aviva.ca — AI Site Grade

Analysis

Aviva.ca — AI-Visibility Audit

Every major AI crawler (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Applebot-Extended, Bytespider) receives a 403 Access Denied from the Akamai edge, while the same page serves 139KB of full HTML to browsers and the anthropic-ai user-agent. This is not a robots.txt block — the robots.txt is permissive (allows /, no AI-bot directives) — but the WAF/CDN layer (Akamai, via errors.edgesuite.net) actively rejects AI crawlers at the network edge, making the entire site invisible to the models that power ChatGPT, Google AI Overviews, Perplexity, and Apple Intelligence.

Crawler Access

compare_bot_access on the homepage returned 403 for 9 of 11 tested user-agents. Only Browser and anthropic-ai received 200 responses with full content. The 403 pages are 373-byte Akamai-denial shells with no content. The robots.txt contains no User-agent: GPTBot, Disallow: / for AI bots, or any AI-specific rules — the block is entirely at the CDN/WAF layer, not signalled in the crawl contract. llms.txt returns 404. The sitemap.xml exists with 1,883 URLs and is well-formed, but most AI crawlers cannot reach it.

Cold-Knowledge Gap

The LLM queried cold about "Aviva Canada insurance" recalled the brand as a major P&C insurer, mentioned the "Aviva Advance" product, the "Aviva Drive" telematics app, and a 2023 data breach with class-action scrutiny. The actual site makes no mention of "Aviva Advance" anywhere on the homepage, auto page, or home page. The telematics product is called "Aviva Journey" (not "Aviva Drive") — the model's product name is stale. The site also makes no mention of any data breach or class action in its public-facing content, fraud hub, or FAQ. The model's prior knowledge of a reputational vulnerability has no counterpart on the live site, creating a gap where AI-generated summaries could surface negative history the brand does not address.

Schema Posture

The homepage and the business cyber page contain zero JSON-LD schema. The auto insurance page and home insurance page each contain a single FAQPage schema with well-structured Q&A — this is the only structured data present across all pages checked. No Organization, Product, InsuranceProduct, WebSite, BreadcrumbList, or LocalBusiness schema exists anywhere. The blog has no schema at all. The Aviva Journey page has a FAQPage schema but no MobileApplication or SoftwareApplication schema for the app itself.

Content & Navigation Fragmentation

The navigation links to /en/tools-and-resources/blog/ but that URL returns a 404 (the actual blog lives at /en/blog/). The FAQ link at /en/tools-and-resources/faq/ also 404s. The "About us" link at /en/about-us/ 404s. The "Tools and resources" hub at /en/tools-and-resources/ 404s. These are navigation dead ends that appear in the global header — any AI crawler that does get through (e.g., anthropic-ai) will hit broken paths. The blog itself is active with recent posts (May 2026) and good content, but has no schema markup and no blog-specific JSON-LD.

Findings

1. All major AI crawlers blocked by Akamai WAF with 403 errors High

   GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Applebot-Extended, and Bytespider all receive 403 Access Denied from the Akamai edge, while browsers and anthropic-ai get full HTML. The robots.txt is permissive with no AI-bot directives, so the block is invisible to crawlers until they hit the CDN.

   What to change: Allow AI crawlers through the Akamai WAF by whitelisting their user-agent strings, or serve a static HTML snapshot to blocked bots.

2. llms.txt file returns 404 Medium

   The standard llms.txt file, which helps AI crawlers discover key content, is not present on aviva.ca.

   What to change: Create an llms.txt file listing the most important pages for AI consumption.

3. LLM cold knowledge uses stale product name 'Aviva Drive' instead of 'Aviva Journey' Medium

   When queried cold, the LLM recalled the telematics product as 'Aviva Drive', but the actual site calls it 'Aviva Journey'. This mismatch means AI-generated summaries may reference a non-existent product name.

   What to change: Ensure the site prominently uses the correct product name 'Aviva Journey' and consider adding schema markup for the mobile application.

4. LLM cold knowledge includes unaddressed data breach and class action Medium

   The LLM recalled a 2023 data breach and class-action scrutiny, but the aviva.ca site makes no mention of any breach, fraud resources, or remediation. This creates a gap where AI summaries could surface negative history without the brand's response.

   What to change: Publish a dedicated page or FAQ addressing the data breach, security measures, and class-action status to provide authoritative context for AI crawlers.

5. Homepage and business cyber page lack any JSON-LD schema High

   The homepage and the Aviva cyber insurance for business page contain zero structured data. Without Organization, WebSite, or Product schema, AI crawlers cannot confidently extract entity information.

   What to change: Add Organization, WebSite, and InsuranceProduct JSON-LD schema to the homepage and product pages.

6. Blog pages have no structured data markup Medium

   The blog index and individual blog posts lack any JSON-LD schema, such as BlogPosting or Article, reducing their visibility in AI-generated summaries and rich results.

   What to change: Add BlogPosting or Article schema to all blog pages.

7. Aviva Journey page missing MobileApplication schema Low

   The Aviva Journey page has a FAQPage schema but no SoftwareApplication or MobileApplication schema for the telematics app, making it harder for AI to understand the app's functionality.

   What to change: Add MobileApplication or SoftwareApplication schema to the Journey page.

8. Multiple navigation links return 404 errors High

   The global header links to /en/tools-and-resources/blog/, /en/tools-and-resources/faq/, /en/about-us/, and /en/tools-and-resources/ all return 404 pages. This creates dead ends for any crawler that gets through, wasting crawl budget and degrading user experience.

   What to change: Fix or redirect all broken navigation links to their correct destinations (e.g., /en/blog/ instead of /en/tools-and-resources/blog/).

9. Robots.txt lacks any AI-bot directives Low

   The robots.txt file contains no User-agent lines for GPTBot, ClaudeBot, or other AI crawlers, and no disallow rules for AI bots. While the CDN blocks them anyway, the lack of explicit directives means crawlers waste time attempting to fetch.

   What to change: Add explicit User-agent and Disallow rules for AI crawlers in robots.txt to signal the block at the crawl stage.

10. Sitemap inaccessible to most AI crawlers due to CDN block High

    The sitemap.xml exists with 1,883 URLs, but since AI crawlers are blocked at the CDN, they cannot discover the sitemap or the URLs within it.

    What to change: Allow AI crawlers through the CDN to access the sitemap, or serve a separate sitemap for AI bots.

11. No Organization schema present on any checked page Medium

    None of the audited pages include Organization JSON-LD, which helps AI crawlers confirm brand identity, logo, and social profiles.

    What to change: Add Organization schema to the homepage and all major landing pages.

12. No BreadcrumbList schema on any page Low

    BreadcrumbList schema is absent from all checked pages, reducing the clarity of site structure for AI crawlers.

    What to change: Add BreadcrumbList schema to all pages with breadcrumb navigation.

What's working

• Robots.txt allows all crawlers with no disallow rules — The robots.txt file is permissive, allowing all user-agents to access the entire site. This is a good foundation for AI visibility once the CDN block is resolved.
• Auto and home insurance pages have FAQPage schema — The personal auto insurance and home insurance pages each contain a well-structured FAQPage schema with relevant Q&A, which can power rich results and AI answers.
• Blog is active with recent, relevant posts — The blog at /en/blog/ contains recent posts (e.g., May 2026) with substantive content about telematics and safe driving, providing fresh material for AI crawlers.
• Sitemap exists with 1,883 URLs and is well-formed — The sitemap.xml is accessible (to browsers) and contains a comprehensive list of URLs, which is a strong signal for crawl efficiency once AI crawlers are allowed.
• Anthropic-ai user-agent is allowed through the CDN — The anthropic-ai crawler receives a 200 response with full HTML, meaning Claude can index the site. This is a positive signal for AI visibility from Anthropic's models.
• French version of the site is available and well-structured — The French homepage at /fr/ serves full content, supporting bilingual AI visibility for Canadian audiences.
• Aviva Journey page includes FAQPage schema — The Journey page has a FAQPage schema with questions about the telematics program, aiding AI understanding of the product.
• No AI-bot disallow rules in robots.txt — The absence of AI-bot disallow rules means the site does not intentionally block AI crawlers via robots.txt, making it easier to grant access once the CDN issue is fixed.