eversanaintouch.com — AI Site Grade

Analysis

I have enough data to write a thorough audit. Let me compile my findings.

GPTBot Blocked, Blog Archive Rot, and a Schema Desert

The site's most consequential AI-visibility problem is that GPTBot receives a 429 (rate-limited/blocked) on the homepage while every other major AI crawler — ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User — gets a full 200 with identical content. This selective blocking means OpenAI's models (ChatGPT, GPT search) cannot read the site's positioning, services, or case studies directly. The robots.txt at https://www.eversanaintouch.com/robots.txt contains no AI-bot directives at all — just a generic User-agent: * rule disallowing /wp-admin/ and a 10-second crawl delay. No llms.txt exists (404).

Crawler Access

All non-GPTBot AI crawlers return the same 48KB HTML payload as a standard browser — no JS-shell, no UA-based gating. The site runs on nginx behind WP Engine (revealed by x-powered-by: WP Engine header). Security headers are absent: no Strict-Transport-Security, no Content-Security-Policy, no X-Frame-Options. The cache-control: max-age=600, must-revalidate is reasonable but means crawlers get a fresh page every 10 minutes. Bytespider (ByteDance) errors out entirely (status 0), suggesting a network-level block or timeout.

Cold-Knowledge Gap

The LLM queried cold describes EVERSANA INTOUCH as a "healthcare marketing agency" formed through a 2021 merger, mentioning an "InTouch platform suite" and "patient support programs." The actual site tells a different story: it positions itself as "A Full-Service Agency Network" with an "AI-powered operating model" and a tagline of "A V RTFUL I NGENUITY" (intentional spacing). The cold model knows nothing about the AI agency platform, the "Launch to Leadership" messaging, the 10x Agency of the Year claims, the Mike Guarino CEO appointment, or the Healthware/Med Comm/Market Access sub-brands. The model also hallucinates a product suite ("InTouch platform") that does not appear anywhere on the live site.

Schema Posture

Every page audited carries only a single WebPage schema type with audience: "Pharmaceutical companies" and a potentialAction of ReadAction. There is no Organization schema, no LocalBusiness, no Service, no FAQPage, no Article schema on blog posts, no BreadcrumbList, no Review schema. The JSON-LD is minimal and identical in structure across all pages — just a name, description, and logo reference. The blog index page (/blog/) has zero JSON-LD at all. Given the site's heavy emphasis on awards, case studies, and services, the absence of structured data for those entities is a significant gap for AI-driven search and knowledge panels.

Content Archaeology

The sitemap reveals 989+ URLs, the vast majority being blog posts. A random sample shows posts dating back to 2016-2017 (e.g., "How Railroads Changed Healthcare" from December 2017). The blog is a mix of genuinely fresh content (May 2026 posts about AI Overviews, Adobe Summit) and years-old articles about Facebook algorithm changes, Google interstitials, and iOS 10 email updates. This creates a stale-content tail that AI crawlers indexing the full sitemap will encounter — outdated references that contradict the brand's forward-looking AI narrative. The /blog/ hub page itself is thin (29 words of visible text) and loads as a JS-dependent shell, though individual posts render full HTML.

External Signals

Web searches returned zero results for queries about EVERSANA INTOUCH reviews, client feedback, or recent press — a striking absence for a 7,000-employee agency claiming 10 Agency of the Year wins. The site's external link profile is limited to parent company eversana.com, social profiles (Facebook, X, Instagram, LinkedIn), and a Pardot-hosted gated content domain (go.eversanaintouch.com). DNS records show Pardot (Salesforce) for email marketing. No third-party review sites, Glassdoor pages, or industry analyst mentions surfaced in search.

Findings

1. GPTBot receives 429 rate-limited response on homepage High

   GPTBot is blocked with a 429 status on the homepage, while all other major AI crawlers (ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User) receive a full 200 response. This prevents OpenAI models from reading the site's content.

   What to change: Remove the rate-limiting or block on GPTBot, or ensure GPTBot receives the same HTML content as other crawlers.

2. Robots.txt lacks AI-bot directives Medium

   The robots.txt file contains no rules for AI crawlers like GPTBot, ClaudeBot, or Google-Extended, only a generic User-agent: * rule disallowing /wp-admin/ and a 10-second crawl delay.

   What to change: Add explicit directives for AI crawlers, e.g., allow GPTBot and other AI bots to access the site.

3. No llms.txt file published Low

   The site does not provide an llms.txt file, which is a recommended standard for guiding AI crawlers to key content.

   What to change: Create an llms.txt file listing important pages for AI crawlers.

4. Bytespider (ByteDance) fails to access the site Medium

   Bytespider receives a status 0 error, indicating a network-level block or timeout, preventing ByteDance's AI from indexing the site.

   What to change: Investigate and resolve the network block or timeout for Bytespider.

5. Only WebPage schema present across the site High

   Every page audited contains only a single WebPage schema type with audience and potentialAction. No Organization, Service, FAQPage, Article, BreadcrumbList, or Review schema exists, missing opportunities for rich results and knowledge panels.

   What to change: Add Organization schema to the homepage, Service schema to services pages, Article schema to blog posts, and BreadcrumbList to all pages.

6. Blog index page has zero JSON-LD Medium

   The /blog/ page lacks any structured data, making it invisible to AI-driven search features that rely on schema.

   What to change: Add a CollectionPage or BlogPosting schema to the blog index.

7. Blog contains years-old posts contradicting AI-forward narrative Medium

   The sitemap includes blog posts from 2016-2017 about outdated topics (Facebook algorithm changes, iOS 10 email updates), creating a stale-content tail that AI crawlers may index, undermining the brand's modern AI positioning.

   What to change: Remove or noindex outdated blog posts, or update them to reflect current relevance.

8. Blog hub page is content-thin with only 29 words Medium

   The /blog/ page contains only 29 words of visible text and appears to rely on JavaScript for full content, which may hinder crawlers that do not execute JS.

   What to change: Add substantive text content to the blog hub page and ensure it renders without JavaScript.

9. Zero external review or press results in web search Medium

   Web searches for agency reviews, client feedback, and recent press returned no results, indicating a lack of external signals that AI models use for authority assessment.

   What to change: Encourage client reviews on platforms like G2 or Clutch, and pursue industry press coverage.

10. LLM cold knowledge hallucinates a product suite not on the site High

    When queried cold, the LLM describes an 'InTouch platform suite' and 'patient support programs' that do not appear on the live site, while missing the actual AI-powered operating model and agency network positioning.

    What to change: Ensure the site clearly and consistently communicates its core offerings and unique value proposition to align with AI training data.

11. Security headers absent (HSTS, CSP, X-Frame-Options) Low

    The site lacks Strict-Transport-Security, Content-Security-Policy, and X-Frame-Options headers, which are not directly an AI visibility issue but indicate low security posture that may affect trust signals.

    What to change: Add HSTS, CSP, and X-Frame-Options headers to improve security and trust signals.

What's working

• All non-GPTBot AI crawlers receive full HTML content — ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, and ChatGPT-User all receive the same 48KB HTML payload as a standard browser, with no JavaScript shell or user-agent gating.
• Sitemap is accessible and contains 989+ URLs — The WordPress sitemap at /wp-sitemap.xml is properly indexed and includes a large number of URLs, helping crawlers discover content.
• Individual blog posts render full HTML content — Unlike the blog hub, individual blog posts like the Railroads article contain substantial text (876 words) and are fully accessible to crawlers.
• Recent blog posts cover current AI and pharma topics — Posts from 2026 about AI Overviews and Adobe Summit demonstrate ongoing content creation relevant to the agency's AI focus.
• Awards page contains detailed content (565 words) — The /work/awards/ page provides substantial text about corporate and creative excellence awards, which can support authority signals.
• Leadership page has rich biographical content — The /about/leadership/ page contains over 1000 words of team member information, providing depth for entity recognition.
• Cache-control headers set to 10-minute freshness — The cache-control: max-age=600, must-revalidate header ensures crawlers get reasonably fresh content without excessive server load.