Sangue e Grafi: Teaching a Small Model to Read the Bloodline
Frontier LLMs fall for the story; a small model reads the graph. How ontology-guided GRPO taught Gemma 4B knowledge graph reasoning, field notes included.
Ontologies for the Agentic Web
How LLMs are changing ontology design from shared conceptualization to executable memory.
Perception Graph: How AI Models See Brands Before They Answer
AI Visibility is only the surface. Perception Graph reveals how AI models represent brands, Signal Graph maps the evidence they rely on, and Agent-Oriented Ontology Engineering (AOOE) turns perception gaps into governed knowledge, making the Knowledge Graph the memory layer for reliable AI agents.
Structure Is the Moat: How AI Finds, Navigates, and Ranks Your Content
Explore the SEO Week deck by Andrea Volpini, where he shares insights on how AI is reshaping search, content, and digital discovery.
From Frame Semantics to Natural Language Autoencoders: How AI Models Perceive Brands
What does a language model internally associate with a brand before it generates an answer? Using Natural Language Autoencoders, and Gemma 3, we explore how latent semantic representations shape AI Visibility and brand perception for Renault.
Do We Need LLM For Every Query? Separating Discovery from Ranking in the Era of Agentic RAG
Optimize Agentic RAG by separating discovery from ranking. Learn how RLM-on-KG and selective escalation scale Knowledge Graph search performance.
Sara Is All You Need: How Slow Shopping Shapes AI-Powered Decision-Making
Discover Slow Shopping: a philosophy for intentional AI decision-making. Learn how WordLift’s SARA agent uses multi-turn reasoning to prioritize human-centered commerce over speed.
Your Knowledge Graph Is Now a Search Space: How AI Agents Navigate, Not Just Retrieve
AI visibility is shifting from retrievability to navigability. Learn how AI agents use Knowledge Graphs as search spaces and explore the RLM-on-KG architecture.
Structured Data Is Not Enough: Why AI Search Needs a Memory Layer
Our latest research reveals Schema.org alone isn’t enough for generative engine optimization. Learn how structured entity hubs improve AI accuracy by up to 29.8%.
Why AI Cites Some Pages and Ignores Others
How AI systems like Google, OpenAI, and Perplexity AI retrieve, rank, and cite web content and how to structure pages so they remain visible in AI-powered search.
The Full Stack of the Agentic Web: Why WebMCP is the New Schema.org Moment
Until now, agents have relied on brittle techniques like visual scraping to guess at actions. This breaks easily and doesn’t scale. This is why the standardized Web Model Context Protocol (WebMCP) is critical.
RLM-on-KG: Recursive Language Models and the Future of SEO
Recursive Language Models (RLMs) treat prompts as environments to explore, not consume. We adapted this for Knowledge Graphs and discovered why structure, not bigger context windows, is the key to AI accuracy and search visibility.