pensioncorporation.com — AI Site Grade

Analysis

The live site is entirely inaccessible to AI crawlers — every page, every bot, every path returns a 1,922-byte Akamai cryptographic challenge shell with zero readable content.

Crawler Access

The entire domain pensioncorporation.com sits behind an Akamai WAF that serves a JavaScript-based cryptographic challenge page (Challenge Validation) to every request. All eleven tested user-agents — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended, ChatGPT-User, Perplexity-User, anthropic-ai, and a standard browser — receive the identical 1,922-byte HTML shell with zero visible text, no <meta> tags, no JSON-LD, no headings, and no links to real content. The robots.txt and llms.txt endpoints both return HTTP 403 Access Denied with the same Akamai error fingerprint (0.e847b5c.*). The sitemap.xml also returns the challenge page, not an actual sitemap. DNS records confirm the site runs on Azure DNS with Akamai edge infrastructure (via akamai-cache-status headers).

Cold-Knowledge Gap

A frontier LLM queried cold describes Pension Corporation as a UK-based specialist insurer founded in 2006, offering bulk purchase annuities (buy-ins and buyouts), regulated by the PRA and FCA, and notes a £1.6 billion ICI Pension Fund transaction in 2023. The model knows the brand exists and its core business. However, the live site provides zero of this information to any AI crawler. The model's knowledge comes entirely from pre-training data and third-party sources — not from the site itself. The site's actual content (visible via Wayback Machine) reveals the company now brands as Pension Insurance Corporation (PIC), reports £51.5bn portfolio and 399,200 policyholders (as of June 2025), and has been acquired by Athora Group — a major corporate event the cold model did not mention.

Schema Posture

The live site contains zero structured data. No JSON-LD of any type — no Organization, FinancialService, FAQPage, Product, or WebSite schema — is present on any page tested. The Wayback snapshots of the real site also show no JSON-LD on the homepage, about page, or trustees page. The only semantic signals are HTML headings (H1: About us., H1: Trustees.) and basic meta descriptions. For a regulated financial institution managing tens of billions in pension assets, the absence of Organization schema with FCA/PRA registration numbers, FinancialService schema, or FAQPage schema for the buy-in vs. buyout comparison is a significant gap.

External Signals

The domain pensioncorporation.com has zero indexed pages in web search results — no search engine has crawled or indexed any live content. The Wayback Machine holds 1,445 captures spanning 2007–2026, confirming the site has existed for nearly two decades with substantial content. The real site (via Wayback) shows the company has been rebranded under Athora Group, with navigation links to "Athora to acquire Pension Insurance Corporation Group" and "About Athora" sections. The site's <title> tag reads "Pension Insurance Corporation - PIC" and the og:site_name is "pic Corporate" — a brand name mismatch with the domain pensioncorporation.com. The cold LLM knowledge is stale: it describes the pre-acquisition, pre-rebrand entity and misses the Athora acquisition entirely.

Findings

1. Akamai cryptographic challenge blocks all AI crawlers High

   Every page on pensioncorporation.com returns a 1,922-byte JavaScript challenge shell with zero readable content to all tested user-agents, including GPTBot, ClaudeBot, and Google-Extended. The robots.txt and llms.txt endpoints return HTTP 403.

   What to change: Remove or bypass the Akamai challenge for known AI crawler user-agents, or serve a static HTML version of the site to those bots.

2. Robots.txt and llms.txt return HTTP 403 High

   The robots.txt and llms.txt endpoints are blocked by the Akamai WAF, returning 403 Access Denied. This prevents crawlers from discovering allowed paths and AI-specific guidance.

   What to change: Serve a standard robots.txt and llms.txt that are accessible to all crawlers, and allow AI bots to read them.

3. Sitemap.xml returns challenge page instead of actual sitemap High

   The sitemap.xml endpoint returns the Akamai challenge page, not an XML sitemap. This prevents crawlers from discovering the site's URL structure.

   What to change: Serve a valid sitemap.xml that lists all public pages, and ensure it is accessible to crawlers.

4. Zero indexed pages in web search results High

   A site: search for pensioncorporation.com returns zero results, indicating no search engine has indexed any live content from the domain.

   What to change: Remove the Akamai challenge for search engine bots to allow indexing.

5. No structured data (JSON-LD) on any tested page High

   The live site contains zero JSON-LD structured data. Wayback snapshots of the real site also show no JSON-LD. Missing Organization, FinancialService, and FAQPage schemas for a regulated financial institution.

   What to change: Add Organization schema with FCA/PRA registration numbers, FinancialService schema for buy-in/buyout products, and FAQPage schema for common trustee questions.

6. Cold LLM knowledge is stale and misses Athora acquisition Medium

   A frontier LLM queried cold describes Pension Corporation as a pre-acquisition entity, missing the rebrand to Pension Insurance Corporation (PIC) and the acquisition by Athora Group. The live site provides no information to correct this.

   What to change: Ensure the live site is accessible to AI crawlers so that current information (rebrand, acquisition) is available for model training and retrieval.

7. Brand name mismatch between domain and site content Low

   The domain pensioncorporation.com hosts content for Pension Insurance Corporation (PIC), with og:site_name 'pic Corporate'. This mismatch may confuse crawlers and users.

   What to change: Consider aligning the domain with the current brand (PIC) or use canonical tags to clarify the relationship.

What's working

• Wayback Machine shows nearly two decades of rich content — The Wayback Machine holds 1,445 captures from 2007–2026, confirming the site has historically contained substantial content about PIC's services, trustees, and corporate information.
• Real site content visible via Wayback Machine — Wayback snapshots reveal a fully functional site with navigation, headings, and detailed pages about buy-in/buyout services, trustees, and corporate information.
• Basic HTML headings present on real site pages — Wayback snapshots show H1 headings like 'About us.' and 'Trustees.' providing basic semantic structure.