digitas.com — AI Site Grade

Analysis

The digitas.com domain redirects to digitas.co.uk, creating a fragmented multi-site architecture where the US site (digitas.com/en-us) is a JavaScript shell invisible to AI crawlers while the UK site (digitas.co.uk) serves static content — but both share a single, AI-bot-agnostic robots.txt with no AI-specific directives.

Crawler Access

All AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, ChatGPT-User, Bytespider, Applebot-Extended) receive HTTP 200 from both digitas.com and digitas.co.uk with identical byte sizes to browser traffic. No UA-based blocking exists. However, the US site at digitas.com/en-us returns zero words of visible text to any crawler — it is a JavaScript-rendered single-page application (Drupal-based, served via AWS CloudFront + Varnish). Every US page (homepage, work, what-we-do, perspective, pressroom, case studies) has < 100 words extracted. The UK site at digitas.co.uk serves server-rendered HTML with 150-400 words per page. The robots.txt contains zero AI-bot-specific rules — only a generic User-agent: * with Drupal admin paths disallowed. No llms.txt exists (404).

Cold-Knowledge Gap

The LLM prior describes Digitas as a "global digital marketing and advertising agency" founded in 1980, specializing in data-driven marketing, with clients like American Express, Delta, and Samsung, and a proprietary "DigitasLBi" analytics tool. The actual site never mentions DigitasLBi, American Express, Samsung, or the 1980 founding year. The UK site brands itself as "The Connected Experience Company" with a "B2C, B2B, and B2Ai" positioning. The US site calls itself "The Networked Experience Agency." These two distinct taglines — "Connected Experience" vs. "Networked Experience" — create brand inconsistency that AI models must reconcile. The cold knowledge also references layoffs and restructuring (2023-2024) that the site does not address.

Schema Posture

The UK site (digitas.co.uk) uses a single Organization schema type on every page with minimal properties: name, URL, sameAs (LinkedIn, Twitter), and a contact point with a generic UK phone number. No WebSite, WebPage, Article, FAQPage, BreadcrumbList, or Service schemas exist anywhere. The US site (digitas.com/en-us) has a richer Organization schema on the homepage (with Boston address, telephone, and @graph structure) but zero schema on all subpages — case studies, perspective articles, and press releases have no JSON-LD at all. No CreativeWork or CaseStudy schema marks the portfolio work.

Content Architecture Fragmentation

The sitemap at digitas.com/sitemap.xml contains 6,500+ URLs — almost entirely locale-specific duplicates of the same ~20 case studies repeated across 50+ country/language paths (en-us, en-pa, en-nl, en-lt, en-be, en-lv, en-ee, en-pt, en-gr, en-th, en-id, en-ph, en-vn, en-kr, en-tw, en-tr, en-ro, en-gt, en-ng, es-co). This creates massive canonical ambiguity: the UK case study at digitas.co.uk/crocs has no canonical tag, while the US version at digitas.com/en-us/work/kid-crocs is a JS shell. The UK homepage carries a noindex, nofollow robots meta tag — meaning the primary entry point for the UK market is deliberately excluded from search indexes. The UK careers page is also noindex.

Findings

1. US site renders as JavaScript shell with zero visible text for AI crawlers High

   The US site at digitas.com/en-us is a JavaScript-rendered single-page application that returns zero words of visible text to any crawler, including AI bots. Every US page (homepage, work, what-we-do, perspective, pressroom, case studies) has fewer than 100 words extracted, making content invisible to AI models.

   What to change: Implement server-side rendering or static generation for all US pages to ensure content is accessible to crawlers without JavaScript execution.

2. UK homepage carries noindex, nofollow robots meta tag High

   The UK homepage at digitas.co.uk includes a noindex, nofollow robots meta tag, deliberately excluding the primary entry point for the UK market from search indexes and reducing discoverability for AI crawlers.

   What to change: Remove the noindex, nofollow meta tag from the UK homepage to allow indexing and crawling.

3. Robots.txt lacks AI-specific directives Medium

   The robots.txt file contains only a generic User-agent: * rule with Drupal admin paths disallowed and no AI-bot-specific rules for GPTBot, ClaudeBot, or others. No llms.txt file exists (404).

   What to change: Add AI-specific directives to robots.txt and create an llms.txt file to guide AI crawlers to key content.

4. US subpages have zero JSON-LD schema markup High

   The US site's subpages (case studies, perspective articles, press releases) contain no JSON-LD schema at all. Only the US homepage has a rich Organization schema. No CreativeWork, CaseStudy, Article, or WebPage schema marks portfolio work or articles.

   What to change: Add appropriate JSON-LD schema (e.g., CreativeWork, CaseStudy, Article) to all US subpages to improve structured data understanding.

5. UK site uses minimal Organization schema with no page-level markup Medium

   The UK site uses a single Organization schema type on every page with only name, URL, sameAs, and a contact point. No WebSite, WebPage, Article, FAQPage, BreadcrumbList, or Service schemas exist anywhere on the UK site.

   What to change: Add WebSite, WebPage, Article, and BreadcrumbList schemas to UK pages to provide richer structured data.

6. Sitemap contains 6,500+ locale duplicates with no canonical tags High

   The sitemap includes over 6,500 URLs, mostly locale-specific duplicates of the same ~20 case studies repeated across 50+ country/language paths. The UK case study at digitas.co.uk/crocs has no canonical tag, and the US version is a JS shell, creating massive canonical ambiguity.

   What to change: Implement canonical tags on all pages, consolidate duplicate locale paths, and reduce sitemap bloat to only unique content.

7. LLM cold knowledge contradicts site content on founding year, tools, and clients Medium

   LLM prior knowledge mentions DigitasLBi analytics tool, American Express and Samsung clients, and a 1980 founding year, none of which appear on the site. The site also does not address recent layoffs or restructuring mentioned in cold knowledge.

   What to change: Update site content to accurately reflect current tools, clients, and history, or add a 'About' page that aligns with the brand's actual narrative.

8. US and UK sites use different taglines creating brand inconsistency Low

   The UK site brands itself as 'The Connected Experience Company' with a 'B2C, B2B, and B2Ai' positioning, while the US site calls itself 'The Networked Experience Agency.' These distinct taglines create inconsistency that AI models must reconcile.

   What to change: Align taglines across all regional sites to present a unified brand identity.

9. UK careers page is noindexed Medium

   The UK careers page at digitas.co.uk/careers carries a noindex meta tag, reducing its discoverability for AI crawlers and potential candidates.

   What to change: Remove the noindex tag from the careers page to allow indexing.

10. Case studies lack structured data markup Medium

    Neither the US nor UK case study pages use CreativeWork or CaseStudy schema, missing an opportunity to clearly signal portfolio content to AI models.

    What to change: Add CreativeWork or CaseStudy schema to all case study pages.

11. Site has minimal web search presence for key queries Medium

    Web searches for 'Digitas agency 2024 2025 reviews clients', 'Digitas digital agency Publicis Groupe', and 'site:digitas.com Digitas agency' returned zero results, indicating poor search engine visibility.

    What to change: Improve SEO by ensuring content is indexable, adding meta tags, and building backlinks.

12. No llms.txt file available for AI crawlers Low

    The llms.txt file returns a 404, missing an opportunity to guide AI crawlers to key content and provide a summary of the site.

    What to change: Create an llms.txt file with links to important pages and a brief site description.

What's working

• All AI crawlers receive HTTP 200 with full content — All tested AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, etc.) receive HTTP 200 responses with identical byte sizes to browser traffic, indicating no UA-based blocking.
• UK site serves server-rendered HTML with substantial text — The UK site at digitas.co.uk serves server-rendered HTML with 150-400 words per page, making content accessible to crawlers without JavaScript.
• US homepage has rich Organization schema with address and contact — The US homepage includes a rich Organization schema with Boston address, telephone, and @graph structure, providing detailed business information.
• UK site consistently uses Organization schema with sameAs links — Every UK page includes an Organization schema with name, URL, and sameAs links to LinkedIn and Twitter, providing basic structured identity.
• UK case studies provide substantive text content — UK case study pages (e.g., Crocs, H&M, EE) contain 150-400 words of descriptive text, offering meaningful content for AI models to understand the work.
• Sitemap is available and contains 80 URLs — A sitemap exists at digitas.com/sitemap.xml with 80 URLs, providing a crawlable index of pages.
• DNS and hosting infrastructure is stable with CloudFront CDN — The site uses AWS CloudFront CDN with multiple A records and TXT records, indicating a robust hosting setup.