Preparing UNICEF's 180+ Drupal Sites for AI-Driven Discovery

UNICEF wanted to: 

• Understand how major AI platforms, including ChatGPT, Gemini, Claude and Perplexity were interpreting and representing its content. 

• Identify what was preventing its pages from being cited or surfaced in AI-generated answers 

• Build a scalable approach to AI readiness that hundreds of independent country teams could execute without centralizing control. 

 Achieving this was not straightforward. Significant variance existed across markets: some pages were well-structured and crawl-friendly, others nearly invisible to LLMs. Structured data coverage was inconsistent, internal linking hierarchies varied, and E-E-A-T (Experience, Expertise, Authoritativeness and Trustworthiness) signals that AI engines rely on when deciding what to cite were unevenly applied across the 180-site network. At its core, UNICEF had no baseline view of how its content was being represented or omitted across major AI platforms.  

UNICEF engaged with Material to assess that risk, identify the gaps and build a roadmap for AI readiness across one of the world's largest decentralized content ecosystems. 

Our Solution 

Material designed and executed a four-stage AI Readiness and AEO/GEO Optimization program across UNICEF’s global digital estate. The program focused on identifying the structural and technical gaps affecting how AI systems interpret and surface content. 

• AI and technical diagnostics: Material audited 30–50 high-value pages per market, expanding across the broader ecosystem using Google Search Console, Google Analytics, Screaming Frog and Acquia SEO Conductor. The audit evaluated 28+ parameters including indexability, crawl budget, XML sitemaps, robots.txt directives, structured schema markup, internal linking, canonicalization, hreflang, semantic clarity, E-E-A-T signals and Core Web Vitals. The findings revealed significant variance across markets, which formed the baseline for the roadmap that followed.  

• Prompt & topic intelligence analysis: Material built a prompt intelligence framework using 400–500 real user prompts across 50+ topic clusters. The analysis mapped branded and generic information needs by persona and market and conducted cross-market intent analysis to understand how AI models summarize, rank and omit UNICEF content across different contexts.  

• LLM visibility assessment: Using manual prompt evaluations alongside AI-powered tools, Material measured website citations, topic-level citations, brand mentions, authority scores, and competitor and comparator visibility across various LLMs like ChatGPT, Gemini, Claude and Perplexity. Acquia SEO Conductor was used to track and analyze performance across the broader ecosystem. The result was the first unified view of how a decentralized humanitarian organization appears in AI-generated content, summaries and answers. 

• Roadmap and implementation planning: Findings were translated into a 12–24-month roadmap with prioritized technical and content recommendations, market-level scorecards, schema implementation plans, governance guidance and execution plans for distributed country teams. Material also established a measurement framework covering AI crawler behavior, LLM visibility, schema health and technical SEO performance to support ongoing monitoring and optimization. Material also delivered 10 capacity-building workshops to help country teams independently maintain and improve AI visibility. 

Outcome 

This engagement gave UNICEF a structured framework for managing AI-driven discovery across its global digital ecosystem. Material’s work: 

• Stronger citation readiness across major LLMs: Delivered 33 technical and AI audit reports alongside 100% completion of the strategic roadmap.  

• A repeatable AI readiness model across 180+ sites: Established a repeatable LLM visibility and governance model that can be adopted independently across UNICEF’s country sites.  

• Improved discoverability across AI platforms: Improved structured data, topic alignment and crawlability across the ecosystem, strengthening visibility and citation accuracy on major LLMs. 

• Consistent AI interpretation across markets and languages: Unified UNICEF's content through a shared knowledge graph, giving AI systems a coherent structure for interpreting it. 

• Greater control over AI-generated accuracy: Reduced the risk of misinformation and inaccurate AI summaries. This was critical for an organization whose content informs global policy, donor decisions and public understanding.  

• Scalable governance for distributed country teams: Delivered a scalable governance model with implementation guidance, shared playbooks and 10 capacity-building workshops for distributed teams.  

• Long-term ownership of AI visibility and discovery: Equipped UNICEF to monitor and manage how its content is interpreted, cited and surfaced as AI reshapes information discovery.