employmenthero.com — AI Site Grade

Analysis

Employment Hero — AI-Visibility Audit

The site has an exemplary llms.txt (139KB, the largest observed in any audit to date) and zero AI-bot blocking across all major crawlers, yet the cold LLM knowledge of the brand is stuck on a 2021-era description — missing the entire "AI Employment Operating System" repositioning and HeroForce EOR product that the site now leads with.

Crawler Access

All 11 AI crawlers tested (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Bytespider, Applebot-Extended, anthropic-ai, Perplexity-User, and a browser baseline) receive a 200 status with identical 502KB payload from employmenthero.com. No UA-based blocking, no Cloudflare challenge, no JS shell. The robots.txt has a Crawl-delay: 10 directive and a * rule disallowing /wp-admin/, /wp-json/, /bk/, and zip files — but no AI-bot-specific rules at all. No GPTBot, ClaudeBot, or Google-Extended section exists. The site runs on nginx behind AWS CloudFront, with PHP 8.3 and Jetpack Boost caching. The anthropic-domain-verification and openai-domain-verification TXT records are present, confirming the brand has proactively verified with both Anthropic and OpenAI for API/retrieval use.

llms.txt and Content Signals

The /llms.txt at 139KB is a standout — it contains a comprehensive, well-structured directory of every product, solution, service, industry page, and resource, each with a one-line description and URL. This is the strongest llms.txt seen in this audit series. The homepage carries FAQPage schema with 12 Q&A pairs covering core products (HeroForce, Employment OS, AI agents). The /ai/ page and /products/heroforce/ page also carry FAQPage and SoftwareApplication schema. However, the homepage's WebPage schema has a dateModified of 2026-05-19 — a future date — which may confuse crawlers about content freshness. The blog contains stale posts from 2019 (Series C announcement) alongside fresh February 2026 press releases, creating a mixed freshness signal.

Cold-Knowledge Gap

The LLM's prior knowledge describes Employment Hero as "an Australian-based HR, payroll, and benefits platform for SMBs" with products called "HeroPay" and "HeroHR" — names that no longer exist on the site. The model references a $1.1B valuation from 2021, a Series E round, and a data breach reputational signal. The site today positions itself as the "world's first AI Employment Operating System" with HeroForce (Employer of Record), Recruitment Agent, Payroll Agent, and HR Agent as AI-powered products. The cold knowledge contains zero mention of HeroForce, the AI agents, the Employment OS framing, or the 180-country EOR capability. The gap is structural: the LLM knows the brand as a legacy HR SaaS tool, not as the AI-platform/EOR hybrid the site now sells.

External Signals

The site prominently displays 4.5/5 on Trustpilot, 4.4/5 on Capterra, and 4.3/5 on G2, with a "G2 2026 Best Software Awards Top 10 in ANZ" badge. DNS records show integrations with Stripe, Drift, Mixpanel, Pendo, Jamf, and Canva — a mature SaaS stack. The blog references a "Series F funding announcement" and "unicorn status and acquisition of KeyPay," but the cold LLM only knows up to Series E. The brand's external reputation is current and strong, but the AI models' training data cutoff means they describe a company that no longer exists in its current form.

Findings

1. No AI-bot rules in robots.txt — The file has a catch-all * section but zero explicit directives for GPTBot, ClaudeBot, Google-Extended, or any AI crawler. While nothing is blocked, the absence means AI crawlers rely on the generic Crawl-delay: 10, which may slow down indexing of the 1,255+ sitemap URLs.

2. Future-dated schema dateModified — The homepage WebPage schema lists dateModified: 2026-05-19. This is a future date at time of audit and may trigger freshness skepticism from crawlers that validate timestamps.

3. Cold LLM knowledge is 3-4 years stale — The model describes "HeroPay" and "HeroHR" products that do not exist on the site, a $1.1B valuation from 2021, and a Series E round. The site now sells an "AI Employment Operating System" with HeroForce EOR, Recruitment Agent, Payroll Agent, and HR Agent — none of which appear in the model's prior.

4. llms.txt is best-in-class — At 139KB with structured product, solution, service, industry, and resource sections, this is the most comprehensive llms.txt encountered. It should be referenced as a benchmark.

5. Stale blog content mixed with fresh — The sitemap includes a 2019 Series C funding post alongside February 2026 press releases. The 2019 post has been updated (dateModified: 2026-03-02) but still references "3,500 businesses" — the homepage now claims "350k+ happy customers," a 100x discrepancy that could confuse crawlers doing entity resolution.

6. No pricing page in sitemap — The FAQ references pricing plans ("HR Essentials", "HR Engage", "HR Elite", "Employment Unlimited") but no dedicated pricing page was found in the sitemap or URL sample. AI crawlers seeking to answer "how much does Employment Hero cost?" would need to infer from FAQ schema alone.

Findings

1. No AI-bot rules in robots.txt Medium

   The robots.txt file has a catch-all '*' section but zero explicit directives for GPTBot, ClaudeBot, Google-Extended, or any AI crawler. While nothing is blocked, the absence means AI crawlers rely on a generic Crawl-delay: 10, which may slow indexing of over 1,255 sitemap URLs.

   What to change: Add explicit directives for GPTBot, ClaudeBot, Google-Extended, and other AI crawlers, either allowing them with a reasonable crawl-delay or removing the generic delay to speed up indexing.

2. Future-dated schema dateModified on homepage Medium

   The homepage WebPage schema lists dateModified as 2026-05-19, a future date at the time of audit. This may trigger freshness skepticism from crawlers that validate timestamps.

   What to change: Update the dateModified to the actual last modification date or remove it if not accurately maintained.

3. Cold LLM knowledge is 3-4 years stale High

   The LLM's prior knowledge describes Employment Hero as an Australian HR platform for SMBs with products 'HeroPay' and 'HeroHR' that no longer exist, a $1.1B valuation from 2021, and a Series E round. The site now sells an 'AI Employment Operating System' with HeroForce EOR and AI agents, none of which appear in the model's prior.

   What to change: Publish a dedicated AI knowledge page or update the llms.txt with a comprehensive brand summary that includes the current product names, valuation, and AI positioning. Consider submitting to LLM knowledge bases or using structured data to reinforce the new narrative.

4. Stale blog content mixed with fresh posts Medium

   The sitemap includes a 2019 Series C funding post alongside February 2026 press releases. The 2019 post has been updated (dateModified: 2026-03-02) but still references '3,500 businesses' while the homepage claims '350k+ happy customers', a 100x discrepancy that could confuse crawlers doing entity resolution.

   What to change: Update or remove the outdated blog post, or add a prominent note that the numbers have changed. Ensure all blog content reflects current metrics.

5. No dedicated pricing page in sitemap Medium

   The FAQ references pricing plans ('HR Essentials', 'HR Engage', 'HR Elite', 'Employment Unlimited') but no dedicated pricing page was found in the sitemap or URL sample. AI crawlers seeking to answer 'how much does Employment Hero cost?' would need to infer from FAQ schema alone.

   What to change: Create a dedicated pricing page with clear tiers and include it in the sitemap and llms.txt.

What's working

• Best-in-class llms.txt at 139KB — The /llms.txt file is 139KB with a comprehensive, well-structured directory of every product, solution, service, industry page, and resource, each with a one-line description and URL. This is the strongest llms.txt encountered in the audit series.
• Zero AI-bot blocking across all major crawlers — All 11 AI crawlers tested receive a 200 status with identical payload from the homepage. No UA-based blocking, no Cloudflare challenge, no JS shell. The site is fully accessible to AI crawlers.
• Anthropic and OpenAI domain verification TXT records present — The DNS TXT records include anthropic-domain-verification and openai-domain-verification, confirming proactive verification with both Anthropic and OpenAI for API/retrieval use.
• FAQPage schema on homepage, /ai/, and /products/heroforce/ — The homepage carries FAQPage schema with 12 Q&A pairs covering core products. The /ai/ and /products/heroforce/ pages also carry FAQPage and SoftwareApplication schema, helping AI crawlers understand product offerings.
• Strong external reputation signals with high ratings — The site displays 4.5/5 on Trustpilot, 4.4/5 on Capterra, and 4.3/5 on G2, with a 'G2 2026 Best Software Awards Top 10 in ANZ' badge. DNS records show integrations with Stripe, Drift, Mixpanel, Pendo, Jamf, and Canva, indicating a mature SaaS stack.