The Wikipedia page (last modified May 2026) is also stale — it still describes the company as "human capital management software" and lists customers like Symantec and Dell, while the site now emphasizes AI-native recognition intelligence.

Schema Posture

Despite excellent AI-crawl access, the site has a critical structured-data deficiency. The homepage and all key pages carry only a Corporation schema type with logo, address, and social links. No SoftwareApplication, Product, FAQPage, HowTo, or WebSite schema with potentialAction (site search) is present. The blog listing page has only a BreadcrumbList. The /pricing-advantage/ page is marked noindex,nofollow. The /human-intelligence/ page — the flagship product narrative — has no product schema at all. This means AI engines consuming the site's rich HTML content get no structured signals about what the software does, its features, or its pricing model.

External Signals

The Wikipedia article is the primary off-domain signal shaping AI knowledge, and it is significantly outdated: it still uses the Globoforce origin story, lists old customers, and categorizes the company as generic HCM software rather than an AI-native recognition platform. No recent press, Gartner quadrant mentions, or Reddit discussions surfaced in search. The blog contains comparison guides ("Workhuman vs. Vantage Circle," "Workhuman vs. Reward Gateway") but one fetched guide returned a 404, suggesting content maintenance gaps. The blog does publish fresh content (2026 dates visible), but the comparison pages may be broken links that AI crawlers will index as dead ends.

Findings

1. Key pages lack SoftwareApplication and Product schema High

   The homepage and flagship product pages (e.g., /human-intelligence/) carry only Corporation schema. No SoftwareApplication, Product, FAQPage, or WebSite schema with potentialAction is present, preventing AI engines from understanding the software's features and pricing.

   What to change: Add SoftwareApplication schema to /human-intelligence/ and Product schema to /pricing-advantage/. Add WebSite schema with SearchAction to the homepage.

2. Wikipedia article is outdated and misrepresents the company High

   The Wikipedia page (last modified May 2026) still describes Workhuman as 'human capital management software' and lists old customers like Symantec and Dell. It does not reflect the current AI-native recognition platform branding, which shapes AI knowledge bases.

   What to change: Update the Wikipedia article to reflect the current product suite (Human Intelligence, AI Assistant, etc.) and remove outdated customer references.

3. LLM prior knowledge is thin and uses obsolete product names High

   The LLM prior knows Workhuman as a rebranded Globoforce and mentions 'Workhuman Cloud' — a term no longer on the site. It is unaware of the ROI Guarantee, $1B+ rewards marketplace, Inclusion Advisor, and Human Intelligence 2025 Release.

   What to change: Ensure the llms.txt and /llm-info/ pages are indexed and cited by AI crawlers. Consider submitting the site to AI training data sources.

4. Pricing page is marked noindex, limiting AI visibility Medium

   The /pricing-advantage/ page has a noindex,nofollow meta tag, preventing AI crawlers from indexing its content. This hides the ROI guarantee and pricing details from AI engines.

   What to change: Remove the noindex directive from /pricing-advantage/ to allow AI crawlers to index the page.

5. Blog comparison guide returns 404 Medium

   The blog post 'Workhuman vs. Vantage Circle 2026 Comparison Guide' returns a 404 error, indicating a broken link that AI crawlers will index as a dead end.

   What to change: Restore the comparison guide or set up a 301 redirect to a relevant page.

6. No FAQPage schema on pages with Q&A content Medium

   Pages that contain FAQ-style content (e.g., /human-intelligence/) do not use FAQPage schema, missing an opportunity to generate rich results in AI responses.

   What to change: Add FAQPage schema to pages with question-and-answer content.

7. WebSite schema missing potentialAction for site search Low

   The homepage lacks WebSite schema with SearchAction, which would enable AI engines to perform site search directly.

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

8. Blog listing page has minimal schema Low

   The blog listing page only has BreadcrumbList schema, missing BlogPosting or CollectionPage schema that could help AI engines understand the blog structure.

   What to change: Add BlogPosting schema to individual blog posts and CollectionPage schema to the blog listing.

What's working

• Comprehensive llms.txt file (23KB) with product and governance info — The /llms.txt file is one of the most thorough observed, covering company overview, product suite, governance posture, and verifiable claims. It is regularly updated (last modified February 2026).
• Dedicated /llm-info/ page as authoritative source for AI assistants — The /llm-info/ page mirrors the llms.txt content in HTML and is explicitly positioned as the authoritative source for product definitions for AI assistants.
• All major AI crawlers allowed unrestricted access — robots.txt explicitly allows GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and ChatGPT-User to crawl the entire site. No AI bot is disallowed from any content path.
• All AI bots receive identical full HTML payload — Every AI bot tested returns a 200 status with the same byte payload (974,654 bytes), indicating no UA-based blocking, JS shell, or Vercel WAF discrimination.
• Flagship product page has rich, detailed content — The /human-intelligence/ page contains 4,392 words of detailed product narrative, providing AI engines with substantial text to understand the platform.
• Blog publishes fresh content with 2026 dates — The blog contains recent posts, indicating ongoing content creation that can attract AI crawlers and provide up-to-date information.
• Customers page is accessible and lists social proof — The /company/our-customers/ page is accessible to AI crawlers and provides social proof, which can be used by AI engines to validate the company's credibility.
• Platform page describes product suite in detail — The /platform/ page provides a comprehensive overview of the product suite, helping AI engines understand the range of offerings.